A major misunderstanding
A chatbot similar to ChatGPT was recently developed by a Chinese hedge fund and released openly to the public. They claimed that it cost about $5.6M to produce, dramatically lower than the cost of similarly capable models (OpenAI’s GPT-4 reportedly cost $60M to $80M).
The market quickly ran with one simple implication—if AI models are dramatically cheaper to produce, then there will be much less demand for AI hardware (Nvidia was down 17% in one day on that premise).
The claim that it only cost $5.6M is misleading at best. That figure is missing several important costs like prior research efforts, and the cost of buying all the hardware in the first place (which may cost $1B or more). Even worse is the snap conclusion that cheaper AI means less demand for AI computing hardware. Nothing could be further from the truth.
The cost to produce a given AI model has been collapsing dramatically for years. Even if DeepSeek’s cost claims were true, it would be just another example of an ongoing and continuing innovation cycle. Rather than reducing the need for AI hardware, more efficient models drive further adoption, pushing innovation even faster.
Cheaper AI models will mean they get used more, and this drives further demand for AI compute capacity. Just as cheaper electricity causes people to use more electricity, cheaper costs for computing will drive new use cases that were previously uneconomic.
The world is getting increasingly tech-y. We all see it. Understanding technology requires a lot of background knowledge, and it’s important to know the lessons of its short history. We’re sympathetic to those who feel like they’re struggling to understand. Technology is both complicated and moving fast.
So let’s simplify:
The demand for intelligence is limitless. What will be the demand for the compute that drives artificial intelligence?
Best regards,
Evan McGoff
P.S. If you would like to dive deeper into DeepSeek and what happened earlier this week, here is an article from Tae Kim, the author of The Nvidia Way.
Disclosure: Dock Street Asset Management, Inc. and our clients own Nvidia (NVDA). This article is not intended to be used as investment advice.
Dock Street Asset Management, Inc. is an investment adviser registered with the U.S. Securities and Exchange Commission. You should not assume that any discussion or information contained in this letter serves as the receipt of, or as a substitute for, personalized investment advice from Dock Street Asset Management, Inc.
It is published solely for informational purposes and is not to be construed as a solicitation nor does it constitute advice, investment or otherwise.
To the extent that a reader has questions regarding the applicability of any specific issue discussed above to their individual situation, they are encouraged to consult with the professional advisor of their choosing.
A copy of our Form ADV Part II regarding our advisory services and fees is available upon request.
Our comments are an expression of opinion. While we believe our statements to be true, they always depend on the reliability of our own credible sources. Past performance is no guarantee of future returns.
