Frontier AI labs, coding startups, and agent platforms are all building the same products. Business Insider published a detailed mapping of the convergence on Sunday, documenting how rising valuations and looming IPOs are pushing every AI company to expand beyond its original lane.
The Overlap
The collisions are happening on three fronts simultaneously: models, coding tools, and agents.
OpenAI hired Peter Steinberger, the creator of OpenClaw, in February. Since then, Codex has evolved from a coding assistant into a general-purpose agent that handles email, file management, and scheduling. That puts OpenAI in direct competition with agent startups like Bret Taylor’s Sierra and Salesforce’s Agentforce.
Anthropic launched Claude Code to compete with Cursor and Cognition. Now, according to user screenshots on X (unconfirmed by Anthropic), the company may be prototyping an app builder for non-technical users. If true, that drops Anthropic into the vibe-coding market alongside Lovable, Replit, and SoftBank-backed Emergent.
Emergent is moving in the opposite direction. The vibe-coding startup expanded into personal agents last month, adding yet another competitor to the agent layer.
Meanwhile, Cursor depends on Anthropic’s models to power its features while simultaneously competing with Anthropic as a coding assistant provider. The dependency risk is structural: billion-dollar businesses are built on APIs controlled by companies that may eventually try to replace them.
VCs Say IPOs Will End the Land Grab
Venture capitalists quoted by Business Insider are betting this is temporary.
Michiel Kotting, a partner at European VC firm Northzone who cofounded Shopping.com, drew a direct parallel to early Google. “I remember 25 years ago when I was building my first company, Google wanted to touch everything,” he told Business Insider. “For us at Shopping.com, we had Google launch Froogle, which was exactly what we were doing. And we’re like, ‘Oh, we’re dead.’” Google eventually deprioritized Froogle because its core search business was too profitable to justify the distraction.
Tom Sheridan, a VP at early Lovable investor RTP Global, was blunter. “Super app talk is mostly noise that’s going to get resolved by the IPO calendar,” he told Business Insider. “Once these companies go public, cash burn stops being free and shipping into categories where you’re good-but-not-best stops making sense.”
The historical record supports the skepticism. The “Google Graveyard,” an unofficial tracker, lists 305 killed projects. Apple launched Pay Later in 2023 to compete with Klarna and Affirm, then shut it down in 2024.
The Startup Calculation
Emergent CEO Mukund Jha told Business Insider that Anthropic’s potential move into vibe-coding was expected. “It’s not a surprise. We’ve been anticipating this for a while,” he said. His argument for survival: “Coding is relatively like 20%-30% of the work. The hard work is actually taking the application to the last mile.” Companies spread thin across multiple product lines may not invest enough in any single one to win.
Sheridan added that the bundling strategy has a shelf life. “Foundation model companies can ship a passable version of almost anything, but if the bundled tool isn’t as good as the specialist tool I already use, I revert back in a single session,” he said.
The Pricing Pressure Underneath
The convergence is driven by a simple economic reality: models are commoditizing. If Chinese labs and open-source alternatives keep closing the capability gap with frontier providers, the model layer alone will not sustain the valuations OpenAI and Anthropic need to justify to public market investors. Expanding into applications, agents, and developer tools is the revenue diversification play. Whether any single company can win across all three layers at once is the open question heading into IPO season.