The race: labs, money, and adoption
In plain termsWhen ChatGPT launched in late 2022, it showed the world what AI could already do, and companies responded by pouring enormous amounts of money into building better models. That race explains why AI tools change so fast — features you learn this semester may look different next semester. The skill that lasts is knowing how to learn and evaluate a new tool.
ChatGPT sparked a trillion-dollar race
When ChatGPT launched in November 2022, it didn't just go viral. It triggered one of the largest capital mobilizations in history. OpenAI, Google, Anthropic, Meta, and xAI began racing toward one goal: AGI, a machine that can outthink humans at everything.
Takeaway: The intelligence race didn't just change tech. It changed what's possible for everyone.
In plain terms“Frontier labs” are the few companies — such as OpenAI, Anthropic, and Google — that build the most powerful models. Most other AI apps are built on top of their models. When you use a creative or study tool, it helps to ask: whose model is underneath, and what happens to the information I share?
Who builds frontier AI, and who pays for it
A handful of labs build the most capable models. Two of them, OpenAI and Anthropic, took about 90% of the $240B that frontier labs raised in 2026 through September.
OpenAI
Raised $122B in Mar 2026 at an $852B valuation. About $182B total.
Anthropic
Raised $65B in May 2026 at a $965B valuation. Total roughly $125B to $145B.
Google DeepMind
Funded by Alphabet; 2026 capex guided to $195B to $205B.
Meta
Funded by Meta; 2026 capex guided to $130B to $145B.
xAI
Raised $20B in Jan 2026; reported to have joined SpaceX in Feb 2026.
Mistral AI
Roughly $6.6B raised in total (estimates vary).
Also in the race: Safe Superintelligence (Ilya Sutskever), Thinking Machines (Mira Murati), DeepSeek, and Moonshot AI. Figures are estimates and change often. Google and Meta fund their labs themselves, so their numbers are capex, not venture capital.
Sources: BlockWest Research (Sep 30, 2026); OpenAI (Mar 31, 2026); Morningstar (May 29, 2026); PitchBook (Jan 2026); Epoch AI via Affarslivet; company earnings guidance via Yield Theory and TMT Finance.
In plain termsHundreds of millions of people now use AI chat tools every week, including many of your classmates, instructors, and future employers. Wide use doesn’t mean every use is good or allowed. It does mean knowing how to use AI responsibly is quickly becoming a basic expectation, like knowing how to search the web.
ChatGPT now reaches 1.2 billion people every week
From 200 million to 1.2 billion weekly users in about two years — roughly one in seven people on Earth.
Weekly users, as reported by OpenAI
- Aug 2024200M
- Dec 2024300M
- Feb 2025400M
- Aug 2025700M
- Oct 2025800M
- Feb 2026900M
- Aug 20261B
- Sep 20261.2B
Takeaway: Company-reported weekly users count people, not prompts, and definitions vary.
Sources: OpenAI DevDay 2026 recap; The Verge; Axios via The Verge (Jul 2025); TechCrunch/a16z; Pew Research Center (Feb 2026).
Check your understanding
Sign in to save your quiz results and earn your certificate.