- Joined
- Jan 22, 2026
- Messages
- 101
- Reaction score
- 920
The company massively deletes developer accounts following the Anthropic.
Google is massively blocking the accounts of its paid subscribers, including those who pay $250 per month for the AI Ultra tariff. The reason was the use of the company's services through third-party tools like OpenClaw and OpenCode. The blockages occurred without warning and caused a wave of outrage among developers who are discussing what is happening on the forums.The problem arose due to the fact that some users connected third-party agent shells to Google's internal backend, the Antigravity platform and Gemini CLI. In fact, the subscription turned into a cheap proxy for heavy offline tasks, for which the service was not originally designed either technically or for the price. The workload increased dramatically, and Google decided to act quickly.The co-founder of Windsor and the current engineer of DeepMind, Varun Mohan, called the incident a "malicious use" that seriously worsened the quality of service for ordinary users. In a post on the social network, he promised to return some of the blocked accounts - those whose owners did not know that they were violating the rules. However, users themselves disagree with this interpretation."People paid for the quota, used the quota within the limits, and got banned. This is not malicious use, this is the use of a product that was sold to them," wrote AI engineer Mohan Prakash in his post. He also pointed out that the rules of the service did not explicitly prohibit integration with OpenClaw, which means that users reasonably considered it acceptable. He cited Anthropic as an example of correct behavior: in similar cases, it simply returns an error instead of banning accounts.The story with Google repeats what happened with Anthropic a little earlier - it also blocked users who connected subscriptions to third-party services instead of the more expensive API. Both cases point to the same systemic problem: AI companies sell tokens significantly below cost in an effort to capture the market in the expectation that they will be able to raise prices later. When users start using the service too actively, it turns out that the business model simply cannot stand it. The price to pay for this is the trust of users and the reputation of the company.
Google is massively blocking the accounts of its paid subscribers, including those who pay $250 per month for the AI Ultra tariff. The reason was the use of the company's services through third-party tools like OpenClaw and OpenCode. The blockages occurred without warning and caused a wave of outrage among developers who are discussing what is happening on the forums.The problem arose due to the fact that some users connected third-party agent shells to Google's internal backend, the Antigravity platform and Gemini CLI. In fact, the subscription turned into a cheap proxy for heavy offline tasks, for which the service was not originally designed either technically or for the price. The workload increased dramatically, and Google decided to act quickly.The co-founder of Windsor and the current engineer of DeepMind, Varun Mohan, called the incident a "malicious use" that seriously worsened the quality of service for ordinary users. In a post on the social network, he promised to return some of the blocked accounts - those whose owners did not know that they were violating the rules. However, users themselves disagree with this interpretation."People paid for the quota, used the quota within the limits, and got banned. This is not malicious use, this is the use of a product that was sold to them," wrote AI engineer Mohan Prakash in his post. He also pointed out that the rules of the service did not explicitly prohibit integration with OpenClaw, which means that users reasonably considered it acceptable. He cited Anthropic as an example of correct behavior: in similar cases, it simply returns an error instead of banning accounts.The story with Google repeats what happened with Anthropic a little earlier - it also blocked users who connected subscriptions to third-party services instead of the more expensive API. Both cases point to the same systemic problem: AI companies sell tokens significantly below cost in an effort to capture the market in the expectation that they will be able to raise prices later. When users start using the service too actively, it turns out that the business model simply cannot stand it. The price to pay for this is the trust of users and the reputation of the company.
