Embracing Disruption
DeepSeek R1 – what just happened?
A private Chinese company, DeepSeek, recently launched a new “large language model” (LLM) – DeepSeek R1 – that resulted in a significant negative reaction to stocks that have previously been seen as artificial intelligence (AI) winners. The reaction can largely be attributed to the view that DeepSeek R1 is able to deliver best of breed results, but at a much lower price point thanks to using fewer and cheaper semiconductors in the training process.
The huge attention that DeepSeek has garnered over recent days is of no surprise, given the current fever pitch level of interest in AI; however, in terms of identifying the ultimate winners of these recent developments, we are mindful of extrapolating from limited data points and claims. Generative AI remains in its infancy, both in terms of its development and deployment, and it is quite normal for nascent technologies that are benefiting from huge investments to show continuous and impressive improvements. Indeed, this ties in with a key part of our investment philosophy: at the early stages of new technology development, it is normal to see standards changing or performance improving persistently – we must thus be very selective about seeking exposure to a new technology during a period when it is more challenging to identify the long-term winners.
While we are certainly going to continue seeing many claims of having “the best” AI or LLM model currently available for the foreseeable future, the developments around DeepSeek R1 are notable in that they are directly relevant to questions surrounding the level of capital expenditure required in the coming years. While these issues have led to the recent short-term volatility seen in some share prices, our view remains that AI is a long-term structural megatrend. It is also worth noting that this news flow has come at a time when many Asian markets are already closed or about to close for their Lunar New Year celebrations, and many Western companies are in their quiet period ahead of Q4 reports. These firms have thus been unable to provide their immediate opinion on the latest developments in the AI market.
Overall, the arrival of DeepSeek R1 reinforces our view that, when looking at the technology sector, we want to focus on companies where we can take a longterm view. It also emphasises the point that multiple tech ecosystems are emerging, and investors should thus maintain a portfolio of well-diversified opportunities.
