Stripe's $7B OpenRouter Deal and Anthropic's First Profitable Quarter Mark an AI Infrastructure Inflection

Stripe is acquiring OpenRouter for roughly $7 billion — a company valued at just $1.3 billion four months earlier, implying about 70x trailing revenue — in a move framed as Stripe absorbing model-routing complexity the way it historically absorbed payments complexity. The same 20VC x SaaStr roundup reports Anthropic posted its first quarterly profit on $11.5 billion in Q2 revenue, with gross margins improving to roughly 30% and an expectation of reaching 40% by year-end, driven by revenue growth outpacing expense scaling rather than new operational discipline. Elsewhere in the same wave of AI infrastructure consolidation: Silver Lake is circling Workday with a $43 billion bid priced at 5.3x revenue against 13% growth, structured as a leveraged buyout targeting roughly 20% IRR, and SpaceX is acquiring Cursor for $60 billion in an all-stock deal with a $10 billion breakup fee. The piece also cites an estimated $100,000 in annual AI token spend per engineer and a claim that agentic workflows are enabling 30-40% smaller teams.

The pattern of infrastructure vendors reaching profitability and consolidating through large acquisitions while AI usage drives their underlying volume isn’t unique to model providers. A separate SaaStr breakdown of Palo Alto Networks traces the security vendor’s run to a roughly $265 billion market cap — the stock doubling over twelve months — after spending about $29 billion on 20+ acquisitions to build a five-pillar platform, including a $25 billion purchase of CyberArk in February. CEO Nikesh Arora’s framing is that AI is raising cybersecurity’s terminal value rather than compressing it: AI agents generate far more east-west network traffic requiring inline inspection, which drove hardware’s best quarter in ten years even as the company’s software platforms scaled. Both stories describe the same dynamic from different infrastructure layers — vendors who sit underneath rising AI usage capturing outsized value and rolling that value into acquisitions rather than fielding new competitors.

For revenue and GTM leaders tracking where AI infrastructure spend actually concentrates, the throughline across payments routing, foundation models, and security platforms is the same: the picks-and-shovels layer beneath agentic AI usage is where the profit and the M&A activity are showing up first.