OpenAI and Anthropic File for IPOs in Historic AI Market Moment
In what may be the most consequential week for AI industry financing in history, both Anthropic and OpenAI have now confidentially filed S-1 registration statements with the SEC, setting the stage for an unprecedented dual IPO season.
Anthropic moved first, submitting its draft S-1 on June 1, fresh off closing a massive $65 billion Series H round that valued the company at a staggering $965 billion — surpassing OpenAI's private valuation for the first time. The Claude maker reports an annual run-rate revenue of $47 billion and expects to reach profitability soon. Apollo Global Management and Blackstone structured a $36 billion private credit deal for TPU chip procurement as part of the round.
OpenAI followed suit on June 8, filing its own S-1 at an $852 billion valuation, with Goldman Sachs and Morgan Stanley leading the process. Analysts project a September–November 2026 listing window. The combined private valuation of these two companies alone now exceeds $1.8 trillion — more than the GDP of most countries.
The timing is no coincidence. Both companies are riding a wave of explosive growth: ChatGPT reached 1 billion monthly active users in May 2026, while Claude's web traffic has surged 640% year-over-year, growing from 203 million visits in January to 824 million in April.
Adding context to the moment, today also marks the first day of trading for SpaceX on Nasdaq under the ticker SPCX — a $1.77 trillion IPO that underscores how tech giants are rushing to public markets while conditions remain favorable.
Great American AI Act: Congress Proposes First Federal AI Regulation Framework
On June 4, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a discussion draft of the Great American AI Act (GAAIA), the first bipartisan attempt at comprehensive federal AI governance in the United States.
The bill targets "large frontier developers" — defined as entities with over $500 million in annual revenue that have trained models using more than 1026 floating-point operations. These developers would be required to:
- Disclose information about frontier models to the government
- Obtain third-party audits through Independent Verification Organizations (IVOs)
- Report critical safety incidents
- Provide whistleblower protections for employees
Perhaps the most controversial provision is a three-year preemption of state AI laws that specifically target model development. This would override emerging state regulations like Colorado's comprehensive AI law, set to take effect on June 30 — though it would not affect state laws of "general applicability" like the CCPA.
The bill also includes a workforce protection measure requiring employers to give 60 days' advance notice when AI is a "substantial factor" in a mass layoff, along with civil penalties of up to $1 million per violation per day.
While the bill has not yet been formally introduced in Congress, it signals a clear shift toward federal oversight of AI development — a topic that gains urgency as the EU AI Act's bulk provisions begin applying on August 2, 2026.
GitHub Copilot's Token Billing Switch Triggers Developer Revolt
GitHub's June 1 switch from flat-rate subscriptions to token-based "AI Credits" billing for Copilot has generated fierce backlash from the developer community, with some users reporting cost projections of 10x to 50x their previous bills.
Under the old system, developers paid a predictable monthly rate with fallback to lower-cost models when premium requests were exhausted. The new system charges based on actual token consumption — and the safety net of automatic model downgrading is gone.
The impact has been particularly brutal for users of agentic coding sessions, where a single extended session can consume $30–40 in tokens. One developer on Reddit reported that their projected monthly costs would jump from $29 to $750, while another estimated an increase from $50 to $3,000 for their team.
GitHub has attempted to ease the transition with temporary promotional credits — an extra $30/month for Business and $70/month for Enterprise customers through August — but developers remain skeptical. The move comes as competitors like Cursor, Windsurf, and Claude Code continue to gain market share in the AI-assisted development space.
The billing change reflects a broader industry tension: as AI coding tools become more capable through agentic workflows, the computational costs become harder to absorb under flat-rate pricing — but usage-based billing can create unpredictable costs that undermine developer trust.
The AI Price War Goes Structural
June 2026 marks the moment the AI pricing competition shifted from tactical discounting to structural market reshaping. DeepSeek's V4 Pro model received a permanent 75% price cut effective May 31, dropping to $0.003625–$0.87 per million tokens while maintaining a 1-million-token context window. These prices now sit below every major Western frontier model.
Meanwhile, Google's Gemini 3.5 Flash — which already surpasses previous Pro-tier performance — is positioned as the "default tier for agentic-first developers," pushing premium capabilities into the commodity pricing layer. The highly anticipated Gemini 3.5 Pro, featuring a 2-million-token context window and "Deep Think" reasoning mode, still awaits general availability after Sundar Pichai's Google I/O promise of "next month."
On the Microsoft side, Build 2026 showcased seven new in-house MAI models, including MAI-Code-1-Flash with its 60% token efficiency improvement. These first-party models represent Microsoft's growing independence from OpenAI, following the April 2026 lifting of partnership exclusivity restrictions.
The competitive dynamics are reshaping how enterprises think about AI procurement. With DeepSeek undercutting on price, Google pushing capability downmarket, and Microsoft diversifying its model portfolio, the era of any single provider commanding premium pricing across the board appears to be ending.
Quick Hits
Claude Opus 4.8 launched — Anthropic's latest flagship is four times less likely to miss code flaws and introduces Dynamic Workflows for parallel sub-agent processing. More →
OpenAI + Oracle integration — Enterprise customers can now access OpenAI's frontier models and Codex through existing Oracle Universal Credits, leveraging the Stargate infrastructure commitment. More →
Colorado AI law takes effect June 30 — Unless preempted by the still-unintroduced GAAIA, Colorado's comprehensive AI legislation will begin enforcement in 18 days, creating urgency for companies that delayed compliance. More →