Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Which is the better investment: a big, established tech group at a relatively modest valuation or an unproven fledgling at hundreds of times forecast revenue? In China’s AI sector, many investors are plumping for the latter.
Shares in Chinese AI start-up Zhipu, listed as Knowledge Atlas Technology in Hong Kong, have more than quadrupled this year to give it a market value of almost HK$250bn, or over $30bn. Those in Minimax Group have more than doubled. Meanwhile, Alibaba and Tencent, local tech giants that are reporting record engagement across their AI platforms, are lower than where they started this year.
That looks counterintuitive. It is true that AI start-ups are a pure-play bet on potentially groundbreaking technologies: Zhipu recently released its new flagship model GLM-5, which it says approaches Anthropic’s Claude Opus 4.5 on some capabilities. And it is also true that the big established platforms’ heavy investment in AI infrastructure could pressure current margins before producing returns.
But they still have their advantages. Alibaba and Tencent, which have embedded large language models across their cloud, advertising, commerce and social platforms, already reflect slower growth in core businesses. Tencent, for example, trades at 15 times forward earnings, more than a third lower than its level three years ago before the current AI wave, providing a buffer against disappointment in earnings or slower than expected AI monetisation.
Meanwhile, both companies remain deeply embedded in China’s digital economy. Alibaba’s vast ecommerce and cloud networks generate recurring revenue and steady cash flow. Baidu’s search business has long funded its investments in AI. With or without AI delivering an immediate profit boost, their core franchises are a cushion that newer AI entrants lack.
That means even modest gains in ad pricing, transaction volumes or cloud usage can translate into meaningful profit growth when applied across Alibaba’s or Tencent’s scale. The newer AI entrants, however, still need to prove that their technology can translate into real returns. Their valuations assume rapid customer acquisition and continued access to capital to fund model development and computing costs.
Does this mean the giants will lose ground to these challengers? It’s more likely that investors are increasingly willing to take risks on potentially groundbreaking technologies, even where profitability is distant. Scarcity is amplifying the effect. Most of the best-known global AI developers are privately held, so the few publicly traded Chinese AI model developers have become rare vehicles for public-market investors to make that bet.
Read the full article here