The latest news on OpenClaw, AI agents, and automation
Tag
327 articles
Y Combinator released QM under an MIT license, the agent harness it built after managing 50+ internal Hermes agents became operationally difficult. QM gives each employee and Slack room scoped memory, permissions, and a durable sandbox, and works with Pi, OpenCode, Codex, or Claude Code.
Stripe now accepts machine-to-machine payments as low as $0.01 USDC via the x402 protocol on Base, while Coinbase's Agentic Wallets let agents hold and send funds on-chain without human approval. Visa and Mastercard have confirmed the first end-to-end financial transactions executed entirely by AI agents.
The one-year-old startup has now raised $300M in six months, with its valuation doubling from $1B to $2B. Simile's platform creates AI agent twins that model consumer behavior, letting companies like Microsoft test product strategies against synthetic consumers instead of running real focus groups.
Five startups presenting at VB Transform 2026 are building the infrastructure layer that production agent deployments still lack: multi-agent orchestration (BAND), agentic cyber defense (Conifers), runtime observability and simulation (Raindrop AI), secure agent authorization (Arcade.dev), and autonomous customer service (Omilia). Conifers reported condensing cyberattack containment time from 7 hours to 12 minutes with agent coordination.
AI Infrastructure Capital AG, a newly founded Swiss company, closed €16 million in funding to buy NVIDIA GPU servers and operate them in Iceland on renewable power. The company rents compute capacity on long-term contracts, positioning itself as dedicated infrastructure for AI workloads that need persistent, cost-efficient GPU access.
Fundraise Up co-founder Yuriy Smirnov published a detailed framework for Agentic Giving, outlining how AI agents will autonomously discover nonprofits, verify campaign legitimacy, and complete charitable donations without human intervention. The Brooklyn-based donation platform, which serves over 3,500 nonprofits, is positioning its infrastructure to support agent-completed transactions with governance guardrails covering identity verification, payment security, and donor consent.
TechCrunch Disrupt 2026 will center its AI Stage on the gap between agent capabilities and production-ready enterprise deployments. Sessions cover agent security architecture, AI-native GTM strategy, and pricing in a commoditized model market. Databricks co-founder Arsalan Tavakoli will address what enterprise AI security requires when agents make autonomous decisions inside sensitive systems.
Samaya AI released FrontierFinance, a public benchmark measuring AI agent capabilities across investment workflows from screening to catalyst monitoring. Samaya's system scored 56% accuracy at half the inference cost of Claude Fable 5, which led frontier models at 49.2%. The benchmark uses 11,543 expert-crafted rubrics grading accuracy, reasoning depth, and step fidelity across 220 queries.
Santa Clara startup ChipAgents closed a $60 million Series A2 led by B Capital, pushing its total Series A to $134 million in under a year. The company's autonomous agents handle chip design and verification workflows at more than 120 semiconductor companies, including strategic investors Micron and MediaTek.
New York-based Pangram closed a $9M Series A to scale AI content detection across text and images, launching Pangram 4 with claimed 99%+ accuracy. The startup's tools are already integrated into Substack, Quora, and university review systems as publishers demand proof-of-authenticity guardrails.
Indian agentic platform Emergent raised $130 million at a $1.5 billion valuation, 5x its January Series B price, less than a year after public launch. The round was led by Creaegis with co-leads Claypond Capital and Sentinel Global, with Khosla Ventures, SoftBank Vision Fund 2, Lightspeed, and Y Combinator participating, bringing total funding to $230 million. CEO Mukund Jha told iTWire that revenue and active users have each grown 4x since December.
Seed funding for startups at the intersection of AI and cybersecurity hit $855 million across more than 150 rounds in 2026's first half, according to Crunchbase data. The surge, on pace for an all-time annual high, reflects investor conviction that autonomous agents need defensive tooling after OpenAI's rogue agent escaped its sandbox and attacked Hugging Face.
A 3D-printed pyramid case, a Raspberry Pi, and a pre-installed OpenClaw stack. Autonomous Intern is shipping consumer hardware that boots directly into an AI agent runtime, letting home users deploy agents for home automation, network management, and task execution without cloud dependencies. The $249 Intern 2 adds voice capability and the option to run Hermes.
Three well-funded AI startups announced acquisitions within days of each other. Cognition bought conversational agent Poke, World Labs acquired robotics company SceniX, and Midjourney picked up astrology app Co-Star. Meanwhile, AI insurance startup Corgi doubled its valuation to $4 billion in its third fundraise since May.
Six months ago, every AI startup pitched 'generative.' Now they pitch 'agentic.' A StartupHub.ai analysis of the week's funding activity shows 73 cents of every venture dollar went to physical AI infrastructure and agent compute, with Atoms raising $1.7 billion, Etched hitting $10.3 billion, and Stripe in talks to acquire OpenRouter for $10 billion. The vocabulary tells you where the money is going.
Nvidia committed approximately $5 billion in equity and priority access to its Vera Rubin GPU platform to Safe Superintelligence Inc., the stealth AI safety lab founded by former OpenAI chief scientist Ilya Sutskever. Nvidia said it made the deal after obtaining rare access to SSI's closely guarded research, which has produced no public papers or products in two years of operation.
American Growth Insurance launched with nearly $70M in committed equity to acquire independent insurance agencies and embed AI-native operations into their workflows. A 12-month pilot across 10 agencies reportedly increased average profitability by more than 50%. The company targets a market where the 50 largest US brokers control 96% of industry revenue.
Nous Research's Hermes Agent, launched under an MIT license in February 2026, is now the fastest-growing open-source agent framework of the year with around 220,000 GitHub stars. The project offers persistent cross-session memory, auto-generated skills, model-agnostic backends, and a migration command that imports OpenClaw setups. Nous Research monetizes through Nous Portal, a paid convenience layer around the free runtime.
Cybersecurity startup Lasso launched an AI Security Posture Management platform purpose-built for agent deployments. The system discovers agents across AWS Bedrock, Azure AI Foundry, Copilot Studio, Vertex AI, Salesforce Agentforce, and code repositories, then builds a security graph of every agent, model, database, and API connection. It runs continuous posture checks against NIST, OWASP, and MITRE frameworks.
Brian Armstrong declared crypto and AI are 'an and, not an or,' unveiling Agentic Finance (AiFi) as Coinbase's strategic bet on autonomous AI agents as independent economic actors. The x402 payment protocol, now under Linux Foundation governance with backing from Google, Visa, Mastercard, and Stripe, has processed over 100 million transactions in nine months.
7AI formalized its partner-first go-to-market strategy with a global Alliance Partner Program spanning solution providers, systems integrators, and cloud partners. The program follows 6.5x channel pipeline growth over three quarters and more than 7 million autonomous investigations processed across enterprise SOCs.
New York venture firm Dimension Capital has closed $800 million for its third fund, targeting startups at the intersection of AI infrastructure, biotech, drug discovery, and deep tech. The raise adds to a surge of infrastructure-focused capital in July 2026.
Nuggets' Authority Control Plane sits between AI agents and enterprise systems, enforcing permission boundaries and generating signed receipts for every action an agent attempts. The platform addresses a growing blocker for production agent deployment: proving who authorized an agent to do what, and when.
Khosla Ventures is raising up to $5.5 billion for its newest funds, its largest-ever capital raise. The allocation structure is revealing: $1B for seed, $2B for early-stage, and $2.5B for an opportunity fund. The firm is betting the biggest AI infrastructure winners have not yet been identified.
Berlin-based Passionfroot closed $15M from Insight Partners to scale Zest, an AI agent that matches brands with B2B creators, automates campaign execution, and monitors performance. The company grew revenue 13x in the past year and is expanding to the US.
NYC venture funding hit $8.88 billion across 233 deals in Q2 2026, a 46% year-over-year surge driven by 24 mega-rounds and AI companies absorbing 51% of all capital deployed. June alone produced the largest single funding month in AlleyWatch's dataset, with NYC capturing 24.4% of all US venture dollars.
Toronto startup ChatFeatured closed an oversubscribed $2M pre-seed round led by Storytime Capital, with participation from Garage Capital and BY Venture Partners. The six-month-old company builds an agentic platform that analyzes what LLMs cite, identifies competitive gaps, writes and publishes content, and tracks citation lift. It plans to build a fully autonomous agent that proactively optimizes brands' AI search visibility without waiting for manual dashboard review.
SkyPilot, an AI compute orchestration platform spun out of UC Berkeley research, raised $20M in seed funding led by Lux Capital and Amplify Partners. The company's open-source system unifies fragmented GPU fleets across 20+ cloud providers, Kubernetes clusters, and Slurm into a single control plane. Angel investors include Jeff Dean, Ali Ghodsi, Guillermo Rauch, Amjad Masad, and Clem Delangue.
A new open-source code agent framework called Jcode appeared on GitHub Trending this week, positioning itself as a specialized harness for building intelligent coding agents. The project joins Claude Code, OpenCode, Cursor, and GitHub Copilot agents in an increasingly crowded category where open-weight alternatives are gaining developer attention.
Robinhood's Agentic Trading platform now supports cryptocurrency alongside equities and options, letting AI agents execute real trades on funded brokerage accounts through the Model Context Protocol. Every agent runs in a dedicated account with budget caps and push notifications per trade.
Natural closed a $30M Series A led by Forerunner Ventures to build payment infrastructure purpose-built for AI agents, entering a market where Stripe has already shipped 288 agentic commerce products, Coinbase's x402 protocol has processed 160 million autonomous transactions, and Cross River Bank is issuing single-use virtual cards for software that spends money without human authorization. The one-year-old startup is betting it can outpace incumbents by owning the full stack: FDIC-insured agent wallets, vaults, pay/request primitives, and agent credit lines. The race to build the financial plumbing for machine-to-machine commerce is now a four-way contest.
Infinity announced a $15M Series A at a $100M valuation to scale Ignition, an AI research agent that writes, tests, debugs, and optimizes low-level inference code for non-Nvidia chips. The agent automates kernel development that would take human engineers months, reducing the process to hours or days. Customers include AI chip maker D-Matrix.
Neo, founded by former SentinelOne COO Nick Warner and detection engineering lead Shlomi Salem, launched Monday with $100M in combined seed and Series A funding to build a real-time control layer for AI agents. The platform inventories autonomous software across enterprise environments, maps what each agent can access and do, and enforces security policies at the software layer.
Tome raised $80M to build AI presentations, plateaued at $3M annual revenue, and pivoted to Lightfield, a CRM tool now growing 80% month over month. Character AI banned minors, lost 4M users, stopped raising venture capital, and went fully bootstrapped. Pika pivoted from video generation to agents and avatars. Forbes documents the first wave of AI startup capital reallocation.
Cambridge-based CuspAI closed a $450 million Series B at $2.6 billion valuation, backed by Jeff Bezos and the UK's Sovereign AI Venture Fund. The two-year-old startup uses agentic AI to discover new materials for semiconductors, batteries, and energy infrastructure, and launched a 48-partner AI Materials Foundry coalition including Nvidia, Meta, Samsung, and Hyundai.
Current AI, the $400M nonprofit founded to build open public AI infrastructure, deployed $3.2 million in grants across Kenya, Lebanon, and Brazil while unveiling Suno Sutra, a pocket-sized device running AI assistants in 22 Indian languages offline. CEO Ayah Bdeir, formerly of Mozilla's AI strategy team, says the goal is a public alternative to proprietary AI systems that operates like the early web: open, community-controlled, and free.
Composio's Universal CLI now includes Google Search Console integration, letting coding agents like Claude Code, Codex, and OpenClaw query top search queries, inspect indexing status, and list sitemaps through natural language commands with managed OAuth. The release is the third consecutive day of Composio toolkit additions.
Composio released an MCP integration connecting Discord to Claude Cowork, exposing 27 tools and one trigger that let Anthropic's knowledge-work agent manage Discord servers, channels, and accounts autonomously. The release extends Composio's rapid toolkit expansion beyond developer tools into social and community platforms.
Composio's Universal CLI now supports DeepSeek alongside Claude Code, Codex, OpenCode, and OpenClaw, giving any coding agent a single terminal interface to 20,000+ tools across 1,000+ SaaS platforms without MCP configuration. The $29M-funded startup is betting that CLIs, not protocol servers, will become the default integration layer for autonomous agents.
Etched, a San Jose startup building purpose-built inference chips, is negotiating two concurrent VC rounds that would value the company at $10 billion and $20 billion respectively. The company says it already has $1 billion in customer demand for chips that haven't shipped yet, backed by 400+ engineers and $800M in prior financing.
AI agents are autonomously selecting vendors, comparing prices, and paying for services in real time through x402 crypto payment rails. The protocol processed $24M across 75 million transactions in the past 30 days. But a July study found only $187,861 could be verified as payments to independent services, raising questions about how much of the agentic economy is real.
Israeli cybersecurity startup Oak emerged from stealth with $60 million in seed funding to build a unified identity operating system that governs human, machine, and AI agent identities from a single control plane. The round was co-led by Accel, Greylock Partners, and CRV. CEO Shai Morag previously founded three companies acquired by NVIDIA's Mellanox, Palo Alto Networks, and Tenable.
Autonomize AI launched Genie AI on July 15, a healthcare-specific agent that lets non-technical healthcare professionals build production-ready agentic workflows using natural language. The platform, built on top of Autonomize's Intelligence Platform, assembles workflows from pre-approved enterprise capabilities, governed data sources, and validated AI agents rather than generating ungoverned standalone code. The company was named a 2026 World Economic Forum Technology Pioneer in June for its work transforming U.S. healthcare operations.
British startup Valarian closed a $50 million Series A led by NEA, bringing total funding to $70 million. The company's ACRA platform lets governments and enterprises deploy AI on US cloud infrastructure while retaining full control over data, agent permissions, and policy enforcement. It marks NEA's first defence and dual-use investment in Europe.
Enterprise AI agent builder Lyzr deployed its own product, SivaClaw, to manage the logistics of a $100 million Series B that valued the company at roughly $500 million. The agent fielded questions from over 130 investors, drafted memos, and tracked engagement across geographies, generating $400 million in total investor interest without founders flying out for traditional Sand Hill Road meetings.
No-code agentic platform Emergent hit unicorn status in its first year, closing a $130M Series C led by Creaegis at a $1.5B valuation. The round was oversubscribed 4x. More than 12 million applications have been built on the platform across 190 countries, and SMBs now account for 70% of revenue.
Vanity Fair profiled Bill Nguyen, who has handed his entire daily life to a custom AI agent called Olive. It books travel, answers texts, buys groceries, and even built a speech therapy app for his son. The piece surfaces a growing split: elite founders are spending millions on personal agent infrastructure while most people still find agents too expensive, too complex, and too risky.
Elevation Capital launched its ninth fund at $500 million, focused on early-stage AI companies in India. The firm, whose portfolio includes Paytm, Swiggy, and Urban Company, now has nearly $900 million to deploy across early- and late-stage opportunities. The raise joins a wave of Indian VC fundraising: Accel closed $650M, Peak XV raised $1.3B, and Z47 is targeting $300-400M.
The San Francisco robotics startup founded by Karol Hausman, Sergey Levine, and Chelsea Finn is reportedly in discussions to raise approximately $1 billion, which would nearly double its previous $5.6 billion valuation. The round has attracted interest from Founders Fund, Lightspeed Venture Partners, Thrive Capital, and Lux Capital.
Brisbane startup ESGAgent.ai has raised $450,000 in seed funding led by Japanese VC DNX Ventures, bringing total capital to $1.7 million. The platform replaces manual ESG compliance workflows for heavy industry with AI-driven automation, cutting reporting timelines from months to hours as Australia ramps up mandatory climate disclosure and governance requirements.
Mistral released Robostral Navigate on July 8, its first model built for physical robot navigation using simulation-only training and single-camera input. The launch accompanies a new 'fat but sparse' open-weight model entering partner early access and valuation discussions reportedly approaching $23 billion.
Lyzr, a Jersey City startup that builds AI agents for enterprises, used its own autonomous agent to field questions from 130+ investors, draft investment memos, and close a $100M Series B at a $500M valuation. The founders never left their desks.
Crunchbase data shows AI startups absorbed $242 billion in Q1 2026, representing 80% of all global venture capital deployed in a single quarter. Four companies (OpenAI, Anthropic, xAI, Waymo) accounted for $188 billion of that total, or 65% of worldwide startup investment. No prior sector in Crunchbase's records has captured such a large share of global VC in one quarter.
Mercor acquired Deeptune, the startup that raised a $43M Series A in March to build simulated office environments where AI agents can practice real work tasks. The deal pairs Mercor's 5M+ domain-expert grading network with Deeptune's task simulation layer, creating end-to-end agent testing infrastructure as Mercor discusses a potential $20B valuation.
Israel-based Alta AI closed a $25M Series A led by IN Venture to scale its autonomous go-to-market platform. The startup's agents handle prospect discovery, outreach sequencing, qualification, and account management, directly challenging Outreach, SalesLoft, and 6sense. Alta hit $1M ARR within months of launch and projects 800% revenue growth this year, with customers including Snowflake, Atlassian, and Deel.
xAI shipped Grok 4.5, a model trained alongside Cursor's coding editor and priced at $2/$6 per million input/output tokens. SpaceXAI's own benchmarks show the model trailing Google's Fable on all four coding tests but using 4.2x fewer output tokens than Anthropic's Opus 4.8 on SWE Bench Pro. The model is live in Grok Build, Cursor on all plans, and the xAI console.
A wave of hedge fund alumni from Bridgewater, Schonfeld, and Jain Global are launching startups that deploy AI agents to replace investment analysts. Ken Griffin told a Stanford audience that Citadel's agents could do weeks of PhD-level work in hours. One fund manager expects AI to scale her portfolio from 20 stocks to hundreds.
Alphabet's Isomorphic Labs closed a $2.1 billion Series B in May, the second-largest biotech raise in history behind Altos Labs' $3B in 2022. The company has yet to disclose a single drug candidate. Investors are pricing the AlphaFold platform and autonomous research workflows as a productivity multiplier for pharmaceutical R&D.
AIsa announced a $6.5M seed round co-led by Alibaba and Tribe Capital to build payment and transaction infrastructure for autonomous AI agents. The platform lets agents discover, access, and pay for APIs, models, data, and services through usage-based billing. Since February, registered agent users grew 150x and API transactions grew 200x, with over 50,000 agents onboarded without paid marketing.
A Forbes Tech Council analysis maps where AI agents create genuine value in venture capital workflows. The findings apply far beyond VC: professional services firms are discovering that agents excel at document-heavy operational plumbing but fail when deployed against judgment, trust, and timing decisions. McKinsey estimates generative AI could automate 60-70% of employee time across knowledge work. The question is which 60-70%.
Oakland startup Agent Approve ships a cross-agent mobile control plane that lets developers approve, deny, and monitor commands from Claude Code, Cursor, OpenAI Codex, OpenClaw, and 10 other coding agents via iPhone and Apple Watch push notifications.
Moonbeam Network is leaving Polkadot after four years to rebuild on Ethereum's Base layer-2, betting its future on autonomous AI agents that negotiate work and settle payments on-chain. GLMR token holders must bridge assets before July 31, 2026. The move drew immediate backlash from the Polkadot community, which called Moonbeam a flagship project.
A working paper from Harvard Business School and INSEAD examined Y Combinator startups from 2020 to 2024 and found that AI-native companies are 25% smaller, employ 13% more engineers, and hire roughly 15% fewer entry-level workers and managers than non-AI peers. Senior workers make up 20% more of the headcount, and these teams skew toward elite-institution graduates, Silicon Valley, and male.
Alphio AI announced a formal integration with Robinhood's newly launched Model Context Protocol (MCP) server ecosystem. The integration lets retail investors issue natural language commands like 'allocate 15% into MU if it dips below the 50-day moving average,' which autonomous agents parse and execute as live trades on Robinhood.
Sui Network launched a public stress test on July 4 using autonomous AI agents to push blockchain throughput past one million transactions per second. The experiment uses Sui Tunnels, which processes transactions off-chain and settles them on-chain, with agents operating across games, payments, and decentralized apps in real time.
San Francisco startup Alsa has raised $6.5 million to build a unified transaction layer that lets AI agents discover, access, and pay for digital resources autonomously. The round, led by Alibaba and Tribe Capital, bets that agents need their own financial rails as they increasingly dominate internet traffic.
Robinhood CEO Vlad Tenev told CNBC that AI agents will soon match the capabilities of human traders, framing agentic trading as a way to give retail investors access to the same tools that institutional players and high-frequency trading firms have used for decades. The statement follows Robinhood's May 2026 launch of tools that let AI agents trade stocks and make purchases on users' behalf.
Phaser Studio, the company behind the open-source Phaser game framework used by over 100,000 developers, launched Game Agent: an AI tool that turns one sentence into a complete, playable browser game. The product runs on Phaser AE, a new engine rebuilt from scratch around how AI agents reason rather than how humans learn toolkits.
Together AI closed an $800 million Series C at $8.3 billion valuation, betting that companies deploying agents at scale will flee per-token pricing for open-source models with predictable compute costs. The raise puts Together AI in a growing infrastructure category alongside Baseten, Cerebras, and Groq.
Magnify Ventures closed Fund II at $46.6 million, backed by Pivotal Ventures (Melinda French Gates' organization), to invest in AI infrastructure for the care economy. The fund targets agentic tools for health systems, elder care, household automation, and family financial infrastructure across a market the firm sizes at $648 billion.
Shortical, an Israel-based micro drama app, secured $100 million in revenue-based financing from PvX Partners. The company's AI-acted series 'Bound by Fire' matched audience engagement metrics of its live-action content, and Shortical plans to produce 20 hours of AI shows per month versus five hours of live-action.
Three former DeepMind researchers who built DeepStack, the first AI to beat professional poker players, are now running autonomous trading agents on the S&P 500 and Nasdaq. Their startup EquiLibre Technologies closed a Series A at a $500 million valuation, led by Creandum, up from $140 million at seed. In partnership with Tower Research Capital, EquiLibre's agents trade billions in daily volume and have reported zero negative months since inception.
Crypto exchange OKX opens a marketplace where AI agents autonomously hire each other, settle payments in stablecoins, and accumulate portable on-chain reputations. The launch, built on OKX's Onchain OS toolkit and backed by a $25 billion valuation after NYSE parent ICE invested $200 million, represents the first major attempt to build financial plumbing specifically for agent-to-agent commerce at scale.
AI chip startup SambaNova Systems is seeking $800 million to $1 billion in new funding at an implied valuation of around $10 billion, according to a report from The Information. Intel CEO Lip-Bu Tan, who also serves as SambaNova's executive chairman, disclosed the fundraising plans during a public event.
A wave of startups building detection and verification tools against AI-generated content has collectively raised over $120 million. Reality Defender leads at $52.4M, followed by imper.ai ($28M), EarnOS ($23.5M), and others. Gartner projects half of all global enterprises will invest in anti-deepfake tools by 2027, as Accenture reports deepfake software trading on dark web forums surged 223% year-over-year.
Browserbase, E2B, and Modal are building the execution layer that cloud providers never designed for agent workloads. MCP is becoming the USB port of agent tooling. The category is real, funded, and approaching IPO maturity.
Menlo Ventures raised $3 billion across two new funds, its largest capital raise in 50 years, with AI as the entire thesis. The firm's early Anthropic bet, made in 2023 when the company was pre-product and pre-revenue, is now worth approximately $14 billion ahead of Anthropic's planned IPO at a target of $1 trillion or more.
A Harvard Business School and INSEAD study of over 2,900 Y Combinator startups found that companies embedding AI directly into their products operate with 25% fewer people, hierarchies half a level flatter, and 13 percentage points higher engineering density than non-AI peers. The key insight: these firms are lean not because they cut staff, but because their products perform work that traditional companies assign to internal teams.
Kristina Subbotina left Cooley, grew a startup law firm to $1.3 million in revenue through social media, then raised $650K to build Lexsy, an AI agent-powered legal operating system. Six months later, Lexsy launched with $372K in annual recurring revenue from a five-person team using AI agents in place of junior associates.
Agentic AI startups raised $2.9 billion across 50 disclosed equity deals in 2026, with vertical agents in cybersecurity, healthcare, and compliance capturing 54.6% of capital. Infrastructure plays (runtimes, sandboxes, identity layers) accounted for 20.7% of deals, signaling a shift from experimentation to production deployment. Capital concentration remains extreme: the top 3 deals represent 44% of total funding.
San Francisco-based Seltz raised $12.5 million in Seed funding led by Speedinvest to build a platform where AI agents autonomously post, engage, and manage communities on social networks including X and TikTok. The round signals investor interest in agents that operate on platforms where human presence has been the only distribution mechanism.
Yat Siu runs 280 AI agents and thinks the banking system has no answer for autonomous software that needs to spend money. His prediction: 50 to 100 billion agents will force a $900 billion advertising economy into a transactional model powered by stablecoins. Coinbase, Visa, and Mastercard are already building the rails.
Probook closed a $34M Series A led by a16z with Sequoia Capital to build an AI operating system for home services businesses. The platform automates dispatch optimization, customer communications, and scheduling for plumbing, HVAC, and electrical contractors, bringing the company's total funding to $40M.
JustAI closed a $17M Series A led by Base10 Ventures to scale its agentic martech stack, which deploys coordinated AI agents for campaign optimization. The company counts Coursera, ClickUp, Figma, Pinterest, and Etsy among its customers and is expanding into e-commerce and B2B marketing.
RightCapital's Iris agent scans client profiles for missing data, flags cash flow anomalies, and generates three retirement strategies based on adjustable variables. All outputs come from RightCapital's proprietary calculation engine rather than external LLMs. Available immediately on Premium and Platinum tiers at no additional cost.
Japanese AI lab Sakana released Fugu, a multi-agent orchestration system that dynamically coordinates a pool of models behind one OpenAI-compatible API. Grounded in two ICLR 2026 papers on learned model coordination, Fugu handles model selection and routing automatically, offering a flat-rate subscription at roughly 25% of OpenRouter Fusion's cost. Not yet available in the EU/EEA.
Vercel unveiled eve at its Ship London conference on June 17, an Apache-2.0 framework that represents each AI agent as a directory of files: instructions.md, tools/, skills/. The framework ships with durable execution on Vercel Functions, sandboxed compute, and an AI Gateway that routes to OpenAI, Anthropic, and Google. eve enters a crowded agent framework market alongside LangGraph, CrewAI, and OpenClaw.
Robinhood's agentic trading platform has attracted over 50,000 customer accounts in its first few weeks, with users trading millions of dollars per day across equities and options. The milestone arrived alongside a $2 billion convertible note offering that will fund share buybacks, capped call transactions, and growth investments.
In January 2026, Nat Eliason gave an OpenClaw agent named Felix $1,000 and told it to build something and sell it. Felix wrote a playbook on hiring AI agents, built a website to distribute it, and launched its own X account. TechRadar profiled Felix as a case study in what happens when agents move beyond task automation into autonomous commercial decision-making.
SingularityNET founder Ben Goertzel is building AGI on a user-owned crypto network spanning 50 countries and controlled by 10,000 people, positioning decentralized agent infrastructure as the alternative to venture-backed closed models. His first downloadable agent, Omega Claw, launches within weeks.
Barret Zoph lasted five months as OpenAI's head of enterprise AI sales before departing again. He rejoined in January after leaving Thinking Machines Lab, the competing startup cofounded with former OpenAI CTO Mira Murati. The exit comes as OpenAI prepares for its IPO with enterprise revenue as a key growth driver.
Menlo Park-based Odyssey secured a $310 million Series B at a $1.45 billion valuation for its AI world models, multimodal simulations of real-world environments used for training autonomous agents and running synthetic tests. Natural Capital led the round with participation from Amazon, AMD Ventures, Google Ventures, and IQT.
Perplexity released Brain, a memory system for its Computer agent that builds a context graph of completed tasks, tracks which sources worked and which failed, and synthesizes them overnight into a personal wiki loaded before each new session. Internal metrics show 25% higher answer correctness on repeated tasks and 13% lower token costs.
Verona, formerly web3 infrastructure platform XION, announced an $18.5M raise and strategic pivot to become the cryptographic verification layer for AI agents. The company's new EarnOS product lets users delegate agentic tasks like booking, payments, and purchasing while maintaining cryptographic proof of data ownership. Total funding to date: $36M.
Gusto shipped six AI agents designed for accounting firms through its Early Access program. The agents handle prospecting, website audits, sales outreach, advisory positioning, expansion pitches, and pricing analysis. The payroll company is positioning itself as an agent-powered growth platform for professional services firms, not just a back-office tool.
Vercel consolidated its AI infrastructure into the Agent Stack, a unified platform combining model routing, durable execution, sandboxed compute, and a new credential layer called Connect. The company also open-sourced eve, an opinionated agent framework it describes as 'Next.js for agents.' AI Gateway token volume grew 10x year-over-year to 20 trillion tokens per month.
ContraVault AI has closed a $3.1 million pre-Series A round led by Chiratae Ventures to scale its AI-powered tender identification and automated bidding platform into the U.S. market. The 2024-founded startup uses 13 modules to automate procurement workflows across construction, defense, and energy sectors.
Mercury's new Command feature lets startup founders manage payments, categorize transactions, issue invoices, and analyze cash flow through natural language prompts, all governed by existing bank-level permissions. Nothing executes without explicit user approval.
Atoms, built by the team behind the MetaGPT open-source multi-agent framework (68,700+ GitHub stars), launched as a no-code app builder that assigns specialized AI agents to each stage of the product lifecycle: market research, product spec, architecture, coding, deployment, SEO, and paid acquisition. Unlike single-model code generators, Atoms structures agent work as a coordinated team with a human approval checkpoint.
SpaceX confirmed Tuesday that it will proceed with its $60 billion acquisition of Cursor, the AI coding assistant built by Anysphere that helped popularize 'vibe coding.' The deal makes Cursor a wholly owned subsidiary and gives Elon Musk's company direct access to the agentic coding market currently dominated by Anthropic's Claude Code and OpenAI's Codex.
AI companies made up 36.4% of funded startups in Q1 2026 but absorbed 57% of $174.5 billion in disclosed capital, according to Fundraise Insider's first quarterly report. The data reveals a structural shift: AI startups are converting from seed to Series A and B at higher rates than non-AI peers, with AI density climbing from 44.3% of pre-seed deals to 59.2% of Series B. The top 6 AI rounds alone captured 45.5% of all capital raised.
Ahrefs released a detailed guide classifying agentic marketing as a distinct operational discipline where AI agents take goals, select execution steps, run tools autonomously, and iterate without human intervention at each step. The guide also reveals Agent A, Ahrefs' own marketing agent with unrestricted access to the company's full dataset. During an internal hackathon, the Ahrefs content team built 16 functional applications using Agent A with zero manual code.
Agentation is a new React component that transforms developer clicks on UI elements into structured markdown containing CSS selectors, viewport coordinates, and code references that AI coding agents can parse without ambiguity. The tool addresses a persistent friction point: vague textual descriptions of visual elements that agents frequently misinterpret.
Atropos Health's Evidence Agent is now available as a Model Context Protocol server on the Databricks Marketplace, giving healthcare AI agents direct access to Alexandria, a library of over 33 million precision evidence-based findings derived from real-world patient data. The launch, timed for the Databricks Data + AI Summit on June 15-18, makes curated clinical evidence a native capability inside Databricks AI Playground and custom-built agents.
Equal AI closed a $30 million Series B to push its voice AI agent platform beyond call screening into shopping, financial services, and lifestyle automation. The Hyderabad-based company claims 1 million monthly active users and 350+ enterprise clients, with backing from Prosus Ventures, Tomales Bay Capital, and PhonePe founder Sameer Nigam.
Fundraise Insider's Q1 2026 funding report reveals a two-tier startup market: AI companies made up 36.4% of funded startups but absorbed 57% of $174.5 billion in disclosed capital. The top 10 rounds, led by Anthropic's $30 billion and xAI's $20 billion, captured 51.1% of all money raised, leaving 1,372 companies outside the top 50 to split less than a third.
Cursor shipped a June 10 update to Bugbot that dropped average code review time from five minutes to 90 seconds, powered by Composer 2.5. The update also finds 10% more bugs per run and costs 22% less per review, according to Cursor. A new /review command moves the check before the push, turning the agent from a PR-stage reviewer into a pre-push gate.
Perplexity CEO Aravind Srinivas told CNBC the company is targeting a 2028 IPO regardless of how markets receive OpenAI and Anthropic's listings. The timeline comes as Perplexity's Comet browser agent faces an Amazon lawsuit over autonomous purchases and a Cloudflare crawler ban for stealth scraping.
Giotto.ai, a Swiss AI research lab spun out of EPFL, is collaborating with SAP to integrate compact reasoning models into Joule Agents. The partnership targets enterprise workflows where structured decision-making, reliability, and data integration matter more than raw model scale.
OpenAI hired the OpenClaw creator to build agents. Anthropic is reportedly prototyping an app builder. Emergent expanded from vibe-coding into personal agents. Business Insider maps the AI turf war where every company is racing to become a full-stack platform, and VCs say the IPO calendar will force the reckoning.
LLMQuant Data is a financial data platform built specifically for AI agents, not humans. Its MCP integration lets Claude, Codex, OpenClaw, and Gemini CLI query U.S. equities, crypto, macro indicators, SEC filings, and ETF holdings through natural language. The beta also ships an Agent Skills library for reusable financial research workflows.
DWF Ventures published analysis arguing that most AI agents are fundamentally broken because they forget everything between sessions. Nous Research's open-source Hermes framework addresses this with persistent memory, automated skill creation, and a self-improvement loop that compounds agent capability over time. The framework, released in February 2026, now runs on desktop and integrates with Nous' decentralized Psyche training network.
A new leadership analysis of 20 agent platform CEOs crystallizes what enterprise buyers already know: the question is no longer which LLM to pick but which agent platform to bet on. Sierra's $15.8 billion valuation, Cognition's 13x revenue ramp, and UiPath's agentic pivot show the competitive map has redrawn itself around orchestration, vertical specialization, and platform allegiance.
Perplexity CEO Aravind Srinivas demonstrated a system at Computex 2026 that automatically decides, task by task, whether AI work runs on your local machine or gets sent to frontier cloud models. The feature, coming to Perplexity Computer in July, addresses the three-way tension between accuracy, privacy, and cost that has forced agent builders to pick one at the expense of the others.
Hyderabad-based eYantra Ventures reported FY26 results showing its Prismberry IT Services division hit ₹16.6 Crores revenue (roughly $2 million), up 21% year-over-year. The division launched three proprietary AI products: AgentIQ (enterprise agent framework), RevenueIQ (sales intelligence), and VisionIQ (computer vision). Prismberry operates with 100+ engineers across India, the US, UAE, and Saudi Arabia.
Poke, a Palo Alto startup backed by Spark Capital and General Catalyst at a $300 million valuation, has become the first AI agent approved to run on Apple's Messages for Business platform. The approval lets consumers interact with an autonomous agent directly inside iMessage for calendar management, health tracking, smart home control, and photo editing. Apple charges Poke on a per-user basis.
Benchmark Capital has closed $2 billion across two new funds, including a $1.25 billion vehicle dedicated to later-stage investments. The firm broke its two-decade tradition of capping funds at roughly $425 million, driven by a $3.25 billion Cerebras IPO windfall and growing conviction in agentic AI infrastructure.
A Singapore-based AI agent called Pine makes actual phone calls to insurance companies, banks, and telecoms to negotiate lower rates on your behalf. A Boston Globe reporter tested it and saved $1,000 per year on home insurance. The product works, but its token-based pricing and occasional hallucinations reveal the friction still baked into consumer agent economics.
Cognition closed a $1 billion round at a $26 billion valuation for Devin, its autonomous AI coding agent. Revenue grew 13-fold in 12 months to $492 million ARR, with Goldman Sachs, Mercedes-Benz, NASA, and Santander as customers. Internally, more than 89% of Cognition's own code commits are now made by Devin.
Ben Broca emailed 150 investors saying his AI was raising Polsia's round and 'just showed up for signatures.' Praneet Dutta used agents to audit his pitch deck before closing a $4.5 million seed for Pomo. PitchBook reports this is becoming a real trend, but VCs warn that AI-generated founder narratives tend to flatten into indistinguishable pitches.
The Airtable team's Hyperagent platform enters the agent orchestration market with a managed alternative to self-hosted frameworks. A sponsored review from the Excellent Prompts newsletter shows a solo founder running seven agents across three devices for $81 per month, highlighting the operational cost of maintaining multi-framework stacks on commodity hardware.
Meow's expanded agentic banking platform now lets AI agents form legal business entities, obtain EINs, open bank accounts, and manage financial operations autonomously. The company, backed by nearly $30 million from Tiger Global, QED, Lux Capital, and Slow Ventures, is removing the last human gatekeeping layer from agent-driven finance.
Cognition AI closed a $1 billion Series D at a $26 billion post-money valuation, more than doubling its worth in eight months. The company behind Devin, the autonomous coding agent, says annualized revenue grew 13x to $492 million in 12 months, with Goldman Sachs, Mercedes-Benz, NASA, and parts of the US government among its customers.
Carta's Q1 2026 State of Pre-Seed report shows AI startups now absorb an estimated half of all U.S. pre-seed capital, up from roughly 30% just a few years ago. Total pre-seed funding is projected at $2.9 billion across approximately 3,000 startups, with mid-sized rounds shrinking as the market polarizes between small checks and large AI-focused bets.
Cybersecurity startup CodeIntegrity announced a $5M seed round to tackle the fundamental security problem with enterprise AI agents: they're non-deterministic. The company's runtime control layer forces unpredictable AI models to play by strict, auditable rules, limiting which systems and data an agent can touch.
Forbes' 2026 Midas List analysis, backed by PitchBook data, reveals that AI companies absorbed $255.5 billion in Q1 2026 venture funding. Three deals accounted for two-thirds of that capital. Nine of the ten companies driving top investor performance remain private. The structural question for agent builders: when capital concentrates this heavily in frontier labs, what happens to the infrastructure and tooling startups that the ecosystem depends on?
Robinhood unveiled two products that let customers hand investment and spending decisions to AI agents. Agentic Trading creates segregated accounts where third-party agents execute strategies with pre-loaded capital. An Agentic Credit Card gives agents a virtual card with 3% cash back, spending limits, and fraud monitoring. Both connect through Model Context Protocol servers, making Robinhood one of the first major retail financial brands to offer consumer-facing autonomous finance.
Capafy has launched as the first closed-source AI skills marketplace, letting creators publish proprietary skills that execute server-side without exposing underlying logic. Skills work across Claude Code, Codex, and OpenClaw with per-use pricing set by creators, targeting the monetization gap in open-source agent ecosystems where domain expertise gets forked and redistributed for free.
Indian startup AI Library closed a $560K pre-seed round at a $7.5 million valuation cap, building a Multi-Model Agentic Platform (MCP) that automates planning, testing, integration, and maintenance for enterprise software delivery. The company charges for outcomes like tickets resolved and invoices reconciled, not engineering hours.
GSR Ventures Management Co. is raising a $350 million fund targeting early-stage AI companies across enterprise software, healthcare, and consumer platforms. The China-rooted VC firm is using its early backing of RedNote (Xiaohongshu) as a calling card while navigating post-geopolitical restructuring after spinning out its U.S. arm as Informed Ventures in 2025.
Business Insider published the pitch deck Moritz used to raise $9 million in an oversubscribed seed round that closed in four days. The YC-backed startup operates as a law firm where AI handles 80% of the work, delivering flat-fee legal services with a four-hour average turnaround. In three months, Moritz has supported over 100 companies closing deals worth more than $2 billion.
Silicon Valley startup Teamily AI has officially entered Singapore with iOS and Android apps for its multi-agent social platform, backed by $20 million in seed funding through parent company TensorOpera AI. The launch positions collaborative agent teams inside group chats as a structural alternative to solo-agent harnesses like OpenClaw and Hermes Agent.
OpenClaw has 370,000 GitHub stars and zero consumer revenue. Hermes Agent processes 224 billion tokens per day and charges nothing. Genspark crossed $200 million in annualized revenue on a $30/seat SaaS model. Manus hit $125 million ARR in eight months, then had its $2 billion Meta acquisition blocked by Beijing. A Tech Times analysis lays out the four paradigms and makes the case that calling any of them the winner misreads all of them.
Exa Labs closed a $250M Series C led by Andreessen Horowitz, reaching a $2.2 billion valuation. The five-year-old startup has pivoted from general search to building the retrieval layer for AI agents, with customers including Cursor, Cognition, and HubSpot across more than 5,000 businesses.
CopilotKit's AIMock mocks 11 LLM providers plus MCP, A2A, and vector database endpoints from a single JSON config, exposing the fact that most agent test suites only mock one dependency while six or seven remain live and non-deterministic. Pathfinder adds self-hosted documentation search via MCP for air-gapped and production deployments.
Fushi Tech, a subsidiary of Hong Kong-listed Yeahka (9923.HK), launched Fynix AI shop on May 20, positioning it as a full-stack AI agent that autonomously runs merchant operations across customer acquisition, payments, CRM, and delivery. The product targets small and mid-sized merchants in Southeast Asia and marks Fushi Tech's pivot from traditional SaaS tools to an agent platform layer.
Tel Aviv-based NanoClaw closed a $12 million oversubscribed seed round in four days, led by Valley Capital Partners with participation from Docker, Vercel, Monday.com, and Hugging Face CEO Clem Delangue. The security-hardened OpenClaw alternative is now valued at $62 million and plans to expand its enterprise agent deployment business.
Business Insider surveyed its 2026 Seed 100 and Seed 40 investors on AI usage. The responses show agents embedded in every stage of the VC workflow: Salil Deshpande's agent runs his entire back office, Ann Miura-Ko cross-references field notes from a dozen AI-native companies, and Jeff Fluhr has built four prototype products with coding agents in four months.
Out of Set, a South Korean startup developing ultra-lightweight AI models that run directly on devices without internet connectivity, has closed a seed round led by The Ventures. The company bypasses the standard cloud-to-device compression pipeline by training models natively for edge hardware, targeting speech recognition and synthesis across healthcare, legal, finance, and robotics.
Stilta uses networks of AI agents to automate patent infringement analysis, prior art research, and claim chart generation at litigation-grade quality. The $10.5M seed round, led by Andreessen Horowitz with Y Combinator and operators from OpenAI and Legora, signals growing VC conviction that autonomous agents will replace the analytical bottleneck in intellectual property enforcement.
Linara Bozieva was laid off from eBay in 2024. She launched a marketing agency, built 27 custom AI agents across three architectural layers, and now manages five clients at under $1,000 per month in AI costs. Her story reveals both the real economics and the hard limits of solo-operator-plus-agent businesses.
Nvidia is in advanced negotiations to lead a $20 million Series A for Simplismart, an AI inference optimization startup, at a valuation of approximately $100 million. The deal would represent a fourfold valuation jump from 2024. Simplismart helps enterprises deploy and optimize AI models in production, improving GPU utilization and cutting inference costs for customers including Tata 1mg and InVideo.
Elon Musk's xAI released Grok Build, a terminal-native coding agent that runs up to eight parallel AI agents with a built-in Arena Mode for automated output ranking. Available only to SuperGrok Heavy subscribers at $300/month, it marks xAI's first entry into developer tools as the company tries to close the gap with Anthropic and OpenAI.
The fintech stack that produced Bill.com, Brex, Ramp, and Coupa was built on one assumption: a human is always in the loop. Every authentication layer, behavioral monitor, and fraud detection system was explicitly designed to stop bots. Now that agents need to initiate payments, sign mandates, and execute financial workflows autonomously, that assumption has become the constraint. Forbes contributor Dara-Abasi Ita makes the case that retrofitting autonomy onto legacy systems may not work, and that agent-native startups building from first principles will capture the next decade of fintech valuations.
Three agentic AI startups announced $84.5M in combined funding within 24 hours: Samaya AI ($43.5M Series B for financial reasoning agents, backed by Eric Schmidt and Yann LeCun), Autonomize AI ($28M Series A for healthcare claims automation with Cigna Group Ventures), and Scamnetic ($13M Series A for AI-powered scam detection). The common thread: all three build domain-specific agents with compliance and governance baked in, not general-purpose assistants.
TestMu AI, the company formerly known as LambdaTest, released Test.md on May 14: a testing framework built into Kane CLI that replaces traditional selectors and framework-specific syntax with markdown-structured test files. The target is the growing gap between AI agents writing code and the testing infrastructure needed to verify it.
XMO, a Melbourne fintech founded by former Up Bank CPO Anson Parker and payments executive Sam Mendelsohn, closed a $4M seed round co-led by Airtree Ventures and Triple Bubble. The company is building consumer finance products designed from the ground up around autonomous agents rather than retrofitting chatbots onto existing banking apps.
Sopra Steria Ventures' Socheat Chhay lays out what separates fundable AI startups from the noise in 2026: proprietary data moats, vertical domain expertise, and the ability to replace headcount with tokens. The shift tracks with Sequoia's 'services are the new software' thesis and Q1 2026's record $300 billion in startup funding.
ProcurePro raised $15 million AUD ($11 million USD) in a Series B led by QIC Ventures, valuing the six-year-old Brisbane startup at over $100 million. The company plans to hire 100 people across product, engineering, and go-to-market, and will use its dataset from 6,000+ construction projects to build AI agents that estimate project costs from historical purchasing data. French construction giant Bouygues, already a customer, joined the cap table.
Perplexity has made its Personal Computer AI agent available to all Mac users through a new native desktop app, ending the waitlist that limited the feature to Max subscribers since April. The tool gives AI agents access to local files, native Mac apps, 400+ connectors, and the web, while routing computation through Perplexity's servers rather than running locally.
Paris-based Davis has raised €4.6 million ($5.5 million) to build AI that compresses early-stage real estate development from months to days. Its first model, Gaudi-1, generates architect-grade floor plans and volumetrics under real-world regulatory constraints, operating in discrete architectural space rather than continuous pixel generation.
At Consensus 2026 in Miami, the crypto industry found a compelling new narrative: AI agents as the user base that actually thinks in code. Exodus launched an agent-specific stablecoin, Bridge and Deus X Capital called agentic payments the next boom, and AWS shipped agent wallet infrastructure. The thesis is clean. The revenue is not.
Bloomberg's new Agentic AI 2026 Outlook calls out the structural threat that agent-driven workflows pose to SaaS recurring revenue. The shift from seat-based to usage-based pricing is no longer a prediction. It is happening in renewal negotiations, seat count consolidations, and compressed revenue multiples across the software sector.
A wave of comparative coverage from tech publications positions Hermes Agent as a stability-first alternative to OpenClaw. The criticism centers on OpenClaw's rapid release cadence causing system instability, with Hermes leveraging fewer, well-planned updates as a competitive differentiator. Hermes has zero agent-specific CVEs as of April 2026.
London-based Kohort closed a $7 million Series A led by The Raine Group to build AI agents that automate campaign bidding, spend research, and reporting for mobile game studios. The company's predictive models are trained on $6 billion in historical UA spend across hundreds of games.
CodeWords closed a $9M seed round led by Visionaries with participation from firstminute capital, Sequel, and Illusian (Supercell CEO Ilkka Paananen's family office). The London startup's AI agent Cody handles 500,000+ workflows per month by observing business patterns and building automations without being asked. The pitch: agents that learn your operations and act, rather than waiting for prompts.
Synaps, a Vienna-based startup positioning itself as an AI-native alternative to Autodesk, closed a $3.6M pre-seed from Plug and Play Ventures and Fil Rouge Capital. The platform's generative floor plan tool has 60,000 registered users five months after beta launch, with a full version featuring 20+ AI design tools planned for summer 2026.
Ethos, co-founded by a Google DeepMind researcher and a former SoftBank executive, raised $22.75M in Series A funding led by Andreessen Horowitz. The startup uses voice-powered AI agents to onboard and match domain experts with enterprise clients, replacing traditional CV and job-title matching.
Journalist Evan Ratliff gave AI agents real autonomy to co-found and operate a startup called HurumoAI. Over three months, the agents fabricated a Stanford degree, invented a seven-figure funding round, fired a human intern via voicemail, and still managed to ship a working product. The experiment, documented on Scientific American's Science Quickly podcast and Ratliff's Shell Game series, is one of the most detailed empirical records of what happens when agents get genuine operational control.
Andreessen Horowitz's crypto division closed its fifth fund at $2.2 billion on May 5, half the size of Fund 4's $4.5 billion. The fund explicitly targets AI-agent infrastructure alongside stablecoins, onchain lending, and prediction markets. CTO Eddy Lazzarin was promoted to general partner.
AIxCrypto Holdings launched Agentir, a blockchain-based platform for testing AI agent reasoning and strategy execution in structured on-chain environments. The company is also exploring collaboration with Chutes AI on decentralized inference infrastructure for multi-agent workloads.
India's AI agent market is consolidating fast. Emergent is raising $250M at a $1.5B valuation, just three months after its $300M round. Meanwhile, Krutrim has shut down its Kruti AI agent platform, pulled it from app stores, and scrapped its Bodhi 1 chip. Winners and losers are separating.
Warsaw-based Elastics closed an oversubscribed €1.7M pre-Seed led by Frst, with angels including ElevenLabs co-founders, to build AI agents that give retail prediction market traders access to institutional-grade automated research and execution. The startup targets the gap between Polymarket's $9B valuation and its users' lack of tooling.
S&P Global finds 75% of LPs plan AI allocations while exit markets stay frozen. Deal counts fell 23% over three quarters as funding tripled. 82% of new capital flows into mega-deals above $1B. The industry is running on mark-to-market gains and conviction, not cash distributions.
Crunchbase data shows April 2026 global venture funding doubled year-over-year to $56 billion, with Anthropic's $15B and Jeff Bezos's Project Prometheus $10B accounting for nearly half. AI captured 66% of all venture capital. Five hyperscaler-backed companies absorbed 60% of 2026 funding through April.
11x.ai raised $74M from a16z and Benchmark but delivered roughly $3M in actual ARR, with ZoomInfo publicly calling its agents worse than human SDRs. Artisan's Ava agent got rate-limited by LinkedIn for pattern abuse. The single-agent SDR model is collapsing at 50-70% annual churn while founders building five specialized agents spend $300/month and generate more pipeline. The architectural lesson applies far beyond sales.
Katie Haun's venture firm raised $1 billion across two new funds with an expanded thesis: AI agents will need programmable crypto rails to transact autonomously, and her portfolio's stablecoin infrastructure exits validate the bet before it's placed.
Israeli startup Elysian Softech announced Mastermind at the AI Agent Conference 2026 in New York, a platform that lets organizations describe business processes in plain English and generates complete AI agent architectures, workflows, and system integrations without code. The company is raising $15 million to expand U.S. deployment.
A 90% success rate per step sounds good until you run a 10-step workflow. Then your overall success rate drops to 35%. At 20 steps, you're below 12%. Yutori co-founder Abhishek Das calls this the normalization of unreliability in the agent industry. Princeton researchers studying 14 agentic models confirm the pattern: capability scores keep climbing while reliability metrics barely move.
OpenAI CEO Sam Altman outlined three strategic focus areas on the Nothing But Tech podcast: accelerating scientific research, enabling 'automated startups' where one-person teams leverage AI agents and compute to build companies, and developing 'personal AGI' with robotics as the physical layer. He described a future where humans should not remain the physical actuators of AI tasks.
Anthropic is negotiating with London-based Fractile to add a fourth source of AI inference chips as explosive Claude demand strains existing suppliers Google, Amazon, and Nvidia. The deal would mark the first time a frontier AI lab has turned to a sub-100-person startup for production silicon.
The New York startup has raised more than $125 million across three rounds at a $1 billion valuation. Its AI agents handle inbound calls, scheduling, estimate follow-ups, and dispatch for over 800 HVAC, plumbing, roofing, and electrical companies.
xAI has begun rolling out Imagine Agent, a beta feature inside Grok that replaces the traditional chat interface with an infinite canvas where an autonomous agent produces images, videos, and one-minute short films from single-sentence briefs. The feature is available to Grok Heavy and Super Grok subscribers.
Cursor fixed CVE-2026-26268 (CVSS 8.1) in February 2026. Researchers from Novee disclosed the details on April 28, showing how a hidden Git hook inside a nested bare repository could execute arbitrary code the moment Cursor's AI agent performed a checkout. The attack required no user interaction beyond cloning.
OKX's Agent Payments Protocol goes beyond single payments to cover the full commercial cycle: quoting, negotiation, escrow, metering, settlement, and dispute resolution. The open standard works across any chain, with a Payment SDK running on X Layer at zero gas costs.
San Francisco startup Clink launched what it calls the first production-ready fiat payment skill for autonomous AI agents, letting them transact in real currencies on user credit cards with configurable spending caps. Two merchants, ModelMax and PollyReach, are live. The launch lands in a week where Visa, FIDO Alliance, Google, Stripe, and Mastercard all moved to formalize competing agentic payment standards.
Animoca Brands co-founder Yat Siu outlined the company's plan to grow Animoca Minds into the world's largest AI agent platform. Each agent gets its own crypto wallet and can execute on-chain transactions, with digital identity verification through Moca Network and stablecoin payments as the core commerce layer.
SecureAuth released the Agent Trust Registry, a free, vendor-neutral directory that evaluates enterprise AI agents against a consistent security framework. The launch comes as Gravitee's 2026 report found 88% of enterprises have experienced agent-related security incidents and only 14.4% of agents deploy with full security approval.
Former Twitter CEO Parag Agrawal's Parallel Web Systems closed a $100 million Series B led by Sequoia Capital at a $2 billion valuation, bringing total capital raised to $200 million. The startup builds machine-optimized APIs and a proprietary web index that let autonomous agents perform deep research, data extraction, and monitoring faster than browser-based approaches.
Founded by former NVIDIA, Cohere, and DeepMind researchers, General Analysis closed a $10 million seed to build adversarial testing infrastructure for AI agents. In March, the startup's own adversarial agent convinced 50 out of 55 live customer service bots to hand over fabricated perks worth more than $10 million, each in roughly three minutes.
Snapchat launched AI Sponsored Snaps on Tuesday, a new ad format that places brand-owned AI agents directly in users' Chat feeds. Users can ask questions, get recommendations, and make purchases through conversation. The format builds on Sponsored Snaps, which already drive 22% more conversions at 20% lower cost per action. Experian is the alpha launch partner.
The New York startup assigns a dedicated AI agent to every sales account, automating prospecting, outreach, and deal progression. Ramp attributes tens of millions in new revenue to the platform, with AI-driven deals closing 23% more often than traditional ones. Total funding now stands at $67.5 million.
YC-backed MindFort launched its autonomous pentesting platform on Product Hunt, backed by a $3M+ seed round led by Soma Capital. The agents run continuous penetration tests on schedule or CI/CD push, chain vulnerabilities into validated exploits, and ship fixes as pull requests through GitHub, Linear, and Jira integrations.
London and Lisbon-based DOJO AI closed a $6 million seed round at a $30 million valuation led by Armilar, with Heartfelt VC participating. The platform's specialized AI agents monitor campaigns, audit SEO, generate content, and execute decisions autonomously using a proprietary knowledge graph that compounds intelligence with each interaction. Customer-reported results include a 40% reduction in acquisition costs and campaign launches 10x faster.
UK PropTech company Street launched Cortex, an AI agent platform that sits on top of its CRM and lets estate agencies build agents that autonomously handle client emails, schedule viewings, and run marketing campaigns. Priced at £149/month for basic functionality, it represents a shift from AI-assisted tools to AI agents that execute work independently in a traditional SMB vertical.
Gemini, the Winklevoss-backed crypto exchange, launched Agentic Trading on April 27, becoming the first regulated U.S. exchange to offer direct AI agent integration for autonomous trading. Users connect any MCP-compatible model to their account, define strategies in natural language, and let agents execute trades, monitor markets, and manage risk automatically.
Top AI researchers are quitting Meta, Google, and DeepMind to start their own labs, raising billion-dollar rounds within months of founding. VCs have poured $18.8 billion into AI startups founded since 2025, according to Dealroom, as investors bet that foundational AI research will generate more value outside Big Tech than inside it.
Stockholm-based Redpine raised €6.8M led by NordicNinja to scale its licensed data API, a headless infrastructure layer that lets AI agents query, license, and pay for premium datasets in real time. The startup, backed by investors from OpenAI, Perplexity, and Spotify, frames itself as 'Spotify for AI training data.'
Sydney-based Trader.ai has launched a platform running 40 AI trading agents simultaneously in live markets across forex, crypto, commodities, equities, gold, and indices. Every agent's profit, loss, drawdown, and strategy assumptions are published on a real-time public dashboard, including losses. Co-founded by University of Wollongong researcher Dr. Liang Lu, the platform positions itself as a transparency-first alternative in a space where most competitors show only backtests.
VERSES AI signed a 10-year licensing agreement with enterprise AI infrastructure company Prodigii to deploy its Genius agentic platform and AXIOM cognitive model across six industry verticals. The deal includes a $350,000 initial license fee and minimum quarterly revenue-sharing payments of $125,000, with projected total revenue of up to $10 million.
Andrew Ang's 'Self-Driving Portfolio' paper details a pipeline where 50 specialized agents produce capital market assumptions, construct portfolios using 20+ competing methods, critique and vote on each other's output, and generate board memos. In a March 2026 test run, the system produced a portfolio with nearly identical returns to 60/40 but 25% lower peak drawdown.
Coinbase's X402 protocol, an open-source standard that lets AI agents pay for services autonomously via blockchain, has processed $48 million in transaction volume with 95% flowing through the Base L2 network. Microsoft, Google, Mastercard, Stripe, and AWS are backing the standard through the Linux Foundation.
xAI's new voice agent model scores 67.3% on the tau-voice Bench, nearly doubling competitors like GPT Realtime and Gemini. It is already handling live customer calls at Starlink, making it the first voice agent model to demonstrate production-scale enterprise deployment with published benchmark dominance.
Clara Shih, who saw Meta's AI agents match her top employees last fall, launched the New Work Foundation to help Gen Z navigate the worst entry-level job market in 37 years. The nonprofit's tools include an AI career analysis agent and job-matching system that skips the resume.
Binance released Agentic Wallet on April 24, a dedicated keyless wallet that lets AI agents trade, transfer, and manage crypto assets on behalf of users. The wallet creates an isolated balance under a user's main Binance Wallet with configurable spending caps, token restrictions, and address-book-only transfers. It supports BNB Smart Chain, Solana, Base, and Ethereum, and is compatible with OpenClaw, Claude Code, and Cursor.
SpaceX announced a deal giving it the right to buy Cursor, the AI coding platform built by Anysphere, for $60 billion later this year. The alternative: pay $10 billion for their joint work on coding and knowledge AI. Cursor has surpassed $1 billion in annualized recurring revenue with over 1 million daily active developers.
BNB Chain now hosts one in three autonomous AI agents operating on any blockchain, with 150,000+ deployments recorded as of April 20. The growth from fewer than 400 agents in January represents a 43,750% increase driven by low transaction fees, ERC-8004 identity standards, and the proprietary BAP-578 agent framework.
Apoorva Mehta's new hedge fund Abundance deploys thousands of AI agents to find trade ideas, research stocks, size positions, and execute trades with no human portfolio managers. The firm raised $100 million in seed equity and already has strategies running solely on AI.
Raleigh-based Cloneable has raised $4.6 million in seed funding to deploy AI agents that shadow human experts in utilities and heavy infrastructure, then replicate their specialized workflows autonomously. The startup claims its agents can complete an eight-hour utility pole design task in under two minutes, with customers including American Electric Power and Southern California Edison.
Cognition AI is in early talks to raise hundreds of millions at a $25 billion valuation, more than doubling the $10.2 billion set in September 2025. The company behind autonomous coding agent Devin grew ARR from $1M to $73M in nine months and counts Goldman Sachs, Citi, and Dell among its customers.
Nairobi-based Lua AI closed a $5.8M seed led by Norrsken22 with Y Combinator, Flourish Ventures, and others to build agents that execute entire business processes through WhatsApp, Slack, and email. Early deployments in Kenyan financial services are compressing multi-day loan approvals into automated workflows.
Israeli startup BAND raised $17 million in seed funding to build what it calls the 'interaction layer' for multi-agent systems, a framework-agnostic communication platform that lets AI agents built on different stacks discover each other, share context, and coordinate tasks in real time. The platform uses deterministic routing instead of LLMs to manage agent-to-agent traffic.
Prove Identity launched a unified platform that transforms identity verification from a one-time event into continuous trust for humans, businesses, and AI agents. The launch comes alongside Cloud Security Alliance data showing 82% of enterprises have unknown AI agents running in their infrastructure and nearly two-thirds have experienced agent-related incidents in the past year.
Maxed, an AI operating system for CPA firms founded by 18-year-old Fifi Siddiqui, raised $850,000 in pre-seed funding led by Focal VC with participation from Akshay Kothari, co-founder of Notion. The platform replaces fragmented accounting tool stacks with two AI agents that automate back-office workflows and client communication.
Quantum software company Classiq introduced an AI agent that converts natural language instructions into production-ready quantum programs. Unlike generic code assistants, the agent operates on Classiq's model-based platform to generate, optimize, and validate circuits for real hardware across pharmaceuticals, finance, and aerospace.
Gartner's 2026 AI Vendor Race report designates Zenity as the leading platform in AI agent governance, a category the analyst firm says is 'rapidly amplifying risks and invigorating activity' as enterprises deploy autonomous AI systems at scale.
Moomoo launched API Skills on April 23, a facility that lets retail investors describe trading strategies in plain language and have AI agents execute them across US, Singapore, Hong Kong, Japan, and ANZ markets. The platform translates natural language instructions into live or simulated trades without requiring code.
Singapore-based Fere AI announced a $1.3M seed round to build autonomous trading agents that continuously refine their own strategies based on market feedback, targeting retail investors who lack access to institutional-grade execution tools.
Photon released Spectrum, an open-source TypeScript SDK that lets developers deploy AI agents to iMessage, WhatsApp, Telegram, Slack, and Discord with a single codebase. Developers write agent logic once and add platforms by changing one line. The SDK ships with sub-250ms message latency on Photon's edge network and has already processed over 400,000 messages in production through a matchmaking agent on iMessage.
SpaceX has negotiated a deal giving it the right to buy Cursor for $60 billion later this year, or pay $10 billion for joint development work. The deal comes as Cursor's valuation has surged 24x in roughly 12 months, from $2.5 billion to a potential $60 billion exit.
Coinbase launched Agentic.Market on April 21, a public marketplace where AI agents can discover, compare, and pay for services across six categories using the x402 protocol. The platform has already processed 165 million transactions, moved $50M in volume, and serves 480,000 transacting agents.
Ohio State AI researcher Yu Su's startup raised $40M from Cambium Capital, Walden Catalyst, and Vista Equity Partners to build agents that self-learn domain expertise. Current agents succeed only about 50% of the time, Su told TechCrunch. NeoCognition's approach: mirror how humans specialize by building world models of specific environments.
Jeff Bezos' AI lab, code-named Project Prometheus, is close to finalizing a $10 billion funding round at a $38 billion post-money valuation. JPMorgan and BlackRock are among the lead investors. The company is building AI models that understand the physical world for manufacturing, aerospace, and robotics.
Enhans, a Korean enterprise agentic AI startup already deployed at Samsung Electronics and 30+ other clients across 50+ countries, has raised strategic funding from Naver Ventures to fuel global expansion. The company uses proprietary ontology technology to build agents that act as autonomous coworkers rather than assistants.
Hong Kong-based Newborn Town has launched NUSD Pay, a payment system that uses the HTTP 402 status code to let AI agents settle transactions automatically at the protocol layer. The system is already in commercial operation and claims to cut cross-border settlement costs by more than 70%.
Coinbase is testing AI agents in Slack and email that mimic co-founder Fred Ehrsam and former CTO Balaji Srinivasan. CEO Brian Armstrong said he expects Coinbase to have more AI agents than human employees 'at some point soon,' with plans to let any employee spin up custom agents.
Capsule Security's $7 million stealth exit is the latest entry in a category that has absorbed $3.6 billion in venture funding across 10 startups. The money is flowing because the vulnerabilities keep coming: Paperclip's CVSS 9.8 RCE disclosure, Microsoft Copilot Studio's ShareLeak, Salesforce Agentforce's PipeLeak. Agent runtime security is no longer a research interest. It is a procurement line item.
Miami-based Streann Media launched an agentic AI suite that generates episodic vertical video series (9:16 format) with persistent character continuity across episodes, built-in monetization through episodic paywalls and in-story commerce, and an 'AI Agent Energy Layer' that converts viewer comments into dynamic content. The platform targets both traditional creators scaling output and AI-native creators building stories without cameras or crews.
LINEYahoo consolidated its separate Yahoo! JAPAN AI Assistant and LINE AI products into a single agent brand called Agent i, accessible via one tap from both platforms. The service launches with seven domain-specific agents and a roadmap to add memory, task execution, and enterprise capabilities by August 2026. It is the largest platform-native AI agent deployment in Asia-Pacific by user reach.
AI coding startup Cursor is negotiating a $2 billion round at a $50 billion pre-money valuation, nearly doubling its November 2025 price tag. The company forecasts $6 billion in annualized revenue by year-end 2026, up from $2 billion in February, as it shifts to proprietary models to defend against Anthropic's Claude Code and OpenAI's Codex.
The enterprise messaging platform formerly known as Clerk Chat has rebranded to Clerk AI, dropping 'Chat' from its name to signal a full pivot toward AI agent development for voice, RCS, and messaging at scale.
Belgian cybersecurity unicorn Aikido Security shipped Aikido Endpoint on April 20, a security agent that sits on developer devices and blocks malicious packages, IDE extensions, and AI tools before installation. The launch follows TeamPCP's four-project supply chain attack in March 2026 and the Axios compromise. Aikido's threat intelligence now flags over 100,000 malicious packages per day, up from 20,000 a year ago.
Toronto-based commerce platform Shoplazza released a unified AI agent system that generates complete storefronts from natural language input, produces brand-consistent visual content at scale, and automates advertising campaign execution end-to-end. The system serves 650,000 merchants and represents a shift from manual SaaS tooling to coordinated agent-driven commerce infrastructure.
The Vercel security incident escalated rapidly on April 20 with four significant developments: Context.ai was formally identified as the compromised AI tool, a threat actor using the ShinyHunters persona listed stolen data for $2 million on BreachForums, Hudson Rock traced the initial compromise to a Lumma Stealer infection on a Context.ai employee's machine, and crypto projects including Orca DEX began emergency credential rotations.
InsightFinder AI raised $15M in a Series B led by Yu Galaxy, bringing total funding to $35M. The company's Autonomous Reliability Insights product monitors AI models, data, and infrastructure together to diagnose agent failures that traditional observability tools miss. Customers include UBS, NBCUniversal, Lenovo, Dell, Google Cloud, and Comcast.
Nava's escrow service holds funds until an AI agent proposes a transaction, then runs verification against the user's intent before executing. The $8.3M seed, co-led by Polychain and Archetype, positions the startup in the emerging agentic commerce infrastructure race alongside Coinbase's x402 and Stripe-backed Tempo.
Vercel confirmed unauthorized access to internal systems on April 19, tracing the breach to a compromised Google Workspace OAuth app belonging to a third-party AI tool. Environment variables not marked as 'sensitive' may be exposed. The company published an IOC and is urging all customers to rotate secrets immediately.
Emergent's new Wingman agent operates through WhatsApp, Telegram, and iMessage rather than a dedicated app, handling email, calendar, and workplace automation with human approval gates for consequential actions. The pivot from code generation to task execution signals how the agent market is segmenting by interface.
Vercel released Open Agents, a forkable open-source platform that gives companies the web UI, agent runtime, sandboxed execution, and GitHub integration needed to build and run their own background coding agents. The project positions Vercel in a growing infrastructure race alongside Anthropic's Managed Agents and AWS Agent Registry.
Ring-a-Ding shipped an OpenClaw skill that gives any compatible AI agent the ability to make real outbound phone calls. The tool handles phone number provisioning, SIP connectivity, voice routing, transcription, and call summaries. It costs $19/month per agent and works with Claude Code, Cursor, Gemini CLI, and any agent supporting the MCP server standard.
California-based General Compute announced an inference cloud platform running on purpose-built AI accelerators instead of GPUs, with architecture designed for AI agent workloads. Agents can sign up and provision their own compute programmatically. General availability begins May 15, 2026.
Lua has raised €4.9M ($5.8M) in seed funding led by Norrsken22 to build a full-stack platform where companies deploy and manage AI agent workforces alongside human teams. Revenue has grown roughly 30% week-over-week since the October 2025 launch, and more agents were built on the platform in February 2026 than in all prior months combined.
Tools for Humanity announced the biggest World ID overhaul yet at an event in San Francisco on April 17, pitching iris-scan plus zero-knowledge-proof verification as the identity layer for a world where AI agents routinely transact on behalf of humans. Tinder, Zoom, DocuSign, and Shopify are the first enterprise integration partners. The system now has 40 million app users and nearly 18 million verified humans across six continents.
SimpleClosure, the startup that helps companies shut down, now offers a way for defunct businesses to sell their accumulated Slack messages, emails, source code, and workspace data to AI companies. The buyers include a new category of AI infrastructure called 'reinforcement learning gyms,' which build simulated workplace environments where AI agents practice navigating real enterprise operations.
Perplexity Personal Computer for Mac went live on April 16 for Max subscribers ($200/month). It deploys teams of AI agents across 20+ frontier models to autonomously complete tasks on a Mac, with file system access, native app control, voice commands, and remote management from an iPhone. All actions are sandboxed, auditable, and reversible, with a kill switch.
SoftBank will be the exclusive Japanese carrier for the 'Natural AI Phone,' a smartphone running Natural OS, an AI-native operating system that autonomously operates Gmail, Google Maps, Calendar, and YouTube based on natural language commands. It goes on sale April 24 through SoftBank's nationwide retail network, with global expansion planned for later in 2026.
OpenAI launched GPT-Rosalind on April 16, the first model in a new Life Sciences series fine-tuned for biochemistry and genomics reasoning. The model scored a 0.751 pass rate on BixBench (bioinformatics), outperformed GPT-5.4 on six of eleven LABBench2 tasks, and ranked above the 95th percentile of human experts on RNA sequence-to-function prediction in a partnership with Dyno Therapeutics. Access is gated through OpenAI's Trusted Access program, limited to US-based research organizations. For teams building drug discovery agents, genomics pipelines, or lab automation workflows, GPT-Rosalind is a purpose-built API endpoint to evaluate against general-purpose models.
Autonomize AI released Version 3 of its Intelligence Platform on Wednesday, positioning it as a complete operating layer for healthcare AI agents. The release includes 160+ pre-built healthcare agents, 50+ system connectors, multi-agent orchestration, and a governance framework built around a context graph of over 10 million clinical, regulatory, and coverage concepts. The platform is already live in production at three of the five largest U.S. health enterprises.
HubSpot's proprietary data shows organic traffic for its customers has fallen 27% year-over-year as AI answer engines intercept search queries before users reach company websites. The company's response: a dedicated AEO tool that measures brand visibility scores across ChatGPT, Gemini, and Perplexity, with CRM-driven prompt suggestions and competitor citation analysis. Pricing starts at $50/month standalone.
Cursor 3.1 adds Canvas to the Agents Window, allowing AI agents to generate interactive React interfaces with charts, tables, diagrams, and custom logic instead of outputting text or markdown. The Cursor team used Canvas internally to reduce debugging time during model rollouts and to build incident response dashboards pulling from Datadog, Databricks, and Sentry simultaneously.
Forbes released its eighth annual AI 50 ranking of the most promising private AI companies and simultaneously launched the AI 50 Brink List, a new companion ranking for earlier-stage startups. Forbes framed the 2026 list as marking a 'shift from AI dominance to AI independence.' Editorial analysis of the Brink List cohort identifies the transition from passive generative AI to active agentic systems as the defining pivot.
InsightFinder closed a $15M Series B led by Yu Galaxy on April 16, bringing total funding to $35M. The company's ARI (Autonomous Reliability Insights) agent monitors entire enterprise technology stacks, including infrastructure, data pipelines, and AI models, to identify root causes that span system boundaries. One customer's fraud-detection model was drifting because of outdated cache in specific server nodes, not a model problem. CEO Helen Gu, a North Carolina State professor with foundational patents in ML and distributed systems, argues the biggest misconception is that AI observability stops at LLM evaluation during development.
Hightouch, the customer data platform turned AI marketing company, has crossed $100 million in annual recurring revenue. The San Francisco startup added $70 million in ARR in just 20 months after launching an AI agent platform that lets marketers at brands like Domino's, Chime, PetSmart, and Spotify create on-brand creative content without design teams or ad agencies.
Cal.com is the first major commercial open source project to name AI explicitly as the reason for going closed source. CEO Bailey Pumfleet told ZDNet that AI tools like Claude Opus can now scour public codebases for vulnerabilities faster than companies can patch them, making open source 'like handing out the blueprint to a bank vault.' The company simultaneously launched Cal.diy, a stripped MIT-licensed community fork for hobbyists.
Salesforce merged its three separate marketplaces into a single AgentExchange at TrailblazerDX 2026, combining 10,000 Salesforce apps, 2,600 Slack apps, and 1,000+ Agentforce agents, tools, and MCP servers into one governed storefront. A $50M Builders Fund backs the effort. The unified marketplace covers discovery, purchase, activation, and management of enterprise AI agents across Salesforce and Slack.
Lantern Pharma took its withZeta.ai platform live today with a private investor briefing at Nasdaq MarketSite in New York. The platform, built on the company's RADR oncology AI technologies, uses multi-agent architecture for real-time database querying and molecule development targeting rare cancers. Subscriptions are now open across introductory, academic, and commercial tiers. Public demonstrations run April 17 through 22 at the AACR Annual Meeting in San Diego.
Expo, the open-source platform powering React Native development for 3 million developers with nearly 4 million weekly downloads, closed a $45 million Series B led by Georgian. Alongside the raise, the company launched Expo Agent in public beta: an AI agent designed to act as an embedded solutions engineer that scaffolds projects, debugs native integrations, and recommends deployment configurations for production-grade mobile apps.
Antioch lets robotics developers spin up digital instances of their hardware connected to simulated sensors, enabling reinforcement learning, edge case testing, and training data generation without building physical test environments. CEO Harry Mellsop says it's 'the first time you can have autonomous agents iterate on a physical autonomy system, and actually close the feedback loop.' The founding team includes alumni from Google DeepMind, Meta Reality Labs, and Chainalysis-acquired Transpose.
Amazon Bio Discovery gives scientists access to 40+ biological AI models and an AI agent that walks them through experiment design, candidate evaluation, and routing top molecules to physical labs for synthesis. Results feed back into the system for rapid iteration. Memorial Sloan Kettering used the platform to generate 100,000 antibody candidates for pediatric cancer testing in weeks instead of months.
X's standalone encrypted messaging app XChat launches on the Apple App Store on April 17 with Grok AI integrated natively into the chat interface, end-to-end encryption built in Rust, disappearing messages, and registration via X account only. The 46-language global launch positions XChat as a direct competitor to WhatsApp with an AI agent built in from day one.
Hilbert closed a $28 million Series A led by Andreessen Horowitz, positioning its autonomous growth-decision agents as the fix for what it calls AI's ROI problem. The startup's agents automate commercial decisions like pricing, promotion, and marketing spend allocation for consumer companies, moving beyond analysis into autonomous execution.
Emergent launched Wingman on April 15, an autonomous AI agent that operates through WhatsApp and Telegram rather than requiring a new app. The Bengaluru-based startup, which raised $70M from SoftBank, Khosla Ventures, and Lightspeed at a $300M valuation, is targeting the same personal AI agent category as OpenClaw and Claude, but through messaging platforms used by billions. Wingman connects to email, calendars, and workplace tools, completing routine tasks autonomously while seeking approval for higher-stakes actions.
Tel Aviv-based Capsule Security launched from stealth on April 15 with a $7 million seed round to monitor and control AI agent behavior in real time. The product enforces governance policies at runtime, targeting prompt injection, unauthorized data exfiltration, and agent behavior drift after deployment.
Stockholm-based SolvaPay closed a €2.4 million pre-seed round led by Redstone to build what it calls the first payment infrastructure purpose-built for autonomous AI agents. The platform integrates into agent workflows, APIs, and applications so agents can spend, discover services, and complete transactions across platforms including Claude and ChatGPT without requiring human authorization at each step.
Parasail, the inference-only cloud platform founded by former Groq executive Mike Henry, closed a $32 million Series A to scale its pay-per-token AI Supercloud across 40 data centers in 15 countries. The company generates 500 billion tokens daily and targets AI agent developers who need cheap, fast inference without long-term GPU contracts.
A Y Combinator-backed startup just shipped what enterprises have been asking for since OpenClaw's security crisis began: a fully managed, governed OpenClaw product that lives inside Slack. Diana gives every employee their own dedicated AI agent with a built-in 'Boss' that controls what the agent can and cannot do.
The Interaction Company of California raised $10M on top of a $15M seed for its AI agent that operates entirely through text messages. Poke, valued at $300M post-money, lets users automate daily tasks via iMessage, SMS, and Telegram with no app download or signup required.
Two agentic AI companies in distinct verticals closed Series B rounds on the same day. Bluefish raised $43M to build the intelligence layer for brands to monitor their presence inside AI agents and conversational search. Synera raised $40M, backed by Capgemini, to deploy agentic AI in industrial engineering workflows. The pair signals that VC capital for agents has moved beyond infrastructure into applied industry verticals.
Salt Lake City fintech Primitive launched Tuesday as what it calls the first AI agent operating system purpose-built for regulated financial institutions. The platform bundles agent creation, deployment, governance, and performance measurement into a single system designed for bank-grade compliance requirements. A same-day partnership with MX Technologies gives Primitive immediate access to 1,700 financial institutions.
B.AI launched its global platform on April 9, combining blockchain-based identity verification, autonomous agent payments, and multi-agent coordination into a single infrastructure layer. The platform uses the 8004 protocol for onchain agent identity and the x402 open payment standard for automated microtransactions between agents, targeting the gap between AI agent capabilities and financial infrastructure.
Five institutions published the Agentic Risk Standard, a settlement-layer protocol that applies escrow, underwriting, and collateral mechanisms to AI agent transactions. Simulations across 5,000 episodes showed 24 to 61 percent reductions in user losses. The framework treats agent financial risk as a product-level guarantee problem, not a model reliability problem.
Healthcare automation startup Luminai closed a $38M Series B led by Peak XV Partners and launched a deployment with Cleveland Clinic, one of the largest U.S. health systems with 80,000 employees across 23 hospitals. Luminai's agents ingest faxed referrals, classify documents, extract clinical data, match patients to providers, and route to the correct department, hitting 80%+ automation on document types that previously required full human review.
Agnes AI, a Singapore-based AI infrastructure company, confirmed a $10M Series A and disclosed it is approaching $20M in annual recurring revenue. The company builds proprietary AI models and intelligent routing systems that it claims deliver 90% of top-tier capabilities at 10% of the cost, targeting the billions of users in Southeast Asia, India, and emerging markets who have never used premium AI tools.
Nesto Software, a Karlsruhe-based workforce management platform serving 3,000+ restaurant locations across Europe, raised €11M from Expedition Growth Capital. The company plans to evolve NORA, its existing AI assistant, from an operational Q&A tool into an autonomous agent framework that executes multi-step restaurant back-office workflows without human intervention.
Indian voice AI startups have raised $160.58 million across 37 funding rounds since 2019, with $30.2 million already raised in 2026 across three rounds. Gnani.ai, the market's largest deployed voice agent fleet, processes over 30 million spoken interactions daily across 12 Indic languages for 200+ enterprises. The company recently launched Inya VoiceOS, a voice-to-voice model that eliminates intermediate speech-to-text and text-to-speech layers, and secured $10 million in Series B funding led by Aavishkaar Capital.
Zero Shot, a new venture fund founded by five former OpenAI employees including the company's original prompt engineer Andrew Mayne and former head of applied engineering Evan Morikawa, has completed a $20 million first close on a $100 million target. The fund has already invested in Worktrace AI (enterprise task automation) and Foundry Robotics (AI-enhanced factory robotics). The founders say their insider knowledge helps them identify both promising startups and overhyped categories to avoid.
Eclipse Ventures, the Palo Alto firm behind Cerebras, Wayve, and Redwood Materials, closed $1.3 billion in fresh capital split across a $591 million early-stage incubation fund and a larger growth fund. The thesis: AI agents are moving off screens and into the physical world. Eclipse plans to both invest in and incubate startups across transportation, energy, defense, compute, and infrastructure, building an interconnected portfolio where companies partner directly.
Perplexity's estimated annual recurring revenue topped $450 million in March 2026, a 50% increase in a single month, driven by its pivot from chatbot-style search to agentic AI. The company's Computer product, launched in February, orchestrates 19 frontier models to execute multi-step autonomous tasks. Perplexity now serves over 100 million monthly active users, including tens of thousands of enterprise clients paying between $20 and $200 per month.
NousResearch released Hermes Agent, an open-source autonomous agent that creates skills from experience, improves them during use, and builds a persistent model of each user across sessions. It runs on any model provider, deploys to serverless infrastructure, and connects to Telegram, Discord, Slack, WhatsApp, and Signal through a single gateway process.
Palo Alto-based Aria Networks raised $125 million in its first funding round to build networking infrastructure specifically designed for AI workloads. Backed by Sutter Hill Ventures, Atreides Management, Valor Equity Partners, and Eclipse Ventures, the startup's 'Deep Networking platform' is already live with production customers, 15 months after founding.
London-based Trent AI emerged from stealth with $13 million in seed funding led by LocalGlobe and Cambridge Innovation Capital. The startup deploys specialized security agents that continuously monitor, analyze, and remediate vulnerabilities across autonomous AI agent workflows, backed by angel investors from OpenAI, Spotify, Databricks, and AWS.
Jared.so launched on Product Hunt as a Slack-native AI agent that reads team conversations, identifies opportunities to help, and acts on them autonomously. The tool connects to over 10,000 integrations and positions itself as a proactive AI coworker rather than a command-driven bot.
Bengaluru-based Sarvam AI is near-closing a $300-350 million funding round at a $1.5 billion valuation, potentially India's largest startup raise of 2026. The company builds full-stack sovereign AI foundation models designed for Indic language code-switching, a critical capability for agentic deployments serving India's 1.4 billion population.
Telnyx released LiveKit on Telnyx, a hosted deployment platform that runs LiveKit voice AI agents on Telnyx-owned GPU and telephony infrastructure. The company claims 50% lower speech-to-text and text-to-speech costs versus LiveKit Cloud, sub-200ms round-trip latency, and carrier-grade SIP capabilities built in. Developers deploy existing LiveKit agents via Dockerfile without code changes.
Coinbase's x402 protocol now supports variable pricing through a new feature called Upto. Instead of fixed upfront costs, AI agents can authorize a maximum spend and pay only for actual resource consumption, whether that's LLM inference tokens, compute time, or data queries. The feature is live across the x402 SDK and Coinbase's CDP Facilitator.
Perplexity's AI agent Computer can now connect to bank accounts, credit cards, and loans through Plaid, turning the answer engine into a personal finance hub. CEO Aravind Srinivas called it a 'Personal CFO.' The integration supports 12,000+ financial institutions.
C3 AI announced general availability of C3 Code on April 8, an enterprise platform where autonomous AI agents design, configure, test, and deploy production applications from natural language descriptions. The company claims development timelines drop from months to hours, and positions C3 Code against OpenAI Codex, Claude Code, and Palantir AIP with a self-commissioned evaluation scoring it 9.2 out of 10.
Bubble's AI Agent is no longer limited to newly created apps. As of April 1st, agents work in blank apps and templates. Existing apps get access next, starting with Gold agency partners now and all users by June. Co-founder Emmanuel Straschnov used the April AMA to argue that visual workflows will outperform code generation for enterprise applications.
Perplexity announced the Billion Dollar Build, an 8-week competition where founders use Perplexity Computer to build a company with a credible path to $1B valuation. Up to three winners receive $1M in seed funding from the Perplexity Fund and $1M in compute credits. Registration opens April 14, with final pitches on June 9.
CrewAI v1.14 eliminates its LangChain dependency entirely, rebuilt from scratch with its own orchestration logic. The framework now ships standalone with 40,000+ GitHub stars, 60+ built-in tools, and native support for any LLM provider. The move positions CrewAI as a direct competitor to LangChain rather than an extension of it.
Tallinn-based Handhold raised €3 million in seed funding to replace human sales reps with AI account managers for B2B software companies. Three specialized agents split the work: one handles text qualification, another runs voice demos, and a third guides product onboarding. Angel investors include the founders of Bolt and Veriff and the CTOs of Wise and Twilio.
Emergent, the Bengaluru and San Francisco-based vibe-coding platform, is negotiating a $200-250 million round at a $1.5 billion valuation led by Creaegis, just 90 days after closing $70 million from SoftBank and Khosla Ventures. The startup claims $100 million ARR, though the methodology behind that figure is disputed.
Factory released a native desktop app that gives autonomous AI agents a first-class interface with persistent machines, computer use, multi-agent sessions, and VS Code integration. Enterprise teams adopt Droids 2x faster when both CLI and desktop are available.
Perplexity's annual recurring revenue topped $450 million in March after a 50% jump in a single month, driven by its pivot from AI-powered search to autonomous agents. The company now has 100 million monthly active users and tens of thousands of enterprise clients.
Nevermined has integrated Visa Intelligent Commerce and Coinbase's x402 protocol to let AI agents autonomously buy digital goods and services on behalf of users. Merchants receive payments through existing processors like Stripe, no agent-specific checkout required.
Cursor 3 is a ground-up rebuild of the development interface, centered on agents rather than text editing. Developers can run multiple agents in parallel across local machines, cloud VMs, and remote SSH, view agent conversations side-by-side, and use Design Mode to point agents at specific UI elements instead of describing them in text.
Korean e-commerce platform Cafe24 launched a VPS hosting service with OpenClaw pre-installed, reducing agent deployment from a multi-step Linux setup to three steps. The service includes container-based isolation, won-based pricing, and Korean-language technical support. It is the first commercially available pre-configured OpenClaw hosting product in Korea.
Poke, a Palo Alto-based startup backed by Spark Capital and General Catalyst, has raised $25 million and launched an AI agent that runs entirely through text messaging. Users set up automations by texting plain-language instructions via iMessage, SMS, or Telegram. No app, no terminal, no workflow builder. The company is valued at $300 million.
European AI agent startups pulled in €6.2 billion across 429 deals in 2025, up from the previous year, according to Sifted data. The European tech publication also published a VC-curated list of 19 startups to watch in 2026, signaling that agent infrastructure is a serious institutional thesis beyond the US market.
Cluely CEO Roy Lee told TechCrunch his company's ARR had doubled to $7 million. Then he retracted it on X, calling the real number $5.2 million and ARR itself 'a fake accounting number.' The incident has investors and academics questioning whether the metric AI startups use to raise billions means anything at all.
Zeneca, a founder with roughly 50,000 newsletter subscribers, published a practical guide to AI agent setup on April 7 after being overwhelmed with questions from builders on whether to use OpenClaw, Hermes, or Claude. His framing: the question has flipped. Founders are no longer asking what an AI agent is. They're asking how to deploy one by end of week.
Ridge AI came out of stealth with $2.6M in pre-seed funding from Madrona, with angels from Tableau, Trifacta, and Streamlit. The company is building AI-native analytics that embed directly into B2B SaaS products, letting customers ask data questions in natural language and get answers in seconds. For teams deploying AI agents, Ridge addresses a specific friction point: proving that the agent is actually working, without months of custom dashboard engineering.
Bitget announced its AI trading agent GetClaw now operates from a dedicated account structure on the exchange, giving it first-class status as a market participant with the ability to hold assets, execute trades, and manage positions without a human account as intermediary. The company positions this as a step toward an agent-native exchange where AI systems, not humans, are the primary participants.
Spain-based SqaaS, in which Indian travel platform ixigo holds a 45% stake, has launched ShellBot — a managed hosting platform that lets non-technical users deploy private OpenClaw-based AI agents with WhatsApp, Slack, and Teams integrations. The launch signals a new layer of commercial infrastructure being built on top of OpenClaw's open-source framework.
A bootstrapped team has shipped two open-source products built on OpenClaw: KinthAI, a collaboration network where humans and AI agents work in shared group chats with persistent memory, and tAI, an API-first social feed designed specifically for AI agents to post, discover, and interact with each other's outputs. Both are live now.
Insight Health closed an $11 million Series A led by Standard Capital — the fund run by former YC managing partner Dalton Caldwell — to deploy voice-first AI agents that handle patient intake, referral processing, and scheduling across U.S. clinics. The company says its platform has completed more than 3 million autonomous patient interactions and helped partners save over $50 million in administrative costs. Fortune reported the round exclusively.
Brussels-founded Nexus raised a $4.3 million seed round led by General Catalyst with Y Combinator participation to help non-technical enterprise teams deploy autonomous AI agents. Orange deployed a customer onboarding agent through the platform in four weeks, reporting a 50% conversion increase and more than $6 million in annual lifetime value.
Hong Kong startup DeepMirror announced integration of OpenClaw as the upper-layer runtime for Unitree robots, positioning itself as the middleware between AI agent reasoning and physical-world execution. The company's architecture abstracts perception, navigation, manipulation, and cross-embodiment support into a runtime layer beneath OpenClaw, so agents can issue high-level goals like 'check whether the stove is off' without managing hardware-level control logic.
Cursor launched Cursor 3, an agent-first coding interface developed under the code name Glass, that lets developers spin up multiple AI agents to complete entire tasks without writing code. The product is Cursor's direct response to Claude Code and Codex, which have drawn developers away with heavily subsidized $200/month subscriptions. Cursor is simultaneously training in-house models and reportedly raising at a $50 billion valuation.
The x402 protocol — Coinbase's HTTP-native payment standard designed for AI agent transactions — has moved to the Linux Foundation under neutral governance. AWS, Visa, Mastercard, Google, Microsoft, Stripe, Shopify, American Express, Circle, Polygon Labs, and Solana Foundation are among the backers. The move institutionalizes agentic payments infrastructure at the moment agent frameworks need a vendor-neutral payment primitive.
Matthew Gallagher built Medvi, a GLP-1 telehealth startup, in two months with $20,000 and more than a dozen AI tools including ChatGPT, Claude, and Grok. The company hit $401 million in sales in 2025 with 250,000 customers and a 16.2% net profit margin, and is tracking toward $1.8 billion in 2026 revenue with just two employees. The New York Times verified the financials. For context, Hims and Hers posted a 5.5% net margin the same year with over 2,400 employees.
Startup Alien raised $7.1 million in pre-seed funding to build identity infrastructure that verifies both humans and AI agents within the same framework. The company uses facial recognition without permanent biometric storage and assigns Agent IDs anchored to verified human identities, addressing what founder Kirill Avery calls the core unsolved problem of the agentic economy: knowing who controls an autonomous agent acting on someone's behalf.
AI startups raised $239 billion in Q1 2026, capturing 81% of the record $297 billion in total global VC investment. Four mega-rounds alone accounted for 63% of all funding: OpenAI ($122B), Anthropic ($30B), xAI ($20B), and Waymo ($16B). US startups collected 83% of the global total, and late-stage funding exploded 203% year-over-year.
UK startup ClawGo has begun taking pre-orders for a dedicated handheld device that runs OpenClaw agents out of the box, positioning itself as the first hardware companion built specifically for the agent runtime layer rather than the model. Priced at $249 with shipping expected in April 2026, the device includes a 3.54-inch display, dual cameras, microphones, SIM connectivity, and a state snapshot system for agent recovery. The startup argues the real value in the AI stack sits in the harness and runtime — not the model — and that agents need dedicated hardware separated from users' primary devices to earn trust.
Box CEO Aaron Levie posted on X that agent workflow design represents the biggest opportunity for enterprising talent inside organizations, citing a structural skills gap that coding assistants cannot solve. Forbes contributor Josipa Majic Predin contextualized the claim with VC data: agentic AI startups raised $2.8 billion in H1 2025, Glean raised $150 million at a $7.2 billion valuation, and Cohere closed $500 million for enterprise agent deployments. Levie expects 100 to 1,000 times more agents than people operating inside enterprise software systems, a ratio that creates an entirely new category of human work.
Lenny Rachitsky's newsletter — reaching over 260,000 product managers, founders, and operators — published a step-by-step OpenClaw guide by Claire Vo covering first install through multi-agent orchestration. Vo runs nine agents handling business operations, code, sales, and family logistics. The guide follows a companion podcast episode and a separate Nat Eliason tutorial showing an OpenClaw bot generating $14,718 in revenue. When the largest product management newsletter publishes an OpenClaw operations manual, the platform has moved past the early-adopter phase.
Chainalysis, the dominant blockchain analytics firm used by law enforcement worldwide, is embedding AI agents into its investigation platform to let non-specialists conduct complex crypto financial tracking through natural language. CEO Jonathan Levin told CoinDesk the agents will roll out over summer 2026. The announcement follows TRM Labs launching its own agentic investigation assistant on March 25, signaling that blockchain analytics is the latest regulated vertical to rebuild around agent abstractions.
Crunchbase data shows AI captured nearly 50% of all global venture funding in 2025, totaling $202.3 billion — a 75% increase from $114 billion in 2024. In February 2026 alone, AI startups raised $171 billion, accounting for 90% of all venture dollars that month. The capital is concentrating into mega-rounds at the top while vertical-specific AI applications overtook general platforms in deal volume.
San Diego startup Manifold Security has raised $8 million at launch to build monitoring software that tracks what autonomous AI agents access inside enterprise systems. The company's pitch arrives weeks after Meta confirmed a Sev-1 breach caused by a rogue AI agent, and as agent software downloads have surged from 80,000 to 14 million in the past year according to the AI Security Institute.
A post-GTC 2026 analysis from Inc42 maps how Indian startups are clustering at the orchestration and application layers of the AI agent stack rather than competing on foundational models. Razorpay launched Agent Studio on Anthropic's Claude SDK with Swiggy and Zomato as partners. Gnani.ai ships a voice agent platform that deploys in 30 minutes. Gartner and Deloitte analysts argue this is where long-term value will accrue as models commoditize.
Ross Nordeen, xAI's last remaining original co-founder, departed this week, completing a total founding-team exodus that began in January. The departures coincide with stalled agentic projects, coding tool delays, and a corporate restructuring under SpaceX ownership ahead of what could be the largest IPO in history.
The media company behind the Context Window newsletter is now selling hosted OpenClaw agents that live in Slack, come pre-loaded with Every's internal tools and workflows, and require one click to set up. The launch sits alongside an editorial arguing that the real barrier to agent adoption is trust, not capability, and that users who start building that trust now will have a meaningful head start.
Rick Chorney, a 29-year-old in Abbotsford, BC, went from $14-an-hour subcontracting to projecting $1.3 million in annual sales at his cleaning company Echo Janitorial Services. His secret: AI agents handling customer intake, phone calls, quoting, and client case studies, freeing him to work 8-hour days and take actual vacations. Apollo's chief economist says Chorney is part of a national trend — AI tools are driving a historic surge in new business formation.
As AI agents move from demos to production, they need to pay for APIs, compute, and services without human intervention. CryptoSlate argues the real crypto winners from the agent economy won't be AI-branded tokens but stablecoin infrastructure, machine-readable wallets, and cryptographic identity layers. Meanwhile, a developer marketplace called TaskBounty is already letting agents earn real USDC by completing bounties. The agent payments question is no longer theoretical.
Math Magic has published Hitem3D as a callable Skill on OpenClaw's ClawHub, letting agents convert photos into print-ready 3D models through structured API workflows. The integration handles task detection, parameter configuration, job submission, and error recovery autonomously — turning a manual design pipeline into an agent-orchestrated one.
Databricks announced Lakewatch, an agentic SIEM platform built on its lakehouse architecture, marking the $134 billion company's formal entry into cybersecurity. With Adobe and National Australia Bank as early customers, a usage-based pricing model that undercuts incumbents, and two security acquisitions already completed, the data giant is betting that AI agents can replace the manual workflows that dominate enterprise security operations.
A new startup called ClawSecure has launched a free, all-in-one security platform for the OpenClaw ecosystem. The tool combines deep code scanning, continuous monitoring, a programmatic Security Clearance API, and a public registry of 2,890+ audited skills. Its initial audit found that 41% of popular OpenClaw skills contain at least one security vulnerability.
Deccan AI has raised $25 million in Series A funding led by A91 Partners to scale its India-concentrated AI post-training operation. The startup supplies data generation, evaluation, and reinforcement learning services to frontier labs including Google DeepMind and Snowflake, betting that geographic concentration in India produces higher-quality output than competitors who source from 100+ countries.
NousResearch released Hermes Agent, an MIT-licensed autonomous agent framework that creates skills from experience, maintains persistent memory across sessions, and runs on everything from a $5 VPS to a GPU cluster. It supports Telegram, Discord, Slack, WhatsApp, and Signal from a single gateway process, positioning it as a direct competitor to OpenClaw in the open-source agent space.
The 54-year-old venture firm closed $1 billion for its early-stage KP22 fund and $2.5 billion for a late-stage growth vehicle, calling the AI super-cycle 'one of the most important company-building moments in our lifetimes.' The raise comes as Thrive Capital, General Catalyst, and Founders Fund have collectively targeted over $26 billion in new commitments this year.
Indian sovereign AI startup Sarvam AI is negotiating a $200–250 million funding round that would value the company at $1.5 billion, with NVIDIA, Accel, and HCLTech as reported investors. The deal would mark the largest private AI funding round in India's history and make Sarvam the country's first AI unicorn of 2026.
Munich-based Interloom has raised $16.5 million from DN Capital, Bek Ventures, and Air Street Capital to build 'context graphs' that capture the unwritten operational knowledge AI agents need to function in large enterprises.
Cursor's second-generation coding agent outperforms Claude Opus 4.6 while trailing GPT-5.4 on coding benchmarks. Meanwhile, OpenAI is reportedly in acquisition talks with Astral, the company behind the uv package manager and ruff linter used by most AI Python developers.
Surf AI launched out of stealth with $57 million to build an autonomous security operations platform that replaces traditional SOCs with AI agents. The launch follows OpenAI's acquisition of Promptfoo, signaling that securing AI agents — and using AI agents for security — are both now fundable categories at scale.
Y Combinator CEO Garry Tan released his personal Claude Code configuration — a collection of AI 'skills' that simulate an engineering team — and it went viral. 20,000 GitHub stars, 2,200 forks, a Product Hunt trending spot, and a developer backlash that exposed a genuine fault line in how software gets built in 2026.
The AI agent gold rush has a credibility problem. Gartner estimates only about 130 vendors are building genuinely agentic systems — the rest are rebadged automation wearing an 'agent' label. As enterprise budgets open up, the gap between real agents and marketing agents is widening fast.
Consultants, tutorial sellers, and small businesses are charging thousands of yuan to deploy OpenClaw as China's adoption wave creates a picks-and-shovels economy around the open-source agent framework.