OpenAI Opens ChatGPT to OpenClaw's 3.2M Users While Anthropic Blocks Access

OpenAI has made ChatGPT subscriptions the authentication layer for OpenClaw, the open‑source AI agent framework with 346K GitHub stars and 3.2M users. Anthropic blocked Claude subscriptions from the same platform in April. The split defines two opposing strategies for the agent era.

At 2:33 a.m. on May 2, Sam Altman posted on X, TNW reports: "you can sign in to openclaw with your chatgpt account now and use your subscription there! happy lobstering." The casual delivery masked what TNW calls anything but a minor update: OpenAI has made its ChatGPT subscription the authentication and billing layer for OpenClaw, the fastest‑growing open‑source project in GitHub history .

OpenClaw — created by Austrian developer Peter Steinberger in November 2025 — accumulated 346,000 GitHub stars in under five months, surpassing React's ten‑year record in 60 days. It now has 3.2 million users . Nvidia CEO Jensen Huang called it "the most popular open‑source project in the history of humanity" at GTC in March, as quoted by TNW .

OpenClaw's rapid growth has been accompanied by equally rapid security failures. In late January, a critical remote code execution vulnerability ( CVE‑2026‑25253 ) was disclosed: any website a user visited could silently connect to the agent's local server through an unvalidated WebSocket, chaining a cross‑site hijack into full code execution.

Security researchers audited ClawHub, OpenClaw's skills marketplace, and found 824 confirmed malicious entries out of 10,700 available skills , with 335 traced to a single coordinated attack operation, per TNW . More than 30,000 OpenClaw instances were found exposed on the public internet without authentication. Moltbook, the social layer for agents, suffered a breach that exposed 1.5 million API tokens and thousands of private conversations.

The vulnerabilities have been patched in current versions, but a significant portion of the installed base runs older, unpatched versions. OpenAI 's decision to tie its ChatGPT subscription — and by extension its brand, billing system, and user credentials — to OpenClaw means those credentials now flow through an open‑source platform that has had more security incidents in four months than most enterprise software accumulates in a decade .

OpenAI didn't build the most popular AI agent in the world. It hired the developer, backed the foundation, and opened the login. Steinberger joined OpenAI in February to drive the generation of personal agents, and OpenClaw was moved to an independent foundation with OpenAI's continued support and funding.

The ecosystem is expanding rapidly. Nvidia turned OpenClaw into an enterprise platform with NemoClaw , adding security hardening and compliance features. Tencent launched ClawPro for the Chinese market. Meta launched Manus AI as a desktop agent, a competing approach that runs as a native application rather than through messaging apps.

The foundation structure gives OpenAI deniability — OpenClaw remains open source and compatible with multiple model providers. But with Anthropic blocking access and OpenAI enabling it, TNW notes the practical effect is that OpenClaw's three million users are being funnelled toward ChatGPT as their default model. The foundation creates distance. The subscription integration creates lock‑in.

The OpenClaw split reveals the two opposing strategies that will define the AI agent era. Anthropic 's approach is margin‑first : protect per‑token economics, block flat‑rate abuse through agent frameworks, and grow revenue through enterprise contracts and API usage. OpenAI 's approach is distribution‑first : subsidize agent usage through the subscription tier, acquire users at negative marginal cost, and bet that lifetime value justifies the near‑term compute expense.

For builders using agent frameworks, the practical implications are immediate. If you're building on OpenClaw, ChatGPT is now the path of least resistance — one login, $23 per month, no per‑token charges. Meanwhile, Motley Fool notes that OpenAI is under pressure to grow revenue after missing internal targets — making the OpenClaw distribution play even more strategically important. If you want Claude as your agent backend, you'll need to pay per‑token API rates, which can be significantly more expensive for autonomous agents that make thousands of calls.

Amazon posted $181.5B in Q1 2026 revenue with AWS growing 28%, but $16.8B of the record $30.3B net income came from marking up its Anthropic stake — not from selling products. Free cash flow collapsed 95% to $1.2B as capex hit $44.2B.

OpenAI reportedly missed internal benchmarks on revenue and user growth at the end of 2025. CFO Sarah Friar has expressed concerns affording compute contracts. Anthropic and Google are taking market as the AI leader faces mounting competitive pressure.

The Pentagon signed AI agreements with Google, Nvidia, OpenAI, Microsoft, AWS, Oracle, SpaceX, and Reflection AI for classified networks. Anthropic remains blacklisted as a supply-chain risk after refusing to accept a 'lawful use' standard — the first American company to receive that designation.

Anthropic quietly tested removing Claude Code from its $20 Pro plan, sparking backlash from developers who rely on the terminal-based coding agent. The move signals growing infrastructure strain as Claude's popularity surges, forcing builders to consider alternatives.