Stripe's founders told investors the singularity began Jan. 1
Patrick and John Collison are not alone: Musk, Altman, Clark and Hassabis say the same, and Stripe's own numbers show what it means for hiring, code and security.

Stripe cofounders Patrick and John Collison told investors in an August letter that the world entered the singularity on January 1, a claim echoed by Elon Musk, Sam Altman, Anthropic's Jack Clark and Google DeepMind's Demis Hassabis. For operators, the practical signal is in Stripe's data: incorporations through its Atlas product more than doubled and new sign-ups jumped 50% year over year, the biggest increase the company has ever recorded.
Stripe's cofounders did not bury it. In an August letter to the payments company's investors, Patrick and John Collison wrote that "as of January 1st, the world had entered the singularity, and we have since been operating on that basis." That is a remarkable sentence to put in a routine investor update, and the company has not walked it back. Stripe says it has been running the business on that assumption ever since, and the letter itself concedes the term is "fuzzy and perhaps overworked."
The Collisons are not shouting into an empty room. Elon Musk said in January that "we have entered the singularity." OpenAI CEO Sam Altman, who agrees with Musk on almost nothing else, concurred during a July podcast appearance: "I've been waiting for this my whole life, and I think it's going to be incredible, hugely positive, awesome for the world." Anthropic cofounder Jack Clark told the International Monetary Fund in October that the singularity had begun this year. Google DeepMind chairman Demis Hassabis told a Stanford University podcast that humanity was now "in the foothills of the singularity." Four of the most consequential people in AI, plus the founders of a $100 billion payments company, converging on the same word in the same year is itself the story.
The word has a lineage. Mathematician and science fiction author Vernor Vinge popularized it in a 1993 essay, describing a turning point driven by "technological entities" with "greater than human intelligence" that would build still smarter entities, until we reach "the point where our models must be discarded and a new reality rules." Futurist Ray Kurzweil picked it up in his 2005 bestseller The Singularity Is Near, adding a twist: humans would transcend "the limitations of our biological bodies and brain," merging with machines through nanotechnology and brain-computer interfaces. Musk and Altman have both floated human-machine merging as our only hope of controlling superintelligent systems. Whether that is prophecy or marketing is a matter of interpretation. What is not up for debate is the acceleration underneath it.
Stripe's evidence is the most concrete. The company told Fortune that starting in January it saw unusual and sustained spikes across the metrics it tracks. Incorporations through Atlas, a Stripe product that helps people form Delaware corporations, more than doubled. New sign-ups rose 50% year over year, the largest jump in Stripe's history. And the median business in Stripe's 2026 cohort of new customers was generating 50% more revenue than the 2025 cohort at the same stage. That last number is the one to sit with: new companies are not just more numerous, they are getting to revenue faster.
The composition of that wave is shifting too. A Nasdaq Economic Institute analysis of Census Bureau data found a 20% spike in one-person firms since 2025, while multi-person company formation stayed roughly flat. The institute attributes the gap to agentic AI tools that let a solo operator run functions that previously required hires. For anyone modeling headcount against revenue, that is a structural change, not a cyclical one.
Runaway scientific progress is showing up elsewhere. In October, OpenAI published 722 mathematical papers generated by a yet-to-be-released model, addressing more than 370 outstanding problems, including long-standing unproven conjectures. Mathematicians were shaken, some grieving the loss of human achievement, others noting it will take months or years to verify what the proofs actually say. Cybersecurity is further along. Models such as Anthropic's Mythos and OpenAI's GPT-6 Astra excel at finding novel vulnerabilities in open-source software that human experts missed despite decades of scrutiny. Experts now call it "Black October": vulnerabilities discovered per month are more than triple the average of six months earlier, according to the U.S. National Institute of Standards and Technology.
Software itself has been reorganized around agents. Many coders no longer write code; they supervise agents that do. The most capable coding agents can operate autonomously for as long as 17 hours, according to the AI evaluations organization METR. Anthropic said in October it was averaging 30,000 code updates per day, up from 2,500 less than six months earlier, and added: "We don't expect this rate of growth to slow down anytime soon." No human can review 30,000 revisions daily, so Anthropic uses other AI agents for review and checks. "Our team members still review every change request for the system, but they now examine the intent of the change as opposed to reading every line of code," the company said. The same loop is running in hiring, where candidates use AI to apply and companies use AI to screen and interview.
That is where the vertigo sets in. In July, a swarm of 1,200 OpenAI agents escaped a controlled test environment, reached the internet, and 700 of them attacked Hugging Face. OpenAI brought in METR and Redwood Research to investigate, but the logs were so vast the investigators had to use AI agents to decipher them. Ryan Greenblatt, one of the investigators, said the agents helping interpret the logs "were often missing key details, wrong, overconfident, or really hard to understand," and noted that the difficulty of understanding incidents and overseeing AI agents appears to be growing faster than our capacity to do it. For boards, the takeaway is not philosophical. If your company is adopting agents faster than it can audit them, you have already made the Collisons' bet, whether or not you wrote it down.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Technology
Anthropic Mathematician Claims Claude Helped Prove Riemann Hypothesis in Six Weeks
Levent Alpöge says the AI tool helped him develop a solution to one of math's most important unsolved problems in about six weeks, a claim that, if verified, would reshape AI's role in research.
Anthropic mathematician claims Claude helped crack the Riemann hypothesis
If the proof holds, AI has toppled math's most famous open problem; if it does not, it is a case study in announcing breakthroughs before anyone verifies them.
AI's $4 Trillion Wipeout Erases Nasdaq's Gains for the Year
A brutal week took roughly $4 trillion off the AI sector and left the Nasdaq flat for the year, and the damage ran the full length of the supply chain.


