A growing debate over the cost of artificial intelligence is reshaping the competitive landscape of the industry, with some U.S. companies increasingly turning to Chinese AI developer DeepSeek as businesses search for lower-cost alternatives to premium models from Anthropic and OpenAI.
The trend gained new attention after Microsoft AI Chief Executive Mustafa Suleyman publicly criticized the expense of Anthropic's services, underscoring mounting pressure on companies to justify rapidly rising AI spending as adoption accelerates across corporate America.
Suleyman's remarks came shortly after Microsoft introduced several new AI models at its annual Build conference, where the company emphasized its ambition to compete more aggressively in the market for advanced coding assistants and enterprise AI tools.
"We pay a lot of money to Anthropic-so our goal is to reduce and ultimately eliminate that cost," Suleyman said, according to Bloomberg.
The comments highlight a broader challenge facing businesses. While artificial intelligence promises productivity gains and operational efficiencies, deploying advanced models across large organizations can quickly become expensive. Anthropic's subscription offerings, including premium tiers costing up to $100 per month per user, can generate substantial costs when rolled out across thousands of employees.
For large corporations, those expenses are increasingly becoming part of strategic discussions about vendor selection, infrastructure investments and long-term technology budgets.
The search for lower-cost options has helped fuel interest in DeepSeek, a Chinese AI company backed by hedge fund High-Flyer. According to the South China Morning Post, DeepSeek has attracted growing attention from U.S.-based businesses seeking more affordable AI solutions.
Industry observers say pricing is only part of the appeal.
Unlike many closed commercial systems, DeepSeek has adopted an open-source approach that allows organizations to download, customize and deploy models on their own infrastructure. That flexibility can reduce dependence on external providers and lower operating costs for companies with sufficient technical resources.
Ara Kharazian, lead economist at Ramp Economics Lab, described the shift as a notable development in the AI marketplace.
"In probably the biggest sign that companies are looking for cheaper alternatives to OpenAI and Anthropic, some are willing to use cheaper Chinese models, sending US data back and forth from China-hosted servers," Kharazian wrote in a social media post.
Chinese developers have responded by releasing lighter models designed to operate with fewer computing resources while maintaining competitive performance levels. Industry analysts say that strategy has made such platforms particularly attractive to cost-conscious businesses that may not require the most advanced frontier models.
At the same time, competition is expanding beyond DeepSeek. Companies including Fireworks AI, Fal AI and DeepInfra are also attracting interest from enterprises seeking alternatives to the dominant U.S. providers.
