January 2025 Market Brief – Game Changer?

January 2025 Market Brief

Gamechanger?

“A computer would deserve to be called intelligent if it could deceive a human into believing that it was human.”- Alan Turing

Markets are reacting sharply to a new potential threat: what if the USA is not the leader in AI technology and others have figured out how to do it better and more efficiently? That potential game-changing question was brought up by DeepSeek, a Chinese AI app that was recently launched in the US. DeepSeek is majority owned by Lian Wenfeng’s quant AI focused Hedge Fund, High-Flyer. They announced the development of DeepSeek in March 2023 and by all measures have exceeded expectations. This app was developed as a response to OpenAI’s ChatGPT, which launched in late 2022 and caused a large scramble to keep up by Chinese tech firms. DeepSeek is like ChatGPT but with a few noticeable differentiators. DeepSeek develops open-sourced free AI models, which means multiple developers can inspect and improve on the software. Their first opensource AI model R1 competes with ChatGPT in areas such as reasoning and code generation and has become the #1 download in the iPhone app store as an AI assistant App. In a very short time (days!) they have overtaken ChatGPT as the top-rated free app. China is flexing its AI muscles this week, and US tech companies seem to be on the back foot.

The big gamechanger? DeepSeek was developed on about 1/10^th of the budget and a much smaller and less powerful data center. The quality of the app, and cost efficiency of developing it, have shaken US tech companies; perhaps we are not the leader in the space, and we are not working as efficiently as the Chinese who are developing comparable, or better, AI technology even without some of the high performing chips which we have export bans on. Bloomberg’s Camero Crise summarized it well, comparing DeepSeek to a generic drug, undercutting a pharma company that has spent tens of billions on R&D, just to have a generic that does the same thing for cheaper. Only this time, there are no patents to protect US tech companies, and technology is moving at the speed of light. A new technology utilizing substantially lower processing power and developed & offered at a lower cost is a risk to US AI and technology stocks. This also means there may be lower barriers to entry into the AI fields which may challenge future revenue growth in the fast-changing tech space.

Much of the US stock market’s gains have been driven by large US technology companies, which are trading at stretched valuations and dependent on continued growth of the AI sector. Much of this multiple expansion has been justified by the anticipated exponential expansion of US data centers, power needs, chips, and advanced software. But if China has figured out how to do more with less, partially driven by necessity due to chip curbs, this is seen as a true threat to US technology companies’ growth trajectory. Did they really figure out how to build a better mousetrap? If so, it would make META’s announcement of a planned data center the size of Manhattan look obsolete and outdated. We have seen a sharp reaction in the AI trade, with reverberations beyond just chip companies such as NVDA and AVGO, but power and infrastructure companies got hit as well. What if the anticipated power load and large data center builds are no longer necessary due to better utilization and model building? $500B Stargate could be a dinosaur and outdated by the time ground breaks. The stock market is, and will continue to, move faster than infrastructure and capex – but could fast moving technology render all the planned expansion in data centers, chips and power obsolete? That is what gripped the markets on Monday. There are potential parallels with the tech bubble, as firms were still increasing spending on IP and infrastructure/equipment even after the Nasdaq bubble burst.

For now, we view this as a speedbump, however, it fits with our thesis of “Bullish… with a but.” There is potential this is paradigm shifting technology that changes the growth outlook for multiple US companies as China might have just figured out how to build AI like the Model T Ford – completely changing an industry. But it could also just be a lower cost version that is complementary and not destructive to the rest of the tech space. There are additional questions of exactly what DeepSeek cost to build, as its murky on how many GPUs were used, and which chips were used. It is also, like all Chinese chatbots, censored. MSFT CEO responded saying, “As AI gets more efficient and accessible, we will see its use skyrocket, turning it into a commodity we just can’t get enough of.” AI certainly is not going anywhere, but the profitability and industries supporting it may continue to shift over time. It is up to US technology firms to respond and continue to develop industry leading models, chatbots, and apps. The AI war between the US and China certainly ratcheted up a notch this week, and US companies must respond. A challenge has been issued, and for now, US firms look a bit clunky vs. the newly rolled out Chinese tech. But time will tell if this was a true pivot in the AI story or simply a bump along the way.

I am not convinced the bubble has burst, but one would be silly not to evaluate the potential risks this update brings. We are looking carefully at how the market progresses from here and if action needs to be taken to protect and shift portfolio allocations.

Follow this link to see the full library of Market Briefs.