Anthropic is throttling its Mythos model for certain tasks while building competing applications, prompting pushback from customers, partners, and investors over potential anticompetitive practices.
The AI industry is repeating a pattern Microsoft perfected over decades: leveraging control of a foundational platform to favor proprietary applications.
Anthropicis restricting access to its new Mythos model for specific use cases while simultaneously developing applications that directly compete with companies that depend on its API. This strategy mirrors Microsoft's historical playbook of bundling competitive software into Windows and giving internal products performance advantages.
The complaint
Customers building AI applications on Anthropic's platform say the throttling limits their ability to compete. Partners express concern about investing in Anthropic infrastructure if the company will undercut them with first-party products. Investors worry the strategy could invite regulatory scrutiny and alienate the developer ecosystem.
The parallels to Microsoft are striking. In the 1990s and 2000s, Microsoft bundled Internet Explorer with Windows while making it difficult for Netscape and other browsers to compete effectively. The company optimized Windows performance for its own applications while limiting competitor access to APIs. These practices led to antitrust investigations globally and settlement agreements that constrained Microsoft's behavior for years.
Why it matters
As AI models become foundational infrastructure for software development, control over these platforms carries immense leverage. If one company can simultaneously control access to a critical capability and compete in adjacent markets, it creates structural advantages others cannot overcome through product quality alone.
The AI industry is still early enough to avoid the regulatory and reputational costs Microsoft eventually faced. But the current pushback suggests the sector is watching closely. Companies building on large language models need assurance that platform providers won't exploit their position to eliminate competition.
For Anthropic, the short-term gains from favoring proprietary applications may not offset the long-term cost of eroding developer trust. The company faces a choice: remain a neutral platform provider or accept the constraints that come with competing in its own ecosystem.
Fei-Fei Li, CEO of World Labs, unveiled Atlas, a world model designed to understand and generate 3D environments. Li argues spatial intelligence represents the next major AI frontier beyond language models.
Alibaba unveiled a new AI chip to compete with Nvidia's dominance, signaling a shift in global semiconductor competition. Meanwhile, silicon valley grapples with talent shortages and the race to build advanced AI systems.
European Commission Executive Vice President Teresa Ribiera called for globally respected AI safety standards during remarks to Bloomberg. She emphasized the need for coordinated safety measures across the technology sector.
Anthropic has increased five-hour usage limits across Pro, Max, and Team subscription plans while resetting rate limits for existing users. The changes aim to reduce constraints on paid subscribers.