Companies Reduce AI Costs by Exploring Affordable Chinese Models
AI Expenses: A Rising Concern
Artificial intelligence is proving to be a significant expense for many businesses. Companies are seeking ways to reduce costs by considering more affordable alternatives, such as Chinese AI models.
Example of Lindy.ai’s Approach
Lindy.ai, a startup based in San Francisco, exemplifies this shift. Initially, the company utilized AI models from Anthropic. The expense was greater than that for payroll and rent. Founder Flo Crivello stated, “By far, our No. 1 expense was Anthropic.” This realization led Lindy to switch entirely to DeepSeek-V4, a Chinese AI model, and consequently save millions.
The AI Cost Dilemma
AI expenses are growing rapidly for U.S. businesses, creating a challenging but necessary situation. To manage this, many are transitioning to cost-effective Chinese models. While U.S. companies like Anthropic and OpenAI lead globally, experts suggest Chinese models are behind in capabilities by several months. However, they have excelled in producing open-source models which are free to use and adapt.
“The open-source scene right now is absolutely dominated by the Chinese,” Flo Crivello noted.
Chinese AI Model Impact
The increasing popularity of Chinese models is altering methods companies use AI. For instance, Uber CEO Dara Khosrowshahi disclosed how Uber’s AI budget was depleted quickly and necessitated adjustments. Airbnb’s CEO, Brian Chesky, mentioned the use of Alibaba’s Qwen model, which was noted as “good, fast and cheap.”
Many companies hesitate to publicly announce their use of Chinese models due to political considerations. Yet, these models remain accessible through platforms like Hugging Face and GitHub.
“It’s like the difference between driving a Ferrari and a Honda. You can have the best luxury car, or you can just have a Honda at scale that works,” stated Eugene Cheah of Featherless.
Challenges and Adaptations
Despite their cost benefits, Chinese models may not meet all company needs. Comment.io’s co-founder Jon Gordner expressed reluctance in sacrificing quality for savings. He highlighted that Anthropic and OpenAI provide value by offering token discounts in subscriptions. Yet, he acknowledged that pressure to demonstrate profitability could lead U.S. AI companies to increase prices.
Ramp’s lead economist, Ara Kharazian, believes American companies will adapt by possibly providing competitively priced models. Gordner, however, speculated that major U.S. AI firms might eventually need to raise prices as they approach public offerings.