In 2023, the AI world talked about OpenAI as the dominant force and Anthropic as the thoughtful underdog. In March 2026, the picture is significantly more complex. Anthropic's Claude Sonnet 4.6 is the model professional developers and enterprise buyers prefer for coding and agentic work. OpenAI's ChatGPT is still the consumer product with by far the largest user base — 400M+ weekly active users. But OpenAI has been rocked by controversy: the Sam Altman board drama, the QuitGPT boycott after the Pentagon deal, and an exodus of safety researchers. Understanding these two companies is essential context for anyone who relies on their tools.
The Origin Story: Anthropic as OpenAI's Conscience
Anthropic was founded in 2021 by Dario Amodei, Daniela Amodei, and seven other researchers — all from OpenAI. Dario had been OpenAI's VP of Research. He resigned because he believed OpenAI was moving too fast on safety, prioritising deployment speed over careful safety research. Anthropic was built on the premise that AI safety and capability should be developed together. This is reflected in everything from their Constitutional AI training approach to their Responsible Scaling Policy — both published openly.
OpenAI in 2026: Scale, Controversy, and $157 Billion
- Financials: Valued at $157 billion. Annual revenue approaching $5B. Microsoft committed $10B+ as strategic partner.
- Products: ChatGPT (400M+ weekly users), Sora (video), GPT-4.1 family, o3 and o4-mini (reasoning), OpenAI API.
- Controversies: The Sam Altman board firing and rehiring in 2023. Ilya Sutskever (co-founder) departed to found Safe Superintelligence Inc. Multiple safety researchers left publicly. The Pentagon deal triggering QuitGPT.
- Governance: Transitioned to capped-profit structure. Original safety mission now explicitly balanced against fiduciary responsibility to shareholders.
Anthropic in 2026: The Enterprise Challenger
- Financials: Valued at approximately $60-70B. Amazon invested $4B in tranches. Google also invested. Total raised approximately $7.5B. Public Benefit Corporation structure.
- Products: Claude family (Opus 4.6, Sonnet 4.6, Haiku), Claude Code, Claude.ai consumer product, Anthropic API.
- Differentiation: Constitutional AI research published openly. Safety-first reputation gives credibility with enterprise buyers in regulated industries — healthcare, finance, government.
| Dimension | OpenAI / ChatGPT | Anthropic / Claude |
|---|---|---|
| Consumer mindshare | Clear winner — 400M+ weekly users | Growing but much smaller |
| Developer/enterprise preference | Strong broad ecosystem | Winning — preferred for coding + agents |
| Safety reputation | Declining after controversy | Strong — primary differentiator |
| Technical capability | GPT-5.4 competitive on most tasks | Sonnet 4.6 leads on coding + agents |
| Governance clarity | Complicated by profit transition | PBC structure — clearer mission |
The honest verdict: OpenAI is winning on consumer scale. Anthropic is winning on developer and enterprise trust. Neither is clearly dominant. Anthropic has positioned Claude as the professional's choice while OpenAI's ChatGPT is the everyone's choice — both markets can be very large simultaneously.
Pro Tip: For students choosing which company's tools to learn: use both. OpenAI's and Anthropic's ecosystems have different strengths, and in your career you will almost certainly use both. Understanding each company's philosophy also helps you engage meaningfully with the AI policy debates that will affect your industry for the rest of your career.