The latest news on OpenClaw, AI agents, and automation
Tag
65 articles
Meta is building a personalized AI assistant for everyday tasks. Google is developing a 24/7 Gemini-powered personal agent. Both moves come months after OpenClaw's viral adoption proved consumer demand for AI that acts, not just answers, and OpenAI's February hire of creator Peter Steinberger signaled the stakes.
Google employees are testing an AI agent codenamed Remy that transforms Gemini from a conversational assistant into an autonomous personal operator capable of taking actions across Gmail, Calendar, Drive, and Android. Internal documents describe it as a '24/7 personal agent for work, school, and daily life.'
Google shut down Project Mariner on May 4, ending a 17-month experiment in browser-based AI agents. The project's death signals a broader industry verdict: screenshot-scraping agents that click and scroll are losing to command-line tools that manipulate files and execute code directly. With OpenAI's Operator falling below 1 million weekly users and Perplexity's Comet stalling at 2.8 million, the entire browser agent category is being absorbed into larger platform plays.
Google employees are dogfooding a new autonomous agent called Remy that integrates across Google services and can handle multi-step tasks proactively. It runs inside a staff-only version of the Gemini app and represents Google's most direct response yet to OpenClaw's viral success.
Google used Cloud Next '26 to ship its full agent stack: the Gemini Enterprise Agent Platform for building governed autonomous agents, eighth-gen TPUs designed for agentic compute demands, the Gemma 4 open model, and a free AI Agents Vibe Coding course on Kaggle. Nearly 75% of Cloud customers already use Google Cloud AI.
Google Cloud unveiled the Gemini Enterprise Agent Platform at Cloud Next 2026, positioning autonomous agents as core business infrastructure rather than experimental features. CEO Thomas Kurian framed the shift as moving from 'systems of intelligence' to 'systems of action,' with Agent Identity, Agent Registry, and Agent Gateway providing enterprise governance for fleets of autonomous agents.
Rubrik Agent Cloud adds real-time governance, agent inventory autodiscovery, and an 'Agent Rewind' feature that can undo an autonomous agent's destructive action on Google Cloud's Gemini Enterprise Agent Platform. The system is powered by Rubrik's Semantic AI Governance Engine (SAGE), which replaces static oversight with intent-driven guardrails.
Three agent control planes launched in the same week: Guild.ai with a $44M Series A from Google Ventures, SS&C Blue Prism with WorkHQ for regulated industries, and Google formalizing its Gemini Enterprise Agent Platform with cryptographic agent identities. The convergence signals that the enterprise AI market has shifted from 'can we build agents?' to 'who governs them in production?' Each platform takes a different architectural bet on where the control layer sits, what it governs, and who it serves.
The Department of Defense has signed classified AI agreements with Google, OpenAI, and xAI in the two months since blacklisting Anthropic as a supply chain risk. Pentagon AI Chief Cameron Stanley confirmed the Google expansion to CNBC, while 600+ Google employees signed a letter urging CEO Sundar Pichai to reject the deal. The contracts allow AI use for 'any lawful government purpose' with adjustable safety filters, the exact language Anthropic refused.
Andhra Pradesh Chief Minister N. Chandrababu Naidu laid the foundation stone for Google's $15 billion AI data center in Visakhapatnam on April 28. The 1-gigawatt facility, spanning 600 acres, is Google's largest single FDI project in India and targets eventual expansion to 5 GW, more than triple India's total data center capacity as of late 2025.
Google's new Workspace Studio lets business users build and deploy AI agents across Gmail, Docs, Sheets, Drive, Meet, and Chat by describing what they want automated in plain language. The platform connects to third-party apps including Asana, Jira, Salesforce, and Mailchimp, and rolls out to Workspace business, enterprise, and education customers.
Sundar Pichai disclosed at Google Cloud Next 2026 that 75% of all new code at Google is now AI-generated and approved by engineers, up from 25% in October 2024 and 50% last fall. Google engineers are now orchestrating 'fully autonomous digital task forces' of AI agents. A complex code migration completed six times faster with agents than engineers alone managed a year ago.
Google Cloud demonstrated a live voice agent at Cloud Next 2026 that handles 100% of YouTube TV and NFL Sunday Ticket customer support calls. The agent switches languages mid-conversation without prompting, answers nuanced follow-up questions about streaming restrictions, and was built and deployed in six weeks using CX Agent Studio's visual orchestration builder.
Google's Gemini 3 claims the top spot on 12 of 18 benchmarks including Terminal-Bench 2.0 and SWE-Bench, introduces a four-tier model family spanning reasoning-heavy to latency-optimized inference, and launches Antigravity as its agent-first development platform.
Google Cloud announced the largest single partner investment from any hyperscaler: a $750 million fund to finance consulting firms and systems integrators building agentic AI on its platform. Accenture has built 450+ agents, KPMG committed $100 million of its own capital, PwC pledged $400 million, and NTT DATA dedicated 5,000 engineers to Google Cloud agent development.
Google Cloud has split its eighth-generation TPU into two specialized chips: TPU 8t for massive-scale training with 9,600-chip superpods, native FP4, and 10x faster storage access, and TPU 8i for low-latency inference powering collaborative agents. Both are designed for the computational patterns of agentic AI, including long context windows, sequential reasoning chains, and world model simulation.
CrowdStrike extended its Falcon security platform into Google's Agent Cloud ecosystem as a launch partner and unveiled Project QuiltWorks, a coalition combining frontier AI models with Falcon Spotlight to accelerate vulnerability discovery. CBO Daniel Bernard said one participant found 45 million vulnerabilities and predicted more patching in the next 12 months than in all of prior history.
Google used Cloud Next 2026 to rebrand Vertex AI as the Gemini Enterprise Agent Platform, absorbing Agentspace into a unified Gemini Enterprise product. The result is a single platform spanning agent development (ADK v1.0 in four languages), sub-second cold starts (Agent Runtime), persistent memory (Agent Memory Bank), cryptographic identity (Agent Identity), fleet management (Agent Gateway), anomaly detection, and a 200+ model garden including Anthropic Claude. The A2A protocol is now in production at 150 organizations.
Defense Department workers built more than 103,000 semi-autonomous AI agents and logged 1.1 million agent sessions on the Pentagon's GenAI.mil platform in roughly five weeks, a senior official disclosed at the Box Federal Summit. The pace of adoption reflects a workforce under pressure from staffing cuts and the ongoing Iran conflict.
Deloitte announced a dedicated practice for deploying agentic AI at enterprise scale using Google Cloud's Gemini Enterprise platform. The practice includes over 1,000 prebuilt industry-specific agents spanning retail, healthcare, financial services, and government.
ServiceNow and Google Cloud announced a strategic partnership to combine Gemini Enterprise with the ServiceNow AI Platform, creating a unified environment where AI agents autonomously detect, diagnose, and resolve operational issues across 5G, retail, and IT.
Google will invest up to $40 billion in Anthropic, starting with $10 billion in cash and adding up to $30 billion more tied to performance milestones. The deal expands on Anthropic's existing TPU infrastructure relationship with Google Cloud and follows a similar $25 billion agreement with Amazon weeks earlier. Anthropic's annualized revenue has topped $30 billion.
Rubrik launched Agent Cloud at Google Cloud Next 2026, a governance layer for autonomous AI agents built on Google's Gemini Enterprise Agent Platform. The integration adds auto-discovery of running agents, semantic governance that replaces static rules with continuous behavioral oversight, and a unique Agent Rewind feature that lets administrators undo an autonomous agent's destructive actions. Gartner predicts 40% of enterprise applications will integrate task-specific AI agents by end of 2026.
Google Cloud unveiled the Agentic Data Cloud at Cloud Next 2026, a three-pillar architecture that replaces the traditional data stack built for human analysts with infrastructure purpose-built for autonomous AI agents. The platform introduces a Knowledge Catalog that automates semantic metadata curation, a cross-cloud lakehouse that queries Iceberg tables on AWS S3 with no egress fees, and a Data Agent Kit that drops MCP tools into VS Code, Claude Code, and Gemini CLI. With Vodafone, American Express, and Virgin Voyages already running production agent workloads on the platform, Google is betting that whoever owns the data context layer for agents will control enterprise automation outcomes.
Google Cloud and CVC Capital Partners announced a multi-year partnership at Cloud Next to accelerate agentic AI adoption across CVC's portfolio companies. The deal gives portfolio companies streamlined access to Google's Gemini Enterprise Agent Platform, early access to AI products, forward-deployed engineering teams, and Mandiant/Wiz cybersecurity protections.
Virgin Voyages unveiled Rovey at Google Cloud Next, an AI crew assistant built on Gemini Enterprise Agent Platform and BigQuery. Rovey handles itinerary recommendations, shore excursion suggestions, cabin selection, and booking guidance. It is the first of seven planned AI releases under Virgin's Project Ruby initiative.
Google unveiled Workspace Studio at Cloud Next 2026, a no-code platform that lets business users build and deploy AI agents across Gmail, Docs, Sheets, Drive, Meet, and Chat using natural language. It connects to third-party apps like Asana, Jira, and Salesforce, and rolls out to all Workspace business, enterprise, and education customers.
Check Point will integrate its AI Defense Plane with Google Cloud's Gemini Enterprise Agent Platform to provide three-layer security for autonomous AI agents: full visibility into agent inventories, pre-deployment policy enforcement, and real-time runtime guardrails that detect prompt injection and block data leaks. Available late June 2026.
Google Cloud unveiled a $750 million fund at Cloud Next 2026 in Las Vegas to help its 120,000-member partner ecosystem build, deploy, and scale agentic AI solutions. The fund includes embedded engineering teams, early model access, and co-investment in partner-led agent projects.
Home Depot launched AI voice agents built on Google Cloud's Gemini Enterprise, with a 50-store pilot showing customer intent recognition in under 10 seconds and issue resolution 4x faster than traditional phone menus. The system handles order status, product availability, service requests, and can build shopping carts from project descriptions. Rollout to all 2,359 U.S. stores is planned within the year.
Anthropic has committed to deploying multiple gigawatts of next-generation Google TPU capacity starting 2027, its largest compute commitment to date. The deal with Google and Broadcom complements the $25 billion Amazon/AWS Trainium infrastructure agreement announced days earlier, giving Claude agents dedicated hardware across two separate cloud ecosystems.
HCL Technologies launched a dedicated Gemini Enterprise Business Unit on April 22, the first such unit from a global system integrator. The move formalizes agentic AI as a distinct P&L within a $13B+ IT services firm, coinciding with Google Cloud's $750 million partner investment fund that embeds forward-deployed engineers at major SIs.
Salesforce and Google Cloud announced an expanded partnership at Cloud Next 2026 that lets Agentforce agents execute workflows natively across Slack, Google Workspace, and Gemini Enterprise. Over 1,400 customers already use Gemini within Agentforce, and the integration adds zero-copy data access from Google Lakehouse, eliminating the need to move or duplicate enterprise data.
Google Cloud and Wiz launched Agentic Defense at Cloud Next 2026, deploying three autonomous security agents for threat hunting, detection engineering, and third-party context enrichment. Google's existing triage agent has processed over 5 million alerts in the past year, cutting 30-minute manual analyses to 60 seconds. Wiz expanded AI Application Protection to cover agent studios from Google, Microsoft, AWS, and Salesforce.
Google unveiled TPU 8t and TPU 8i at Cloud Next 2026, splitting its custom silicon into specialized training and inference architectures for the first time. TPU 8t scales to 9,600 chips per superpod with 2 petabytes of shared memory. TPU 8i triples on-chip SRAM to eliminate latency bottlenecks for multi-agent inference. Both connect through Virgo Network for near-linear scaling to one million chips.
Google Cloud formally replaced Vertex AI with the Gemini Enterprise Agent Platform at Cloud Next 2026, a full-stack system for building, deploying, governing, and optimizing autonomous agents at enterprise scale. The platform reported 40% quarter-over-quarter growth in paid monthly active users in Q1 2026, with production deployments at Bosch, NASA, PepsiCo, and Unilever.
Google DeepMind released two autonomous research agents on April 21: Deep Research for fast interactive queries and Deep Research Max for exhaustive, asynchronous analysis. Both run on Gemini 3.1 Pro, support MCP for proprietary data feeds, and generate native charts. FactSet, S&P Global, and PitchBook are already collaborating on MCP integrations.
Trust3 AI announced native integration with Google Cloud's Agent Development Kit and Vertex AI Agent Builder on April 21, adding automated policy enforcement, data lineage tracking, and compliance mapping for enterprises deploying Gemini-powered agents.
Google Chief Scientist Jeff Dean told Bloomberg the company is building next-gen TPU chips specialized for inference, targeting the latency requirements of AI agents. Meta, Anthropic, and Citadel Securities are already using or expanding TPU access. The announcement is expected at Google Cloud Next in Las Vegas this week.
Google DeepMind's Aletheia, built on Gemini 3 Deep Think, autonomously solved 6 of 10 never-before-published research-level math problems in the FirstProof challenge. Expert mathematicians judged the solutions publishable after minor revisions. When Aletheia could not solve a problem, it said so instead of hallucinating a plausible answer. OpenAI attempted the same challenge with human supervision and scored 5 out of 10.
A Johns Hopkins University research team demonstrated that prompt injections in pull request titles and issue comments can hijack Claude Code, Gemini CLI, and GitHub Copilot running in GitHub Actions to steal API keys and access tokens. All three vendors paid bug bounties but none issued CVEs or public advisories.
Google's Android team released a new CLI, skills repository, and knowledge base designed specifically for AI agent workflows. Internal experiments showed 70% token reduction and 3x speed improvement over standard toolsets. The tools work with any agent, not just Gemini.
Google Cloud's Gemini models and Vertex AI will be integrated directly into Avid Media Composer, the dominant professional video editing platform for film and television. The multi-year partnership, announced ahead of NAB Show in Las Vegas, targets automated media discovery, scene tagging, and production workflows. Avid CEO Wellford Dillard said customers 'are asking for intelligent tools that plug into existing workflows and scale with their creativity.'
Google Cloud customers can now run GitLab's AI agent platform using the Vertex AI models they already pay for, counting that usage toward existing GCP commitments. GitLab stock jumped 7% after-hours. The story isn't the product. It's the procurement shortcut.
Google released a native Gemini desktop app for macOS with a floating chat triggered by Option+Space and the ability to share any active window directly with Gemini for real-time analysis. It's not autonomous computer control. It's an ambient AI layer that sees your work when you invite it to.
Addy Osmani, a director of engineering at Google Cloud AI working on Gemini, published a framework for Agentic Engine Optimization (AEO) that defines how web content should be structured for AI agents rather than human readers. The framework covers discoverability, parsability, token efficiency, capability signaling, and access control. Research cited in the framework shows AI coding agents compress multi-page human browsing sessions into one or two HTTP requests, making traditional engagement analytics invisible.
Google DeepMind's new robotics reasoning model can read pressure gauges, sight glasses, and digital readouts without instrument-specific training. Boston Dynamics is already shipping it inside Spot's inspection platform for industrial facilities.
In a recent interview, Sundar Pichai laid out Google's clearest vision yet for search: information queries become tasks executed by autonomous agents running in parallel. 'Search would be an agent manager,' he said, describing a product that orchestrates rather than returns results.
Google Research published two multi-agent frameworks targeting academic publishing bottlenecks. PaperVizAgent orchestrates five specialized agents to generate publication-ready figures from manuscript text, scoring 60.2 on a calibrated evaluation scale where human performance sits at 50.0. ScholarPeer automates peer review using adversarial literature search and multi-aspect verification, producing reviews that are more critical and literature-grounded than existing automated approaches.
Google shipped ADK Python v1.30.0 overnight with Gemma 4 model support, Auth Provider integration for agent registries, artifact streaming in A2A protocol events, and a credential leakage fix. The most active open-source multi-agent framework in 2026 now lets agents pass live artifacts in agent-to-agent communications and manage credentials across distributed agent ecosystems.
Google Colab's new Learn Mode turns Gemini into a step-by-step coding tutor that understands full notebook context. Separately, Gemini CLI v0.33.0 shipped authenticated agent-to-agent communication, addressing a key gap in secure multi-agent orchestration.
Google has committed to using multiple generations of Intel's Xeon 6 CPUs in its AI data centers for training and inference workloads. The companies also reiterated their ongoing collaboration on custom Infrastructure Processing Units, a partnership dating to 2022. Intel CEO Lip-Bu Tan said scaling AI 'requires more than accelerators.'
Google's CEO told the Cheeky Pint podcast that Search will evolve from an answer engine into a task orchestrator, managing multiple AI agents running long-duration workflows. He compared the vision to Antigravity, Google's agent platform, and said information-seeking queries will become agentic.
Broadcom disclosed on April 6 that it has agreed to produce future versions of Google's AI chips and signed an expanded deal with Anthropic for access to approximately 3.5 gigawatts of TPU computing capacity starting in 2027. The deals accelerate both companies' moves toward infrastructure independence from Nvidia. For Anthropic specifically: the same week it ended flat-rate Claude access for third-party agent harnesses like OpenClaw, it committed to the hardware required to serve vastly more inference volume at lower per-token cost.
Three companies that compete fiercely on pricing and capability have quietly begun sharing threat intelligence through the Frontier Model Forum to detect Chinese adversarial distillation: the technique of using API query outputs to train cheaper imitation models. Bloomberg reported the cooperation on April 6. The models being protected are the same ones powering agent harnesses built on Claude, Gemini, and GPT — and an eroded model moat means an eroded foundation for any agent stack built on top.
Google DeepMind researchers have published the first systematic framework for how malicious web content manipulates autonomous AI agents. The paper identifies six 'trap' categories — from hidden HTML instructions that hijack agent perception to systemic attacks that could trigger synchronized crashes across thousands of trading bots. Content injection attacks partially hijacked agents in up to 86% of tested scenarios. A single crafted email caused Microsoft M365 Copilot to leak its full privileged context.
Google released ADK for Java 1.0.0 on March 30 and ADK for Go 1.0 on March 31, making it the only major agent framework with four language SDKs. A new Morph comparison of eight agent frameworks surfaces a widening split: provider-native SDKs from Anthropic, OpenAI, and Google trade model flexibility for deeper integration, while independent frameworks like LangGraph and CrewAI offer cross-provider portability at the cost of native features. Meanwhile, IBM's ACP has merged into Google's A2A under the Linux Foundation, consolidating the agent-to-agent protocol layer.
Three days after Anthropic banned OpenClaw users from Claude subscriptions and required them to pay extra for API access, Google launched Gemma 4 under Apache 2.0 — a capable open-weight model that runs locally via Ollama. LushBinary published a guide pairing the two this weekend. The combination gives builders a zero-cost, privacy-first alternative with no API bills and no data leaving their machine.
A new Google DeepMind paper titled 'AI Agent Traps' presents the first systematic taxonomy of adversarial attacks against autonomous AI agents. The researchers document six categories of traps — from hidden HTML instructions that commandeer agents 86% of the time, to multi-agent chain reactions that could trigger synchronized financial sell-offs. The paper arrives the same week a critical vulnerability was discovered in Claude Code and Permiso launched SandyClaw for agent skill sandboxing.
Google is preparing to finance a $5 billion-plus data center in Texas operated by Nexus Data Centers and leased to Anthropic. The 500-megawatt facility, already under construction on a 2,800-acre campus, represents one of the largest single-tenant AI compute buildouts to date and sets a new floor for the infrastructure investment required to run agent workloads at scale.
Google quietly added 'Google-Agent' to its official crawler documentation and has made WebMCP (Web Model Context Protocol) available for early preview, signaling a fundamental shift in how AI agents will interact with the web. Rather than agents navigating websites the way humans do—pixel-based, slow—WebMCP allows agents to access structured tools directly from HTML forms and JavaScript APIs. Search Engine Journal describes this as 'the biggest mindset shift in SEO history.' For e-commerce and SaaS, the implications are profound: the web is becoming agent-optimized, not human-optimized.
Google built an internal AI coding agent called 'Agent Smith' that became so popular employees couldn't stop using it—so the company had to restrict access. Sergey Brin publicly confirmed agents are a major Google focus for 2026, signaling the company is building a public equivalent.
In a January all-hands meeting, Google DeepMind leaders told employees the company would pursue more national security contracts, including cybersecurity and biosecurity. CEO Demis Hassabis said he was 'very comfortable' with the balance.
Google has reassigned staffers from Project Mariner, its Chrome browser AI agent, to higher-priority projects. The move comes as browser agent adoption stalls industrywide — Perplexity's Comet hit just 2.8 million weekly users, ChatGPT Agent fell below 1 million — while command-line agents like OpenClaw surge.
Google DeepMind appointed Jasjeet Sekhon, previously Chief Scientist at the world's largest hedge fund, as Chief Strategy Officer. The hire signals a hard pivot toward commercial deployment and real-world agent systems.