Almost A$1 trillion (US$600 billion) was erased from the value of artificial intelligence microchip maker Nvidia overnight on Monday, after a little-known Chinese startup, DeepSeek, threatened to shake up the US tech market.
As Nvidia saw the biggest one-day loss in sharemarket history, other tech giants - Microsoft, Alphabet and Amazon, which are also ploughing huge amounts of money into competing AI tools, including ChatGPT and Gemini - suffered as well.
The rout was triggered by the investors' shock at the claimed performance of DeepSeek's new R1 chatbot. The Chinese AI was said to be more advanced than others and cheaper in terms of development cost.
DeepSeek R1 took off, topping the free downloaded app in Apple's app store, as US technology and related stock prices plunged.
The market was surprised by DeepSeek providing what amounts to cheaper technology but comparable performance.
This has dramatically changed the market's expectations of computing power, showing more can be done for less. It has also compromised the competitiveness of the US tech companies' existing AI products and developments.
Market expectations drive stock prices. The announced performance of DeepSeek R1 triggered a major revision of expectations about what was technologically possible and about how cheaply AI could be developed and operated.
Investors have quickly incorporated the news of this low-cost Chinese AI competitor into stock prices, expecting this new entrant to threaten the market and undercut the competitive edge of existing leaders.
DeepSeek was founded in 2023 by Chinese hedge fund High Flyer, which had exclusively used AI in trading since 2021.
DeepSeek develops large language models (LLMs) that can underpin chatbots and other AI-based tools. R1 is the latest iteration of DeepSeek’s chatbot and underlying model. It builds on earlier versions of generative AI models developed by DeepSeek, and considerable amounts of data, but is a surprising leap forward in performance and cost.
Technology investors believe R1 matches or outperforms competitors, including OpenAI’s ChatGPT 4.o1 on numerous benchmarks.
The model underlying R1 works much less intensively. It's much cheaper to develop and operate: it demands less data and computing power.
The training of the model was possible despite the US export ban preventing Chinese companies such as DeepSeek from accessing chips from US companies such as Nvidia.
The Biden administration had introduced laws restricting the sale of certain computer chips and machinery to China, in a move intended to block its rival from accessing some of the world’s most advanced technology.
The training data and data uploaded to R1 sit on servers in China. With concerns about data privacy and intellectual property already raised against US-based companies, it is arguably even more concerning to have data under the jurisdiction of the Chinese Communist Party (CCP).
Besides, the chatbot program code is free to download, read and modify, unlike ChatGPT. That is, however, somewhat false transparency – it is the underlying model that really matters, not the Chatbot code.
R1 is known to censor its responses in consonance with Chinese Communist Party values.
It is yet unknown whether it is an irrational panic that is bound to be reversed or the simple reflection of correct pricing for the future costs and benefits of AI are unknown.
This question is both technical and economic.
In other words, there are as of now no assurances if R1 is to actually require much less computing power and much lesser data to be trained and used in.
Economically, there are winners and losers. The user of AI is likely to benefit from cheaper access to AI, and in particular, to LLMs, which will increase adoption and associated productivity benefits. The producers who exist, like Nvidia, lose out in what was a market with few real competitors.
More generally, a less computationally expensive and thus more energy-friendly AI can only benefit society.
On the other hand, a country capturing a monopoly of this market together with potential risks in regard to data protection, intellectual property rights, as well as censorship risks could weigh in on the downside.
All have their own impact on business, but there is an influence on them.
This is why I co-write a weekly business and economy email newsletter. It brings a curated summary of the week's briefings from academic researchers around the world straight to your inbox. And it's free.
Read also| India's Semiconductor Market Expected to Reach $103.4 Billion by 2030
Read also| India a Top Priority: Samsung’s TM Roh Pledges Increased Investment in R&D Centers