February 20, 2026 | Chinese AI Companies Rock Silicon Valley

The United States is spending hundreds of billions to win the AI race.
China is trying to win by spending far less.

Nvidia’s high-performance chips and multi-billion-dollar data centers have become the foundation of America’s AI strategy. The largest U.S. technology companies are committing extraordinary capital — potentially as much as 500 billion dollars this year — to build ever larger compute clusters.

Several trillion-dollar companies such as Google, Meta, Amazon and Microsoft are effectively betting their future growth on artificial intelligence. Musk’s SpaceX and Ellison’s Oracle are not far behind.

The problem is straightforward: the capital expenditures are certain; the profits are not.

Revenues from AI products remain modest relative to the scale of investment. Yet valuations assume sustained dominance and very high future margins.

What if those margins never materialize?

Or more disruptive still: what if Chinese firms develop competitive models at a fraction of the cost?

Last year at the end of January, DeepSeek surprised markets with an open-source model that appeared far closer to U.S. capability than investors had assumed. Even modest signs of cost compression triggered sharp declines in AI-linked stocks. And DeepSeek claimed to have built their model with less than 1/10^th of the cost and compute power.

Nvidia’s stock just before the news was $144 and plummeted to $94 within two months.

Possibly as soon as this week a new version is expected, reportedly focused on coding — the most commercially promising AI segment. Whether or not it matches top U.S. systems, the signal matters: China is pursuing a different cost curve.

We have seen this movie before.

Chinese electric vehicle manufacturers now produce compelling vehicles at dramatically lower prices than many Western competitors. Solar panels followed the same pattern. Telecommunications equipment did as well.

If AI model training and inference costs fall rapidly — and especially if they fall first in China — today’s valuation assumptions for the largest U.S. technology firms become vulnerable.

This is not an argument that America loses the AI race.

It is an argument that investors may be underestimating how the race could change the economics of the industry.

And when capital intensity meets compressed margins, equity valuations tend to adjust — sometimes abruptly.

Hilliard MacBeth

The opinions expressed in this report are the opinions of the author and readers should not assume they reflect the opinions or recommendations of Richardson Wealth Limited or its affiliates. Richardson Wealth Limited is a subsidiary of iA Financial Corporation Inc. and is not affiliated with James Richardson & Sons, Limited. Richardson Wealth is a trade-mark of James Richardson & Sons, Limited and Richardson Wealth Limited is a licensed user of the mark. Richardson Wealth Limited, Member Canadian Investor Protection Fund.

Hilliard MacBeth February 20th, 2026

Posted In: Hilliard's Weekend Notebook

Next: