2 min read
Chinese AI startup DeepSeek recently made a bold claim about the profitability of its AI models but with some caveats.
In a post on X, DeepSeek asserted that its online services operate with a “cost profit margin” of 545%. However, this figure is based on what it calls “theoretical income.”
🚀 Day 6 of #OpenSourceWeek: One More Thing – DeepSeek-V3/R1 Inference System Overview
Optimized throughput and latency via:
🔧 Cross-node EP-powered batch scaling
🔄 Computation-communication overlap
⚖️ Load balancingStatistics of DeepSeek’s Online Service:
⚡ 73.7k/14.8k…— DeepSeek (@deepseek_ai) March 1, 2025
The company elaborated on these numbers in a detailed GitHub post, where it discussed its methods for achieving “higher throughput and lower latency.” According to DeepSeek, if all usage of its V3 and R1 models over a 24-hour period had been billed at R1’s pricing, the company would have generated $562,027 in daily revenue.
At the same time, the expense of leasing the required GPUs (graphics processing units) would have only amounted to $87,072.
Despite these figures, DeepSeek acknowledged that its actual revenue is “significantly lower” due to several factors, including discounted nighttime rates, lower pricing for the V3 model, and the fact that “only a subset of services are monetized.” Its web and app access, for instance, remains free.
Naturally, if the website and app were not free and if discounts were removed, user activity would likely decrease. This suggests that DeepSeek’s projections are largely speculative—less a reflection of its current financial performance and more an indication of what its profit margins could look like under ideal conditions.
DeepSeek is releasing these figures at a time when there’s growing debate over the costs and potential profitability of AI. The company first gained attention in January with a model that reportedly performed on par with OpenAI’s o1 on certain benchmarks, despite being built at a fraction of the cost and amid U.S. trade restrictions that limit Chinese firms’ access to high-end chips. The development sent ripples through the tech industry, leading to a dip in tech stock prices and renewed scrutiny of AI-related expenditures.
DeepSeek’s advancements didn’t just shake up Wall Street. Its app briefly overtook OpenAI’s ChatGPT as the top-ranking app on Apple’s App Store. However, it has since dropped in rankings and currently holds the #6 spot in the productivity category, trailing behind ChatGPT, Grok, and Google Gemini.
