Chinese AI startup DeepSeek has stirred conversation in the tech community by claiming that its AI models could achieve a “cost profit margin” of 545% under ideal conditions.
In a post on X, the company presented calculations based on “theoretical income” that suggest its online services, if fully billed at premium rates, could generate substantial daily revenue relative to operating costs.
According to DeepSeek, a 24-hour period of usage for its V3 and R1 models would yield an estimated $562,027 in revenue if all usage were billed at R1 pricing. In contrast, the cost of leasing the necessary GPUs for this level of service would have been approximately $87,072. These figures, detailed at the end of a longer GitHub post outlining its strategy for achieving higher throughput and lower latency, imply impressive profit margins when operating under optimal conditions.
However, DeepSeek acknowledged that actual revenue figures are “substantially lower” due to several factors. These include nighttime discounts, reduced pricing for its V3 model, and the fact that only a subset of its services is monetized—with its app and website remaining free to users. Critics note that while the calculations offer a glimpse into potential future profitability, they remain highly speculative given the various discounts and free offerings that currently impact overall revenue.
DeepSeek’s announcement comes amid broader industry debates over the cost structure and profit potential of AI services. The startup previously captured headlines in January when it unveiled a new model that, according to some benchmarks, matched the performance of OpenAI’s offerings despite being developed at a significantly lower cost and under the constraints of U.S. trade restrictions on powerful chips.
The company’s technology has also made an impact in the consumer space. At one point, the DeepSeek app briefly displaced OpenAI’s ChatGPT at the top of Apple’s App Store rankings before falling to a current position of #6 in the productivity category—positioned behind ChatGPT, Grok, and Google Gemini.
As the race for efficient and profitable AI technologies continues, DeepSeek’s bold profit margin claims underscore the high stakes and rapid innovation defining the industry. While the figures may point to an optimistic future, experts remain cautious, emphasizing that real-world performance and market dynamics will ultimately determine the financial viability of such models.
Read next
17:00
Nvidia
During the latest earnings call, Nvidia CEO Jensen Huang reiterated his bullish outlook for the company, dismissing concerns that DeepSeek’s R1 reasoning model might undercut Nvidia’s sales.
16:30
DeepSeek
Hangzhou-based AI startup DeepSeek, known for its groundbreaking large language models that have challenged Western systems at a fraction of the cost, is now making its way into the home appliance sector.
00:10
Chinese companies are rapidly adopting DeepSeek’s latest AI model to improve productivity and services after its impressive debut earlier this year. The artificial intelligence model is being integrated into tech, telecom, finance, healthcare, and energy sectors.
23:00
DeepSeek
Chinese AI startup DeepSeek announced on Wednesday a significant reduction in the cost of its API usage during off-peak hours. Between 1630 GMT and 0030 GMT, developers integrating DeepSeek’s AI models via its API can expect savings of up to 75%.
15:30
Nvidia
Santa Clara, CA, February 17, 2025 – Nvidia founder and CEO Jensen Huang dismissed market concerns over DeepSeek’s recent release of its R1 open source reasoning model, arguing that the technological advancement will ultimately benefit the broader AI ecosystem rather than harm Nvidia’s business.
What is your opinion on this topic?
Leave the first comment