It seems like DeepSeek, a Chinese AI startup, has made a significant claim regarding the cost-profit ratio of its V3 and R1 models. According to the company's disclosure, it is stating a theoretical cost-profit ratio of 545% per day. This suggests that, in an ideal scenario, for every unit of cost incurred, the revenue generated could be over five times that amount.
However, DeepSeek has also cautioned that the actual revenue generated will likely be much lower than this theoretical figure. This means the high percentage could be based on ideal or best-case conditions, and real-world results could vary due to factors like operational challenges, market conditions, or limitations in scalability.
The disclosure of such data could be a way for DeepSeek to highlight the potential profitability of its models, but it's also important to take this with a grain of caution as startups often present idealized projections to attract investment or attention.
If you're following the AI industry or investment opportunities, this type of claim is an interesting indicator of how companies in AI and tech are positioning themselves for growth, but you'd need to consider the practicalities and risks involved.