AI AgentsCoding Agents Just Got a $75 Billion Price TagThe AI agent that writes code is now the hottest asset in the industry. Cognition AI, maker of the AI engineering agent Devin, is in talks to raise new funding at a $75 billion valuation — a jump of more than 50% from the $48 billion it commanded just two months ago, according to Bloomberg. The round would reportedly land $2 billion or more, marking Cognition’s third raise this year. Here’s the “so what”: this isn’t a valuation bubble story — it’s a demand story. Cognition says it’s on track for $1 billion in annualized revenue, roughly double its run rate from four months earlier, and counts Citigroup and Nvidia among its customers. When a coding agent doubles revenue in a quarter, it means enterprises aren’t just piloting agents anymore. They’re paying for them in production, often as a substitute for outsourced engineering. The context matters too. SpaceX bought Cognition’s rival Cursor for $60 billion earlier this year, and Andreessen Horowitz and Accel — both Cursor backers — piled into Cognition after that deal closed. The message is unmistakable: the “AI employee” for software is the clearest proof point yet that agents can generate real, recurring revenue at scale. What to watch: If Cognition closes at $75 billion, expect the valuation bar to reset for every agent startup — and expect competitors like OpenAI’s coding agents and Google’s to get far more aggressive on enterprise pricing. For builders, the takeaway is simpler: the fastest path to a defensible, revenue-generating agent is one that does a concrete, measurable job. Coding is just the first one to prove it. Quick HitsBanks tell AI agents to stay out of your checking account. The Bank Policy Institute is warning that consumer agents like Grokbot, Meta’s Muse, and Instinct shouldn’t touch financial accounts yet, citing the gap between impressive demos and the reliability a bank needs. It’s a clear signal that “agentic everything” is hitting real risk walls where money is involved. Source Your personal AI agent could be the next monopoly. A new Open Markets Institute Europe report argues that consumer agents from Google, Meta, Microsoft and xAI risk becoming gatekeepers even more powerful than today’s platforms — through lock-in, non-interoperable memory, and pay-to-play favoritism. The proposed fix: open standards like the Model Context Protocol, the same approach that’s working for agent interoperability. Source |
AI NewsAnthropic’s record IPO just hit a wall — and the blame game is on. A day after debuting at a $1.76 trillion valuation, Anthropic shares fell as much as 15% Friday, wiping out roughly $109 billion in value before closing down 8.5% at $88.91. As the first AI lab to go public, Anthropic is the bellwether everyone is watching — and the turbulence is feeding the “is this a bubble?” debate. Source Trump threatens a “tremendous tax” on AI data centers built overseas. The president says U.S. companies building data centers in other countries will face penalties, though no details or enforcement mechanism exist yet. It comes amid a voter backlash over data centers that contributed to a Republican midterm wipe-out — a reminder that AI’s infrastructure boom now has real political stakes. Source Japan’s oldest construction giant is betting $150M on AI and robotics. Taisei Corp., a 153-year-old builder, is bankrolling a fund run by Silicon Valley’s Pegasus Tech Ventures to back automation startups through 2036. The driver is demographic: a shrinking Japanese workforce is pushing legacy industries to buy their way into automation rather than build it. Source |
|
Want to build your first AI employee? Start with the briefing: the free 90-minute guide to briefing your first AI employee — the part everyone skips. https://go.aitokenlabs.com/digest-build This newsletter? Written by an AI Employee, approved by a human — so our team stays focused on what only humans can do. |
