
DeepSeek x 100
Open-source AI with state backing? The real threat isn’t power - it’s price. And that could squeeze big tech’s margins faster than their AI models can evolve.
Six months after the DeepSeek moment that rocked Wall Street’s AI trade, China’s Baidu ($BIDU) has another curveball for U.S. tech firms. Their Ernie model – a rival to ChatGPT and local competitor DeepSeek – is now open-source.
That complicates things for other players, since open-source models are all about lower cost and better performance. Massive AI spend for mega-cap tech firms becomes harder to justify: Amazon ($AMZN) alone has a projected US$100m spend on AI in FY25.
AI advisory firm Epic Loot’s Alec Strasmore put it best: ‘Baidu just threw a Molotov into the AI world.’ And Ernie? Baidu claims it matches DeepSeek-R1’s performance at half the cost.
Baidu, Alibaba ($BABA), and Tencent ($TCEHY) are leading members of China’s ‘National AI Team.’ It’s exactly what it sounds like: the government backs them with data and dollars in return for pushing the national agenda to make China the world’s biggest AI innovation centre by 2030.
UBS puts that potential at US$81bn in AI compute opportunity – up from just US$6b today. According to former People’s Bank of China deputy governor Zhu Min, China could see more than 100 DeepSeek-like breakthroughs in the next 18 months.
Competition isn’t necessarily a good thing for the firms involved. Bloomberg analysts warn of a Chinese AI price war. For context, China’s top 10 AI chatbots made just US$1m in revenue from Apple’s ($AAPL) iOS app store compared to ChatGPT’s US$669m.
But China’s rapid advances in AI are a big threat to the U.S., which blocked advanced chip exports to the Chinese market. That’s left Nvidia ($NVDA) with an US$8b-sized hole from the cancelled H20 chip orders to China. And CEO Jensen Huang believes China will move on with its AI ambitions even without Nvidia.
‘The U.S. has based its policy on the assumption that China cannot make AI chips… it’s clearly wrong,’ said Huang.
‘The question is not whether China will have AI. It already does.’
But the bigger question for investors is where does that leave big tech firms that are actively ramping up AI efforts in the U.S.?
Analysts have already priced AI into the valuations of Microsoft ($MSFT), Alphabet ($GOOGL), and Nvidia, but the game is moving faster than the markets might like. If high-end AI becomes commoditised, returns on multibillion-dollar AI bets might shrink.
Investors may need to recalibrate: the real winners in this next phase of the AI race might not be the ones building the smartest brains – but the ones who figure out how to sell the most picks and shovels.
This is not financial advice nor a recommendation to invest in the securities listed. The information presented is intended to be of a factual nature only. Past performance is not a reliable indicator of future performance. As always, do your own research and consider seeking financial, legal and taxation advice before investing.