Anthropic surpassed OpenAI, becoming the leader after the major event HumanX in the field of artificial intelligence

Anthropic Captures Industry Attention at HumanX

At the HumanX conference in San Francisco, where about 6,500 executives and founders gathered, Anthropic drew widespread attention. The main topic of discussion was the Claude Code system – a tool for generating software code that strengthened the company’s position among large clients and placed it at the center of the fastest-growing applied AI segment.

What Experts Say
- CNBC surveyed 19 leaders and investors who identified Claude Code as the most prominent product on the market. They also noted serious competitors: OpenAI, Cursor, and Google.
- For Anthropic, this success is especially important because the company has already proven its effectiveness working with large clients. Successes in creating, editing, and reviewing code increase the chances of securing contracts with the biggest buyers.
- Although OpenAI launched an AI boom with ChatGPT in 2022, Anthropic now appears better prepared to compete for the largest budgets.

Legal Conflict with the Pentagon
Despite a public dispute with the U.S. Department of Defense (DoD) that led to Claude being blacklisted, two opposing court rulings allowed Anthropic to continue working with other federal agencies. The company was founded in 2021 after a group of researchers left OpenAI and is valued at roughly $380 million.

- Claude Code was made available to the public in May 2025.
- As of February this year, current annual revenue exceeded $2.5 million.

New Model – Claude Mythos Preview
On Tuesday, Anthropic unveiled a new AI model, Claude Mythos Preview, focused on cyber defense. It leverages strong programming and reasoning skills:

- At HumanX the novelty sparked notable interest, but only about 50 companies gained access.
- Synthesia CEO Victor Riparbell explained Anthropic’s success by focusing on code generation rather than video and voice AI models. According to him, OpenAI promoted six different products at once, diluting consumer attention.
- One investor noted that the market is still too young and current momentum could easily shift.

How Work in Companies Is Changing
Decagon
President Ashwin Srinivasan said that the emergence of code-writing tools has changed both hiring and organizational work. Candidates are now allowed to use such tools in interviews, and projects that previously required 4–5 engineers can now be handled by two.

Credo AI
Chief Naveen Singh highlighted that tasks which would have taken 10 people last year can now be assembled over a weekend and deployed within the company. However, it has become harder to maintain control over development plans and obligations to large clients who need greater clarity and stability.

Cisco
President Jitu Patel reported that AI is already used by about 85 % of engineering staff (≈18,000 employees). Cisco arrived at this not in the way leadership expected: first the company prioritized self-implementation rather than results. Patel suggested viewing systems as digital colleagues: a team can consist of two people and six software assistants or even an infinite number of such assistants.

Concerns About Open Chinese Models
Most participants in CNBC’s survey express worry about open-weight Chinese AI models – systems whose parameters are publicly available, allowing improvements to answers and predictions during training. As of April, such models (GLM‑5.1, Kimi K2.5, Qwen3.5) led industry tests. American companies are already actively using these developments.

Cursor IDE built its AI model… *(text cuts off)*