Hacker News
points by riknos314 4 days ago | 0 comments
Deepseek prices their models based on a 10 month hardware payback period. They report ~$38 revenue/GPU-hour on V4.1-Flash clusters.

Inference is clearly profitable. The question is whether it's profitable enough to justify the training research costs.

[–] scheme271 4 days ago
Are those statements audited or fact checked? Inference profitability is great but how much of that profit gets eaten up paying back training costs, R&D, and operating expenses.
[–] Gareth321 4 days ago
No, and it’s very rare for Chinese companies to allow independent audits. I would not trust DeepSeek financial disclosures. However from industry data we can see that inference is often profitable. It’s the training which is very expensive. Which is not a big problem for Chinese companies since to date much of their training is just distilling frontier models.
[–] snovv_crash 4 days ago
Their models and inference frameworks are open, you can run them yourselves to check how many tokens per day you can generate, and then validate that against their API prices.

Not sure how they can get more transparent than that.

[–] scheme271 3 days ago
My point was really about profitability in general. The inference portion might be profitable but if training and operating costs are very large, then that can make the whole thing unprofitable. E.g. if it costs you $100B to train a good model but you make $0.00001 per inference token, you may not be able to hit breakeven for a very long time.
[–] mike_hearn 4 days ago
That's conflating different things. Unless you believe you can instantly saturate any amount of GPUs you bring online, 24/7, with paying customers from day one, then theoretical max throughputs on a stack isn't the same thing as revenue generated. Customer acquisition costs are real!

Even if you could acquire infinity custom on day one for zero cost, revenue isn't profit.

[–] javier2 4 days ago
At best, this tells you v4.1 flash is profitable, but can they really get $38 per gpu hour? Sounds like a lot.
[–] riknos314 4 days ago
From their own inference framework docs: "A single cluster currently serves 155 trillion tokens per day, with overall revenue of US$38 per GPU-hour (at valley pricing)."

https://github.com/deepseek-ai/DeepInfer/blob/main/doc/infer...

So this implies they actually hit about double that during peak pricing.

[–] mustyoshi 4 days ago
If they keep it running at near constant usage, you'd only need 80-120 million tokens an hour to hit 38$ of revenue (at off peak pricing, half as much for peak pricing).
[–] margorczynski 4 days ago
V4.1-Flash is probably quite smaller and much more resource effective than the frontier models. Add to that much cheaper energy in China.
[–] calmworm 4 days ago
Clearly potentially profitable. Not clearly profitable.
[–] parineum 4 days ago
How much did it cost to train Deepseek v4.1 flash?
[–] riknos314 4 days ago
Great question. Nobody knows.

The cool thing about models designed for cheaper inference is that post-training (which often requires inference as a part of the feedback loop) becomes cheaper as well, so almost certainly cheaper than models that are less optimized.