Google Suspends OpenClaw Users from Antigravity AI Platform Over OAuth Abuse

Google suspended numerous OpenClaw users from its Antigravity AI platform after detecting OAuth token abuse. Developers used OpenClaw’s plugin to access subsidized Gemini model tokens at scale, violating terms of service. This caused backend service degradation across the platform.

OpenClaw gained massive traction since November 2025 launch. Over 219,000 GitHub stars fueled adoption for local AI agents handling email and browsing tasks. Users authenticated through Antigravity to reach high-end Gemini 2.5 Pro models cheaply.

BEST WINTER DEALS

Editor's Choice

Private Internet Access

Access content across the globe at the highest speed rate.

70% of our readers choose Private Internet Access

70% of our readers choose ExpressVPN

ExpressVPN

Browse the web from multiple devices with industry-standard security protocols.

Nord VPN

Faster dedicated servers for specific actions (currently at summer discounts)

Third-party OAuth bypassed official channels completely. Google’s automated systems flagged “malicious usage” patterns from heavy token arbitrage. Legitimate paying customers faced slowdowns as shared capacity collapsed.

Mid-February suspensions hit AI Ultra subscribers hard. Monthly $249.99 plans lost access without warning. Some saw full Google account restrictions including Gmail and Workspace. Forum posts detail widespread 403 errors and lost account histories.

Varun Mohan, DeepMind product lead, addressed the crisis publicly. He prioritized “actual users” amid capacity limits. Limited reinstatement offered to unaware violators only. This mirrors Anthropic’s recent OpenClaw OAuth ban.

OpenClaw vulnerabilities compound the fallout. Over 21,000 exposed instances risk infostealers targeting configs. Supply chain attacks hit agent deployments too. China’s ministry warned of misconfigured systems enabling broader breaches.

Peter Steinberger, OpenClaw creator now at OpenAI, called bans “draconian.” He plans dropping Antigravity support entirely. Community forks like Nanobot and IronClaw gain steam quickly.

Suspension Impact Table

Developers now scramble for alternatives. OpenAI welcomes third-party harnesses explicitly. This exposes growing AI ecosystem fractures over subsidized access.

Ban waves highlight backend costs of free tiers. Rapid agent proliferation strains shared infrastructure. Enterprises face sudden productivity gaps from tool bans.

Google AI Developer Forum threads show frustration. Users demand clearer ToS and warnings. Capacity issues reveal limits of current scaling approaches.

Security researchers note OpenClaw’s ClawDBot flaw enables RCE. Exposed agents amplify risks beyond ToS violations. Agentic AI demands tighter configs.

Community pivots fast. IronClaw fork adds native auth. Nanobot emphasizes self-hosting. Developers weigh convenience against platform lock-in risks.

This enforcement accelerates closed AI ecosystems. Subsidized tokens carry hidden compliance burdens. Ban waves remind builders of infrastructure realities.

Key Takeaways

• OAuth arbitrage violates ToS across providers.
• 219K GitHub stars drove unsustainable usage.
• Varying ban severity hits free and paid users.
• Forks fill gaps as creator exits support.

AI agent wars intensify. Google prioritizes enterprise stability. Developers adapt through diversification.

FAQ