China Tech – Local Syndrome

The outlook for AI in China is completely unknown.

• Poor results from Alibaba and Tencent are an indication that AI alone is not going to turn the Chinese technology sector around, meaning that a recovery in domestic consumption remains the most important factor in any Chinese technology renaissance.
• Furthermore, how these companies are going to access the compute capacity to support these services is unknown, meaning that they have no real idea how they are going to spend the money that they have allocated to investing in AI.
• Tencent reported results that saw 13% growth and met expectations, but its investment plans mean lower margins and distributions to shareholders in the coming year or two.
• At the same time, Alibaba’s profit missed expectations as it is having to spend hard on promotions to fend off competition as domestic consumption has remained stagnant.
• Both companies signalled heavy investing to capitalise on the potential of a new wave of AI services being made available to Chinese users, but both were very vague on how the investments would earn a return for them.
• This is unfortunate, especially as China is arguably one of the most fertile markets in which to develop and sell AI-powered services.
• Multiple surveys have found that Chinese users are one of the most willing and optimistic about AI services across the world, meaning that if Chinese companies can develop useful services, then they should enjoy rapid uptake.
• This, combined with leadership in algorithmic innovation, open source and energy availability, means that China should be leading the AI race, but the commentary from its leading technology companies indicates that things are not going well.
• Both companies are investing heavily, but their access to leading-edge semiconductors remains limited, meaning that the economics of building data centres at home will be questionable.
• I think that this means that both Alibaba, Tencent and any other company developing AI services for the domestic market will be looking overseas for compute.
• I think that this is the only way that these Chinese companies have any real hope of rolling out AI services at scale in China and doing so economically.
• The situation is also very uncertain, as Nvidia thinks that it will be selling H200 chips to China sometime soon, but permission for the chips to be imported into China has not yet been given by Chinese regulators.
• If Tencent and Alibaba can get access to high volumes of H200, it will substantially improve the economics of their AI infrastructure, but this remains very uncertain.
• Furthermore, the practice of importing compute from overseas is also something that could be targeted by the US Department of Commerce when it updates its export rules in Q3 this year.
• This is why the economics of AI investments made by Chinese companies are so uncertain, which could explain why both Alibaba and Tencent are being so vague about how AI will impact their financial performance.
• Despite this, the real problem that both of these companies face is domestic consumption, which is what ultimately underpins their core businesses.
• Here, the outlook is not good as the economy remains mired in high debt, a real estate crisis, weakening demographics and a high level of unemployment among young people.
• This means that the Chinese technology sector is increasingly looking like a classic value trap where the stocks are cheap but are likely to remain so until domestic consumption turns around.
• The focus of the 5-year plan presented to the National People’s Congress focused on technological independence and loyalty to the party while making one of the weakest economic forecasts seen for many years.
• Hence, I don’t think that the domestic economy is going to see any signs of recovery this year or next.
• I have been holding Alibaba for several years, looking for a recovery, but this outlook is beginning to test my patience.