OpenAI and Anthropic Are Going Public. Together.
Both companies confidentially filed for IPO in early June — the same week, at nearly the same trillion-dollar valuation
It’s the biggest event AI has ever had on Wall Street. Two rivals that spent five years defining the frontier are now heading into the public markets at the same time. Here’s what the timing tells us — and what it changes for everyone using their models.
In early June 2026, both OpenAI and Anthropic quietly filed confidential S-1 paperwork with the U.S. Securities and Exchange Commission — the first step toward a public listing. Neither company has confirmed the specifics, but multiple outlets including CNBC and Reuters reported the filings within days of each other.
What makes it strange isn’t the IPOs themselves. It’s the timing. These two companies are the fiercest rivals in generative AI, and they both decided that right now — the same week — was the moment to head for the public markets. That’s not a coincidence.
Both filed within days in early June
Confidential S-1s at similar ~$1T valuations
Growth rates are near their peak
Analysts say list-now-or-wait-longer window
OpenAI run rate pacing ~$25B
Up from $13.1B in 2025 — nearly 2x YoY
Enterprise customers cutting AI spend
“Tokenmaxxing” era is quietly ending
The current growth rate is the fastest
it will ever be. That’s mostly basic math.
Growth is peaking, and both companies know it
The mathOpenAI’s revenue jumped from roughly $13.1B in 2025 to a run rate pacing around $25B this year, according to reporting cited by CNBC. Anthropic’s growth curve looks similar — enterprise-led, heavy in the coding and knowledge-work stack.
The problem isn’t the numbers. It’s the second derivative. Growth like this comes from a small number of very large customers loading up on tokens. As those customers start looking at their bills, the future growth curve flattens. Filing now locks in a valuation set against the peak, before the flattening shows up in the S-1.
The “tokenmaxxing” era is quietly ending
The pressureCNBC’s late-June reporting cited concrete examples: Lindy CEO Flo Crivello switched his entire company off Claude and moved 100% of traffic to DeepSeek’s cheaper models. “You could see that cost curve go down, like, crash to the ground,” he told CNBC.
That’s one company, but the pattern is showing up across the ecosystem — Microsoft launching low-cost models, GitHub Copilot routing tasks to the cheapest capable model, Amazon and Google pushing their own frontier candidates. The era of paying whatever it costs to use the best model is turning into an era of “use the cheapest model that does the job.”
Filing “together” isn’t cooperation — it’s mirroring
The dynamicThe two labs didn’t coordinate. But they read the same market signals — peaking growth curves, cooling enterprise spend, competitive pressure from open-source and Chinese models, and a Wall Street window that might not stay open forever.
When two rivals arrive at the same conclusion in the same week, that itself is data. It tells you the strategic calculus for frontier AI has shifted from “raise as much private capital as possible” to “get to the public market before conditions change.”
Consumer scale + enterprise depth
ChatGPT is one of the fastest-growing consumer products in history. Enterprise is layered on top. Revenue mix is broader but harder to forecast because consumer subscriptions swing.
Enterprise-first, developer-heavy
Claude’s strength has been API and enterprise integrations — coding tools, regulated industries, consulting partnerships with TCS and DXC. Revenue mix is narrower but easier to project.
- For developers — expect more predictable, but not necessarily cheaper, pricing as public reporting begins
- For enterprise buyers — negotiating leverage improves as public scrutiny raises transparency
- For competitors (Google, Meta, xAI) — a public OpenAI or Anthropic means public financials to benchmark against
- For investors — the first “pure-play” frontier AI listings — nothing else on public markets looks like this yet
- For open-source models — a tailwind, as cost pressure keeps pushing companies toward cheaper alternatives
- For end users — the pace of new model releases likely doesn’t slow — public markets reward growth
⚠️ What’s still uncertain
Confidential S-1 filings don’t lock in a timeline. Either company could delay, pull the filing, or shift strategy based on market conditions. Reported valuations are pre-marketing estimates, not offering prices.
Anthropic declined to comment on the CNBC report. OpenAI didn’t respond. Everything above is based on reporting from CNBC, Reuters, and TechCrunch — not official company disclosures.