China’s AI Models Challenge U.S. Aspiration

AI Strategy

Wall Street has been valuing AI-related companies in the U.S. as if they have infinite market potential.

However, Sam Altman acknowledged earlier this year that the U.S.’s proprietary AI approach may prove to be restrictive. More specifically, OpenAI’s strategy to make ChatGPT closed source was short-sighted.

China is pursuing a commodification strategy for AI, in contrast to the U.S. approach, and DeepSeek demonstrated a different method for building AI models at a much lower cost.

The Chinese are providing an operating system for developers and companies to build on, enabling companies to deploy models on their own private infrastructure and ensure sensitive data remains in-house.

These open-source, more efficient models are being rapidly adopted globally, and the market, including the U.S., is increasingly occupied by AI models developed in China.

Companies like Airbnb (using Alibaba’s Qwen) are attracted to these smaller, more efficient AI models that are highly specific and tailored to them.

AI Capital Expenditure

Markets always have one eye on capital expenditure.

If you compare AI models, it’s not clear what you’re getting for the additional U.S. capital expenditure.

To put this into perspective, while the U.S. maintains an overall lead in AI investment, data center build and talent, China excels in applied AI and is rapidly closing the performance gap.

Notably, China has now surpassed the U.S. in monthly downloads of artificial intelligence models; marking a significant turning point in the global AI race.

If it’s cheaper, as efficient, and it’s good enough, then that is likely the infrastructure or AI ecosystem that the rest of the world will build upon.

In short, the U.S. model is likely to have a smaller market than that being pursued by China.

This shift, if permanent, could eventually affect U.S. AI-related company valuations.

S&P 500 Key Levels

The initial response to the FED rate cut announcement last Wednesday was broadly positive, propelling the S&P 500 to near all-time highs.

However, Friday reversed those gains, and the index closed down 0.5% from the previous week.

While the reason for the correction is not clear, it’s likely a combination of the Bank of Japan signalling an increase in interest rates at their next meeting and news that Oracle had delayed several new OpenAI data centers due to material and labour supply.

The market fell sharply when the Oracle news broke; however, it rallied at the point that Oracle denied there was any delay.

Options high level today sits at 6956, and the low is 6773.

Seasonally, the last two weeks of December and the first two weeks of January are positive for markets.

The 6800 level was a key support level (previously resistance) on Friday and will likely hold this week.

The advanced decline line reached an all-time high last week, suggesting the market is not yet at its all-time high.

The VIX, a measure of stock market volatility, sits around 15, and bond market spreads, which reflect risk in borrowing costs, show little sign of stress.

While the balance of probability points to further upside, the Nonfarm Payrolls report and CPI data will likely be key catalysts that sets the market’s direction.