English
AI Industry Enters a New Era as Anthropic and OpenAI File for IPO
In a historic week for the artificial intelligence industry, two of the world's leading AI companies—Anthropic and OpenAI—have filed confidential initial public offerings (IPOs) with the U.S. Securities and Exchange Commission in June 2026. The filings mark a pivotal shift from AI as a speculative research venture to a mature public market business, with combined valuations exceeding $1.8 trillion. Anthropic, the creator of the Claude AI model, filed on June 1, followed by OpenAI on June 8. The moves come amid a flurry of regulatory activity, including a new U.S. executive order on AI security and the European Union's proposed Cloud and AI Development Act.
Anthropic's IPO: A $965 Billion Bet on Safe AI
Anthropic, founded in 2021 by former OpenAI researchers Dario and Daniela Amodei, filed its S-1 confidentially on June 1, 2026. The company, which raised $65 billion in private capital at a $965 billion valuation in late May, is now seeking a public listing to broaden its shareholder base and gain acquisition currency. According to the Anthropic IPO valuation and details, the company's annualized revenue run rate reached approximately $47 billion by May 2026, though it remains unprofitable due to massive infrastructure spending on AI compute.
Anthropic's focus on 'constitutional AI' and safety has attracted major investors including Google, Amazon, Sequoia, and Coatue. The company recently acquired software startup Stainless and partnered with xAI for access to its Colossus 1 data center. However, the IPO introduces tensions with its safety-first mission, as public markets demand quarterly growth and faster product cycles.
OpenAI Follows Suit: $852 Billion Valuation and Growing Pains
OpenAI filed its confidential S-1 on June 8, 2026, just one week after Anthropic. The ChatGPT-maker, last valued at $852 billion, is working with Goldman Sachs and Morgan Stanley on the offering. CEO Sam Altman struck a cautious tone, noting that 'it may be a while' before listing, as certain strategic goals are easier to execute as a private company. OpenAI projects a $14 billion loss in 2026 and does not expect profitability until around 2030, according to its filing.
The company reported $2 billion in monthly revenue as of March 2025 and targets $200 billion in annual revenue by 2030. ChatGPT now supports over 900 million weekly active users. However, the IPO faces scrutiny over internal governance concerns from the 2023 boardroom coup, multiple lawsuits including one from Florida over alleged harms to children, and the massive cash burn required for AI research computing—expected to reach $122 billion by 2028.
US Executive Order 14409: Promoting AI Innovation and Security
On June 2, 2026, President Donald Trump signed Executive Order 14409, titled 'Promoting Advanced Artificial Intelligence Innovation and Security.' The order, published in the Federal Register on June 5, establishes a voluntary framework for AI developers to share frontier models with the government before public release. Key provisions include:
- Creating a classified benchmarking process under the NSA to assess AI model capabilities and designate 'covered frontier models'
- Establishing an AI cybersecurity clearinghouse between Treasury, NSA, and CISA
- Directing federal agencies to prioritize cyber defense of national security systems within 30 days
- Prioritizing enforcement of federal criminal laws against AI-enabled cybercrimes
The order explicitly avoids creating mandatory government licensing or preclearance for AI models, maintaining the administration's 'America First' approach to AI dominance. It also slashes bureaucratic constraints from the prior administration, according to the White House.
Europe's Response: The Cloud and AI Development Act
On June 3, 2026, the European Commission proposed the Cloud and AI Development Act (CADA), introducing the first EU-wide four-tier sovereignty assurance framework for cloud and AI services. The framework establishes four levels: Level 1 (data residency), Level 2 (supply-chain independence from third countries), Level 3 (EU-domiciled ownership and control), and Level 4 (complete supply-chain transparency for defense workloads).
CADA directly targets exposure to extraterritorial legislation like the U.S. CLOUD Act, creating compliance tensions for U.S.-domiciled cloud providers serving EU public sector customers. The proposal also introduces an 'open source-first' procurement principle, requiring software bill of materials (SBOM) practices and model provenance verification for EU public sector contracts. The act is part of a broader AI Continent Action Plan that includes tripling EU data centre capacity within 5–7 years.
Microsoft's MAI Models: Building a Hill-Climbing Machine
At its Build 2026 conference on June 2, Microsoft launched seven in-house MAI models across reasoning, coding, image, voice, and speech—all trained from scratch with zero distillation from other labs. The flagship MAI-Thinking-1 is a 35B-active parameter reasoning model with 256K context, designed to compete with leading models like Claude Opus 4.8. Other models include MAI-Code-1-Flash (5B agentic coding model), MAI-Image-2.5, and MAI-Transcribe-1.5 (supporting 43 languages).
Microsoft also introduced Frontier Tuning, a reinforcement learning-based customization system that trains models on an enterprise's own operational workflow data inside its compliance boundary. The models run on Microsoft's own Maia 200 accelerators and are deeply integrated into GitHub Copilot, VS Code, and other Microsoft products. This marks a strategic shift toward self-sufficiency, reducing reliance on OpenAI technology.
Google's Gemma 4: AI for Everyone's Laptop
Google DeepMind released Gemma 4 12B on June 3, 2026, a multimodal AI model designed to run locally on laptops with just 16GB of VRAM. Released under an Apache 2.0 license, it features a novel encoder-free unified architecture where vision and audio inputs flow directly into the LLM backbone, eliminating traditional separate encoders to reduce latency and memory usage. This is the first mid-sized Gemma model with native audio input support.
Gemma 4 delivers benchmark performance nearing the larger 26B MoE model at less than half the memory footprint, enabling powerful multi-step reasoning and agentic workflows on consumer hardware. The model supports 140 languages and has surpassed 150 million downloads across the Gemma family.
Impact and Implications
The convergence of IPO filings, new regulations, and product launches in June 2026 signals that AI has entered a new phase of maturity. The AI IPO market trends in 2026 are reshaping the investment landscape, with public market investors gaining access to previously private AI giants. However, tensions remain between safety-focused missions and quarterly growth demands. The U.S. and EU are taking divergent approaches to AI regulation, with the U.S. favoring voluntary frameworks and the EU pushing for binding sovereignty requirements.
The race for AI dominance is now playing out on multiple fronts: public markets, regulatory frameworks, and technological innovation. As these developments unfold, the global AI landscape will be defined by the balance between innovation, security, and governance.
Frequently Asked Questions
What is the valuation of Anthropic's IPO?
Anthropic filed for IPO at a $965 billion valuation after raising $65 billion in private capital. The company's annualized revenue run rate reached approximately $47 billion by May 2026.
When did OpenAI file for IPO?
OpenAI filed a confidential S-1 with the SEC on June 8, 2026, targeting a valuation of up to $1 trillion. The company is valued at $852 billion and projects a $14 billion loss in 2026.
What does Executive Order 14409 do?
Executive Order 14409, signed June 2, 2026, establishes a voluntary framework for AI developers to share frontier models with the government for security testing, creates an AI cybersecurity clearinghouse, and prioritizes enforcement against AI-enabled cybercrimes—without imposing mandatory licensing.
What is the EU Cloud and AI Development Act?
Proposed on June 3, 2026, CADA introduces a four-tier sovereignty framework for cloud and AI services, targeting exposure to non-EU legislation. It aims to triple EU data centre capacity within 5–7 years and mandates open-source-first procurement for public sector contracts.
What are Microsoft's new MAI models?
Microsoft launched seven in-house MAI models at Build 2026, including MAI-Thinking-1 (reasoning), MAI-Code-1-Flash (coding), and MAI-Transcribe-1.5 (speech). All are trained from scratch without distillation from other labs.
Follow Discussion