Key takeaways

  • Stripe says we're living in the singularity, and therefore, regrettably, can't IPO.
  • Token usage is growing 9 percent per week.
  • Attracting investors doesn't seem to be a problem.

What happened

Stripe says we're living in the singularity, and therefore, regrettably, can't IPO. " Stripe argues it's better positioned to navigate the singularity as a private company, able to fund acquisitions and investments without diluting shareholders. " Stripe also confirmed its acquisition of OpenRouter for more than $8 billion. OpenRouter has built the world's largest and most trusted token-routing engine, supporting all major AI models and providers.

Why it matters

Token usage is growing 9 percent per week. Going forward, every developer will need not just payment infrastructure but also a reliable way to manage their "intelligence pipeline," Stripe says. That's what OpenRouter delivers. A more skeptical read is that "the singularity" gives Stripe a handy reason to stay private. The company benefits from less regulation, no obligation to disclose financials, and full founder control.

Attracting investors doesn't seem to be a problem. The numbers speak for themselves. Its stock price has risen 31 percent per year since the Series D, beating both the S&P 500 (14 percent) and Nasdaq (18 percent). First-half revenue rose 41 percent year over year, and free cash flow climbed 43 percent, Axios reports. 88 percent of the Forbes AI 50, including OpenAI and Anthropic, use Stripe.

What to watch

Declaring the singularity is becoming fashionable. The technological singularity is the idea that AI reaches superhuman intelligence and triggers an uncontrollable, self-accelerating cycle of progress. " Most recently, Demis Hassabis, former head of Google's AI lab DeepMind, said humanity is at the foothills of the singularity. OpenAI CEO Sam Altman and Elon Musk have made similar claims.

Altman already went there last summer, writing about a "gentle singularity" that had started and couldn't be reversed. One way to back up these claims would be AI systems that actually start improving themselves at speed. There are serious doubts that this ability exists yet, though early signs suggest it isn't impossible.