The irruption of Chinese newcomer Deepseek into the world of artificial intelligence (AI) software came as a shock to financial markets, exerting brief but intense downward pressure on US technology megacaps. To us, however, the bigger surprise was that investors had not anticipated such a development. Or, put differently, that China’s technological capabilities are still so widely underestimated. The decision to make Deepseek “open source” (meaning that the software code is publicly available) will certainly be a thorn in the side of President Trump and his increasingly powerful Silicon Valley “friends”. Not to mention that Chinese electric car manufacturer BYD is soon to make its “hands-free” driving software available at no extra cost in its models, a move unlikely to please Tesla…
Our reluctance to invest in AI-related stocks because of their sky-high valuations seems rather justified in light of these recent Chinese initiatives. We had no concerns regarding those companies’ qualities and capabilities, but they were being hailed by the market as winners of the entire future AI pie – and more. Forgetting that someone else, somewhere else, may one day challenge their dominance.
That day has now arrived, and yet some observers continue to minimise the threat, arguing that the true cost of Deepseek’s development well exceeds the touted USD 6 million, that China managed to secure a supply of high-end AI chips just before the US export ban, or that no Western user will want to entrust data to a Chinese AI model.
However, Deepseek’s “open source” nature will make it possible for countless companies and start-ups (including Western and Indian ones) to build “apps” upon the code (probably initially without having to pay for a licence) and also develop the code further. Which will eventually make the whole endeavour lose its Chinese character somewhat – and become a formidable competitor to US AI companies. No wonder that a consortium around Elon Musk, clearly recognising the danger, immediately made a bid of no less than almost USD 100 billion for ChatGPT’s code (OpenAI), with the intention of making it “open source” (again) and thus competing with Deepseek in the future. Needless to say, law firms are in the meantime exploring how to protect the existing property rights of their AI clients and on what grounds, if any, the use of Deepseek can be banned. Stay tuned.
For Europe, this evolution towards “open source” AI might be an opportunity to jump back onto the bandwagon. Although our continent has a limited presence in this market today, the roots of AI lie in the Benelux. Indeed, speech technology and natural language processing, the precursors of AI’s large language models (LLM), were developed at the end of last century by Philips, the University of Ghent and Lernout & Hauspie, among others, before later ending up in US hands.
The January correction in equity markets and the big US technology names did not last. As is often the case, the market reacted strongly to the news release, but then soon moved back to its usual (optimist) stance. In our view, however, the true impact has yet to come – for two reasons.
First, Deepseek is but another warning sign that China has become a technology superpower, on the back of massive government investment and a huge pool of science and technology graduates. We already knew that Chinese companies can make superior smartphones, as well as electric vehicles that are both cheaper and more efficient than current Western models. Now comes Deepseek in AI, and odds are that China will soon venture into other fast-growing technology segments, such as advanced cloud computing.
Second, making Deepseek open source was most likely a deliberate tactic by Chinese authorities. No company management is inclined to pursue such a business model of its own accord, given the difficulties it involves in generating revenues. Think only of US pioneer OpenAI which, after a few non-profit years (and then with limited revenue), eventually partnered with Microsoft (and is now coveted by Elon Musk).
In effect, China seems intent on undermining the position of the US high-tech oligopolists – at precisely the moment when they have all chosen to align with the new Trump regime. How great the danger to their current business model remains of course an open question, but at least a significant chunk of their potential market is at risk.
Speaking of the new US President, the sheer volume of decrees (and announcements) issued since his inauguration has been overwhelming. It speaks of thorough and thoughtful preparation, but also an intention to create such a chaotic environment that opposition finds itself dizzied. Moreover, the Trump administration is abusing the National Security Act to circumvent the Congress’ checks and balances on domestic policy. Which leaves only the legal system, with its slow operation, to block the implementation of the many (often illegal) decrees.
On the international front, President Trump has more leeway and trade tariffs are a good example of his “America First” tactics. He announced major hikes (25%) on imports from Canada and Mexico, only to then delay their implementation by one month (so as to negotiate all sorts of matters with these two countries in the interval). He also threatened the European Union with import tariffs but has taken little action (so far), and he pushed through a 10% additional tariff on goods coming from China.
Pointing to the risks – for all parties – of a global trade war, the Chinese authorities have so far remained rather restrained in their retaliatory measures, the most notable being export controls on 25 rare metals needed to make semiconductors and batteries. A means, like Deepseek, to hurt the US technology titans?
In such a quasi-unpredictable economic and geopolitical market environment, maximum diversification remains the key word for investors. Within the equity space in particular, there is certainly no obvious choice today. The US market is, for the time being, where most earnings growth can be expected, but valuations are very rich and there is no longer any premium for risk. European equities are less expensive and investor positioning remains bearish, but their risk premium is low too and the biggest bargains are to be found in the small- to mid-cap segment, i.e. precisely those companies that would be most hurt by an escalating trade war (moving production from one region to another is much more difficult for smaller businesses than for multinationals with globally spread production facilities). Note, however, that investor flows into European equity markets were positive in January, after months in negative territory: might a change be underway? As for Chinese technology companies, they do seem to present opportunities, even after their strong year-to-date rebound (a form of catch-up after many years of underperformance). They are still trading at “only” 19.5x projected 2025 earnings, well below the 43.5x multiple of their US counterparts. But then, of course, one needs to add the China risk factor.
