Microsoft CEO Satya Nadella has publicly criticized Anthropic's restrictions on its AI models, calling them illogical and warning that such practices could foster monopolistic risks in the artificial intelligence industry. Nadella urged enterprises to diversify their AI dependencies rather than rely on a single provider. His comments arrive as the AI sector faces intensifying regulatory scrutiny worldwide.
Why the restrictions drew fire
Nadella argued that limiting how third parties can use AI models goes against the broader goal of advancing the technology. In his view, restrictions that prevent developers from customizing or integrating models freely don't just stifle innovation — they create a bottleneck that could slow down the entire industry. He did not single out specific policies at Anthropic, but the company has been known for tight controls on how its Claude models are deployed, especially in high-risk applications.
The Microsoft chief's criticism is notable because his company is a major investor in OpenAI, a direct competitor to Anthropic. Nadella's stance suggests he sees the market shifting toward a model where openness and flexibility become competitive advantages.
Monopoly fears in a fast-concentrating market
Nadella highlighted the risk of monopolistic behavior in AI, a concern that has bubbled up as a handful of companies — including Microsoft-backed OpenAI, Google, and Anthropic — come to dominate the space. He warned that if enterprises lock themselves into a single AI provider, they give that provider outsized influence over their operations and data. Over time, that could lead to higher costs, less innovation, and a dependency that's hard to break.
His warning echoes broader antitrust discussions in Washington and Brussels. Regulators have already begun probing whether big tech firms are using their control over AI models to squeeze out smaller competitors. The Federal Trade Commission, for example, has opened inquiries into partnerships between major cloud providers and AI startups.
Nadella's push for enterprise diversification
Nadella didn't just criticize — he offered a prescription. Enterprises, he said, should actively diversify their AI suppliers. Building systems that can switch between different models or use multiple models in parallel would reduce risk and keep the market competitive. He pointed to Microsoft's own Azure platform, which offers access to models from OpenAI, Meta, and others, as a model for how companies can avoid vendor lock-in.
His advice is timely. Many businesses rushed to adopt AI tools over the past two years, often signing exclusive deals with a single provider. As the technology matures, those quick decisions are starting to look like liabilities. The cost of switching models can be high, but the cost of not having a backup could be higher.
The regulatory backdrop
Nadella's remarks come as regulators around the world are trying to figure out how to govern AI without killing innovation. The European Union's AI Act is set to impose strict rules on high-risk systems, and the U.S. is considering its own framework. One of the biggest unanswered questions is how to handle the power of a few companies that control the most advanced models. Nadella's intervention puts additional pressure on policymakers to address the concentration issue.
Anthropic has not responded to Nadella's comments. The company has previously defended its model restrictions as necessary for safety, arguing that uncontrolled access could lead to misuse. But Nadella's critique suggests that safety and competition are not always aligned — and that the industry's biggest players are starting to clash over which principle should come first.
The debate over AI model access and market power is unlikely to fade. With regulators watching and customers growing more cautious, the pressure on companies like Anthropic to justify their restrictions — or to loosen them — is only set to increase.
