Chinese AI models can trigger massive capital destruction in US: Chris Wood

Cheaper Chinese artificial intelligence (AI) models can trigger a ‘massive capital destruction’ in the United States (US) markets, warned Christopher Wood, global head of equity strategy at Jefferies in his weekly note to investors, GREED & fear.

 

The best way to play the AI trade, Wood believes, is the picks and shovels trade, in terms of companies who are profiting, or are likely to benefit directly from the hyperscalers’ spending.

 

“The key issue has always been one of timing in the sense of when the market will start to worry about the return on investment (made in AI). It would be interesting to know the extent to which Anthropic and OpenAI’s revenue growth slowed month-on-month in June, as corporates cut back on their previous practice of positively encouraging employees to play around with AI models regardless of the cost,” Wood wrote.

 

 

Kospi tanks 11%

 

On Tuesday, South Korea's Kospi index tanked nearly 11 per cent in trade as investors grew anxious regarding the AI-related boom.  

 

Trading was temporarily halted during the intraday deals as the index dropped to its lowest level since April, closing 10.8 per cent lower at 6,023.66. Samsung Electronics (down 13.4 per cent) and SK Hynix (14.7 per cent) were among the top losers.

 

The rise in AI-related stocks has been too fast, too soon and investors are now starting to question the capex plans of companies and return on capital employed (ROCE), said Jyotivardhan Jaipuria, founder and managing director at Valentis Advisors.

 

“It is a double whammy. The stocks have gone up too fast and also made investors cautious on the capital spending of these AI-related companies. It is natural for investors to take some money off the table given the sharp run in most of the AI and new technology related companies,” Jaipuria added.

 

Token processing

 

Meanwhile, Chinese large language models (LLMs), Wood said, are gaining popularity. The top Chinese AI models, Wood wrote, processed 36.39 trillion tokens on the global aggregator platform OpenRouter in the week ended July 19, up from 4.37 trillion in late April.  

 

“This compared with 7.39 trillion tokens for the top US models, based on the weekly usage of the top nine models on OpenRoute. There is also a growing realisation now that China has become a technological peer to the US in AI, as well as in so many other areas,” he wrote. 

 

 

And all this, Wood believes, has not been priced in yet by the markets. The growing narrative around AI, Wood cautioned, also faces growing credit risks.

 

The AI capex spend over the past year in the US, Wood wrote, has been funded less by hyperscalers’ cash and more by debt. “The hyperscalers have now issued more investment-grade debt than the energy sector in the US, and are now the largest single issuer having raised $194 billion of debt year to date compared with $55 billion of investment-grade issuance in the energy sector,” Wood wrote.

 

This is why his base case remains that the US markets have already made an all-time peak as a percentage of world stock market capitalisation. “This is another reason why the relative and absolute performance of the hyperscaler stocks should be monitored closely,” Wood wrote.