Meta and Microsoft are drastically reducing their use of Anthropic's Claude model, signaling a shift in how major tech companies approach AI tool deployment. Microsoft cut its monthly per-employee budget for Claude in its cloud division from $100,000 to $10,000, a 90 percent reduction. Meta simultaneously halved its Claude Code user base to 30,000 employees.

Both companies are redirecting resources toward their own internal AI systems instead of relying on Claude. Microsoft has invested heavily in its own models and partnerships, while Meta continues developing its Llama architecture and other proprietary tools. The pullback reflects a broader industry trend where large technology firms prefer building and deploying homegrown AI solutions over depending on third-party providers.

For Anthropic, the shift presents a significant business risk. The company's enterprise revenue model has depended on securing major clients like Meta and Microsoft to achieve scale and profitability. With two of its largest customers substantially reducing Claude usage, Anthropic faces pressure to diversify its customer base and demonstrate Claude's value to mid-market and smaller enterprises.

The timing complicates Anthropic's positioning. The startup emerged from OpenAI to compete directly with established players, positioning itself as a safety-focused alternative to larger competitors. However, as Meta and Microsoft develop competitive models, Anthropic transitions from being a trusted partner to these companies into a direct competitor for enterprise AI market share.

The strategic shift also reflects confidence among major tech companies in their ability to develop competitive AI models internally. Rather than licensing Claude for enterprise use, these firms are investing in proprietary systems they can control, customize, and integrate directly into their products without dependency on external vendors.

Anthropic must now compete for enterprise customers on merit rather than partnership relationships. The company's Claude model remains technically competitive, but without anchor customers driving large-scale adoption, building market momentum becomes harder. Smaller