Time and again, the arrival of new technology is underestimated, while the latest technological advancements are adopted at an ever-increasing pace. This makes investing in new technology attractive, as both its profit potential and profit growth are often undervalued.
Due to this underestimation, technology stocks are often relatively inexpensive. This is likely one of the reasons why many tech stocks belong to the small group of shares responsible for the vast majority of returns in the stock market. For this reason, investors are constantly searching for the next Microsoft, the next Apple, or the next Google.
Last week, there was a significant acceleration in the development of artificial intelligence. The market was caught off guard by a Chinese start-up capable of training AI models at a fraction of the cost of comparable American models. While tech experts may have been aware that such technology was on the horizon, the real surprise lay in its substantially lower costs. Additionally, DeepSeek’s prominence was boosted by the fact that many Chinese officials were notably active in the app store, and the launch, which closely followed the announcement of the Stargate project, was perfectly timed. This has strengthened the narrative that the AI race has well and truly begun.
Ultimately, DeepSeek has been a revelation, demonstrating—to use the words of venture capitalist Marc Andreessen—that the world is experiencing a new “Sputnik moment.” AI is evolving at an extraordinary rate, becoming cheaper and more accessible than many had previously imagined. This is likely to trigger an investment race between the US and China, further accelerating advancements. Increased competition will also drive down costs—good news for both consumers and the broader economy.
As a result, we will likely reap the benefits of AI sooner than anticipated. Currently, AI adoption remains strongest in the information sector (publishing, data processing, and so on), but these industries account for less than 2 percent of total employment. Other sectors must find productive applications for generative AI technology to achieve significant macroeconomic impact. Thanks to DeepSeek, it is now far more plausible that more businesses will be able to develop and implement such applications faster than previously expected.
Companies with large workforces are particularly interesting in this regard. Routine and repetitive tasks, in particular, lend themselves well to automation via AI. Furthermore, major industries facing structural labour shortages, such as healthcare and education, also stand to benefit significantly.
Productivity—measured as output per hour worked—is the ultimate driver of economic growth. It enables higher wages while simultaneously reducing per-unit production costs, all without fuelling inflation. Productivity growth has historically depended on new technologies, as seen in the 1920s, the post-World War II era, and the 1990s.
A sustained period of high productivity growth could also ensure that nominal economic growth consistently outpaces nominal interest rates. This, in turn, could reduce debt levels as a percentage of GDP. In short, if we can capitalise on what Einstein referred to as the “eighth wonder of the world” for long enough, AI could offer a solution to debt-related challenges.
A pivotal moment for any new technology comes when it reaches mass adoption. When the first personal computer was introduced in 1968, its inflation-adjusted cost was around 44,000 dollar. It was not until the 1980s and 1990s that costs fell sufficiently to allow widespread adoption. The emergence of the internet in the 1990s further accelerated this trend. Thanks to network effects, the PC became increasingly valuable simply because more users adopted it. The iPhone saw even faster adoption, and when ChatGPT was launched, it shattered all previous records almost immediately.
The advent of DeepSeek means that significantly more businesses and individuals will now embrace AI. Lower costs naturally lead to higher demand. Moreover, concerns about AI’s extreme energy consumption appear to be subsiding.
However, the key question remains: will the major tech companies continue to be the primary beneficiaries? These firms have historically enjoyed a monopoly due to their control over vast amounts of cloud-based data, coupled with deep pockets to finance the most expensive chips and data centres. While demand for these advanced chips still outstrips supply, the once-unassailable monopoly or oligopoly position of these firms is beginning to wane. Consequently, the premium attached to their valuations is likely to erode over time.
Fortunately, the AI landscape extends far beyond ChatGPT and DeepSeek, which focus solely on generative AI. While these AI models can perform many human-like cognitive tasks, there are numerous other forms of artificial intelligence. Applied AI (narrow AI) has long been in use for speech recognition, image recognition, and recommendation systems on platforms like Netflix, Spotify, and Amazon. Google Maps also relies on this type of AI.
Parallel to generative AI, machine learning continues to advance, playing a crucial role in spam filters, pattern recognition in large datasets, and, of course, autonomous vehicles and self-navigating drones. Future innovations will see industrial robots replacing not only human muscle power but also cognitive abilities. Virtually all devices will evolve into smart systems, including medical diagnostic tools and financial risk analysis solutions.
The programming language of the future will increasingly resemble natural human language. Soon, AI will not only predict the weather but also forecast product demand and determine optimal maintenance schedules. For AI, complexity is not a drawback but an advantage. Many complex optimisation problems require vast computational power, and the rapid progress driven by DeepSeek brings such capabilities closer to reality. This means that the search for the next Microsoft, Google, or Apple can proceed at an even faster pace.
Han Dieperink is Chief Investment Officer at Auréus Vermogensbeheer. Previously, he held the role of Chief Investment Officer at Rabobank and Schretlen & Co.