Monday 3rd of February 2025
For the past 40 years, one of the biggest trends in finance has been the rise of “winner-take-all” companies in the technology sector. These are businesses that dominate their industries, making it extremely difficult for competitors to catch up. Think of Microsoft in PC software, Google (Alphabet) in internet search, Amazon in online shopping, Facebook (Meta) in social media, and Apple in high-end consumer gadgets. These companies have benefited from what’s called “network effects”, the idea that the more people use a service, the more valuable it becomes. Because of this, most of the profits in these industries have gone to just one or two major players.
For investors, the key to making money over the past few decades has been betting on these tech giants early. Given this history, many people assumed that artificial intelligence (AI), the next big technology revolution, would follow the same pattern, with one or two companies taking all the winnings.
However, something happened recently that shook up this assumption. It wasn’t a market crash, nor a sudden panic, it was simply the realisation that AI might not be a “winner-take-all” game after all.
The best way to explain this is by comparing AI to past technological revolutions like the invention of cars and airplanes. While these innovations created big, successful companies (like Ford and Boeing), they also led to lots of competition. This meant that while some companies did well, the real winners were consumers, who got access to better products at lower prices.
A Chinese AI company called DeepSeek recently launched an AI model named R1. What makes R1 interesting is that it performs at the same level as models from top companies like OpenAI (the creators of ChatGPT), Anthropic, and Meta, but it was built at a much lower cost. Even more surprising, DeepSeek used older, less powerful computer chips rather than the cutting-edge chips many experts thought were necessary for advanced AI.
Until now, Nvidia has been the go-to company for AI chips, and its most advanced processors have been seen as essential for training top AI models. But DeepSeek showed that good AI models can be built using older, less expensive chips. This opens the door for other companies to enter the AI race without relying on Nvidia’s best technology.
The dominant belief has been that training AI requires massive amounts of data and computing power, which only big companies like Microsoft, Google, and Amazon can afford. DeepSeek’s success suggests that smaller companies might be able to compete with much lower costs.
And up until now, if a company wanted to use AI in their products, they usually had to rely on the big AI providers. But with models like R1, businesses may soon be able to run AI on their own servers, reducing their dependence on tech giants like OpenAI and Google.
These shifts could have big consequences for both technology companies and investors.
For one, Nvidia, the company that makes the best AI chips, saw its stock price fall nearly 17% in a single day. The reason? Investors are starting to believe that Nvidia’s dominance in AI may not be as secure as once thought.
But it wasn’t just Nvidia that was affected. Companies that provide infrastructure for AI data centres (the huge computing hubs that power AI models) also took a hit. This includes utility companies like Constellation, Vistra, and NRG, as well as suppliers like GE Vernova, Eaton, and Quanta Services, which build power systems for data centres.
Interestingly, the biggest tech companies: Amazon, Google, Microsoft, Meta, and Apple, weren’t hit as hard. While they have invested heavily in AI data centres, they may now be able to spend less on infrastructure if AI models become more efficient. In fact, companies like Apple and Microsoft, which specialise in software and applications rather than AI infrastructure, might even benefit from this shift.
It’s too soon to say that big tech companies are losing control of AI. Nvidia, for example, still has a major advantage because its software ecosystem (called CUDA) is the industry standard. And while DeepSeek’s AI model is more efficient, it could still benefit from Nvidia’s high-powered chips.
Even after its steep drop, Nvidia’s stock is still trading at levels that were considered extremely high just a few months ago.
More broadly, while the market’s reaction to R1 was significant, it wasn’t a disaster. The Nasdaq stock index fell about 3% in a day, a bad day, but not a full-blown panic. The idea that AI will be a “winner-take-all” market has been shaken, but not completely disproven.
What does all of this mean for everyday investors and technology watchers? The key lesson is that AI’s future might not be dominated by a single company or a small group of tech giants. Instead, we may see more competition, lower costs, and broader access to AI technology, much like what happened with cars and airplanes in the past.
For consumers and businesses, this is great news. More competition means faster innovation, better AI tools, and lower prices. For investors, however, it’s a reminder that even the biggest trends in technology don’t always play out the way everyone expects.
We would like to thank Dominion Capital Strategies for writing this content and sharing it with us.
Sources: Bloomberg, Yahoo Finance, Marketwatch, MSCI.
Copyright © 2023 Dominion Capital Strategies, All rights reserved.
Disclaimer: The views expressed in this article are those of the author at the date of publication and not necessarily those of Dominion Capital Strategies Limited or its related companies. The content of this article is not intended as investment advice and will not be updated after publication. Images, video, quotations from literature and any such material which may be subject to copyright is reproduced in whole or in part in this article on the basis of Fair use as applied to news reporting and journalistic comment on events.