Spending on AI infrastructure falls into two major categories: cloud providers and AI research labs. Neither of these areas will be affected by Deepseek.
AI infrastructure spending for cloud providers like Oracle, AWS, Microsoft will not be affected at all.
Major cloud providers like Oracle, AWS, and Microsoft will continue investing in AI infrastructure. Their role is similar to utility companies that provide power or internet—these services remain essential regardless of changes in technology. When training efficiency improves, it allows for the creation of even more complex AI models with advanced training methods.
Comparison to energy: Just as electricity providers continue to grow despite the rise of energy-efficient appliances, cloud providers will keep enhancing AI compute power to support growing demand across industries.
Comparison to internet: Faster web applications don’t reduce the need for better internet; instead, they drive the development of more advanced applications like virtual reality and real-time cloud gaming. Similarly, as AI computing power improves, industries will push for even more sophisticated AI models to tackle increasingly complex problems.
Whereas for AI research labs, Meta, OpenAI and Google’s AI Capex is not limited to AI chips only.
Duplicating an existing OpenAI o1 model is not the same as creating a new groundbreaking model, which requires a lot of time, effort, and money.
The AI race is still going strong. Models like OpenAI o1 and Deepseek R1 will be outdated in less than a year. The path towards AGI is not over, and it still requires a lot of cost and even more advanced AI chips to push the frontier. It would be interesting to see what AI models can achieve with the new Blackwell chips.
In reality, Deepseek has zero implications for any of these companies when it comes to justifying AI spending. Meta's advertising AI spending is a totally different category from Deepseek with totally different agendas and features.
These insights are easy for any hedge fund analysts to realize. Yet, the intentional dumping of NVDA stocks yesterday shows clear market manipulation by trying to create fear and chaos out of nothing. The big hedge funds know very well how this will play out and is just playing retail investors and others.
Modify on 2025-01-30 08:24
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
Now, who is going to still want to spend a lot more on AI chips if the time to recoup those cost now is 10x longer than expected. Every company boardroom who has planned for the large AI Capex this year is going to reconsider their planned investment thesis.