DeepSeek’s emergence from stealth mode on US President Donald Trump’s inauguration day has sent shockwaves through the AI sector. The revelation that the China-based platform is a formidable rival to the once-dominant US AI giants wiped out around US$1 trillion ($1.36 trillion) in market value.
However, fund managers and analysts do not suggest that DeepSeek is going to be a true gamechanger and that investors should abandon their AI bets. Manuel Villegas, digital assets analyst at Julius Baer, is maintaining his “constructive” view on cloud computing and AI for now. “DeepSeek’s impact has yet to be tested on congested networks, and social media’s doomsday scenarios on hardware seem far-fetched,” he says.
The declines suffered by the broader technology sector — which, besides Nvidia’s by now infamous US$600 billion hit — the single largest one-day market value slump in history — is shared to a smaller extent by other big tech names, including Microsoft, Alphabet and ASML.
“This seems to be more like profit-taking and a bit of overreacting, as there is still quite a bit of uncertainty as to the overall impact of DeepSeek whether it would be good or not for overall capex spend,” says Ken Wong, Asian equity portfolio specialist at Eastspring Investments Hong Kong, in reference to the dramatic market reaction.
Nonetheless, Eastspring notes that the global AI high-performance computing market, including AI servers, smartphones, smart cars and PCs, is projected to grow at a CAGR of 55% between 2023 and 2026 and reach US$200 billion. Eastspring’s checks of 30 Asian companies suggest that AI chip demand is set to increase as companies increasingly monetise their AI usage.
“Asian tech companies offer investors exposure to the AI theme but at cheaper valuations. This includes companies within the semiconductor upstream supply chain: raw material suppliers, chip vendors, foundries, and outsourced semiconductor assembly and test (OSAT) players. It also includes downstream component and original design manufacturing/original equipment manufacturing (ODM/OEM) vendors, as well as semiconductor production equipment (SPE) suppliers,” says Eastspring in its 2025 Market Outlook.
See also: India remains a compelling investment story despite near-term drop, says aberdeen’s Thom
DeepSeek has shown that AI model training does not require the most advanced Nvidia GPUs and a “brute-force approach”, which has been driving capital expenditure from hyperscalers in the US seeking to build a competitive advantage, says Jamie Mills O’Brien, investment director of equities at abrdn. “It is too early to say what the exact fallout of the development will be, but there is a risk these companies will now need to re-evaluate the justifications for any future spending and how best to compete,” he adds.
AI tools will largely remain solutions in search of a problem. The industry has mainly focused on expanding GPU clusters to train increasingly powerful models. However, DeepSeek has introduced an alternative approach to developing complex models, significantly reducing the need for extensive initial processing.
“This could enable greater competition by allowing more participants to develop custom AI models, potentially accelerating the commoditisation of AI as market participants aim to compete through product differentiation and specialisation,” says O’Brien. “Despite the initial de-rating of many tech stocks across the AI semi supply chain, demand could remain robust, and we remain relatively bullish on its prospects.”
See also: In times of volatility and uncertainty, stay diversified: analysts
Healthy ecosystem
The growing democratisation of AI is driving greater adoption among end-users, sustaining demand for increased processing power throughout the operational phase of the AI model lifecycle. This ongoing need for processing power supports a generally positive outlook for players in the semiconductor, data centre and utility sectors.
“Our thesis, after the release of DeepSeek, is that we were going to see instances of sleeker, more efficient AI models that would not rely on massive clusters of AI GPUs and related hardware,” says Eric Compton, CFA, director, Morningstar. “This was the only way the ecosystem was going to successfully address large numbers of use cases in the long term.”
Compton believes that lower costs — making AI cheaper and, therefore, more economical — will increase the number of use cases it is viable for, thereby lifting demand instead. He compares this to the PC revolution decades ago and sees AI going down a similar path. Computing power became cheap enough that millions of individuals could use it at an affordable cost. He noted that a similar thing happened with the cloud and software-as-a-service (SaaS) revolutions thereafter, where the incremental cost of adding users was close to zero.
“We believe a future where AI is both prohibitively expensive and also ‘taking over the world’ is not likely. As such, we view the advancements made by DeepSeek as promising and healthy for the overall ecosystem,” says Compton, who notes that AI affects several companies across much of the firm’s coverage, including semiconductors, cloud infrastructure, software, utilities and energy.
Many of the firms under Morningstar’s coverage with an “AI premium” were already trading in the one- to two-star territory within a five-star scale. “Our valuations were already positioned for a pullback of this nature, as we were having a hard time justifying the increases in revenue implied by these valuations. We view the current pullback as healthy, even as we remain positive on the long-term potential of AI. We have maintained our fair value estimates across the affected companies,” says Compton.
On the other hand, Vishwanath Tirupattur, strategist at Morgan Stanley, thinks that it is unlikely that the DeepSeek development will meaningfully reduce capex related to AI infrastructure. From a macroeconomic perspective, there is a good case to be made for higher business spending as well as productivity growth. “Obviously, it is still early days and we will see leaders and laggards at the stock level. But the economy as a whole will emerge as a winner, in our view. DeepSeek illustrates the potential for efficiency gains, which in turn foster greater competition and drive wider adoption of AI. We remain constructive on AI’s transformational promise,” he adds.
To stay ahead of Singapore and the region’s corporate and economic trends, click here for Latest Section
US AI agenda
In a Jan 27 note, a team of analysts led by Ed Mills at Raymond James believe that in the near term, DeepSeek’s achievement is likely to pressure the US into increased support for domestic AI development at the federal level. A re-evaluation of current AI development strategies is likely, emphasising not just size and scale of investment, but also efficiency and innovation in AI architectures.
The short-term limitations of the current US export control regime were clearly demonstrated by the release of DeepSeek — opening new questions about the next steps for US tech restrictions. “While we expect to see a narrative emerge that DeepSeek proves the ineffectiveness of current export controls, it will most likely be interpreted by the Trump administration as a reason to tighten controls and further limit and track who has access to leading-edge technology,” say Mills and his team.
“We have long maintained that the new administration is likely to ramp up the aggressiveness of existing tech export controls, and DeepSeek provides a clear catalyst for an acceleration in both scope and enforcement, especially given that the new administration is already conducting a review of existing export control policies with an April 1 deadline,” adds the Raymond James team.
“This announcement also challenges the assumption that the AI diffusion rules, released in the last week of the Biden administration, establishing country-level chip restrictions will be rescinded or fundamentally altered by the Trump administration. The DeepSeek announcement will make it politically difficult to loosen chip restrictions.”
Morningstar’s equity analyst Phelix Lee believes that there is low imminent risk for DeepSeek to be caught in the political crossfire. DeepSeek has recently restricted new registrations to Chinese numbers so that will limit the number of US users. It has not released any commercial versions or monetisation plans, so there is no direct monetary impact even if new restrictions are rolled out, as funding thus far supposedly comes from founder Liang Wenfeng’s hedge fund. “Since DeepSeek is open-source, the most probable scenario is US tech corporations leveraging the public codes to refine their own models, resulting in lower AI cloud and computation costs globally,” says Lee.
DeepSeek shows it is possible to use less powerful chips to create similarly powerful models, but what does this mean for Asian players such as China’s Semiconductor Manufacturing International Corp (SMIC) and Shanghai Huahong Grace Semiconductor Manufacturing Corp?
Morningstar’s Lee believes that these two companies now face less urgency to catch up and produce cutting-edge AI chips, and hence less enthusiastic government support. Specifically, SMIC has enjoyed government subsidies and its output used in domestic appliances, which do not require the latest cutting-edge chips. “The DeepSeek news is the perfect time for some bulls to take profit,” says Lee.
