Google Enforces Terms of Service Amid Antigravity AI Compute Strain

Google has suspended numerous customer accounts accessing its Antigravity AI agent development platform through third-party tools like OpenClaw and OpenCode, citing violations of its Terms of Service (ToS). This enforcement action, impacting users across subscription tiers including $250/month AI Ultra subscribers, follows unprecedented computational demands that degraded service quality. The company characterizes the usage patterns as "malicious," though affected developers contend they operated within purchased quotas.

Compliance Requirements Under Google's Terms

The core violation involves using Antigravity's backend as a proxy for external applications—a practice explicitly prohibited under Section 3.2 of Google's Cloud AI Terms of Service. Key compliance requirements include:

1. Direct Usage Mandate: Services must be accessed via Google-authorized interfaces only; third-party wrappers/harnesses violate ToS
2. Prohibition on Automated Proxy Use: Systems like OpenClaw that route requests through Antigravity constitute unauthorized "service chaining"
3. Commercial Use Restrictions: Subscriptions are for human-supervised development, not autonomous agent deployment at scale

According to Varun Mohan, DeepMind engineer and former Windsurf co-founder, these practices triggered a "massive increase in malicious usage" that overwhelmed infrastructure. "Users in these groups have overwhelmed our compute," Mohan stated, emphasizing that bans specifically target Antigravity misuse while other Google services remain accessible.

Enforcement Timeline and Remediation Path

Google's compliance actions unfolded through February 2026 with immediate effect:

• February 10-20: Unprecedented compute load detected from OpenClaw/OpenCode integrations
• February 21: First wave of account suspensions without prior warning
• February 23: Public acknowledgment of enforcement and degradation of service
• Remediation: Google plans to establish a reinstatement path for users unaware of ToS violations, though capacity remains limited

Affected organizations must immediately:

• Cease all Antigravity integrations with unauthorized wrappers
• Audit development pipelines for indirect API access points
• Migrate autonomous agent workloads to Google's enterprise API tiers with appropriate pricing

Industry Context and Precedents

This enforcement mirrors Anthropic's recent restrictions on token arbitrage, highlighting systemic challenges as AI providers balance growth against operational costs. The incident reveals a broader industry pattern: vendors subsidize token pricing below operational cost to capture market share, creating unsustainable load when developers leverage pricing gaps through third-party tools.

AI engineer Mohan Prakash criticized the abrupt enforcement: "Banning paying customers without warning is how you lose trust faster than you lose capacity." Unlike Anthropic—which blocks unauthorized usage through technical errors—Google's suspension-first approach creates compliance uncertainty.

Practical Compliance Steps

Organizations using Google's AI services should:

1. Review Integration Patterns: Ensure no tools bypass official SDKs/channels
2. Monitor Usage: Track compute consumption against quota limits
3. Document Exceptions: Obtain written approval for unconventional implementations
4. Evaluate API Tiers: Scale to appropriately priced enterprise APIs for automated workflows

Google's Terms of Service remain the binding compliance framework. As Mohan clarified: "It is not intended to use the Antigravity backend as a proxy for other products." Organizations ignoring these provisions risk service disruption regardless of payment status.