Nvidia Weekly Investor Briefing: From Record Highs to China Shock

Nvidia’s market share in China has dropped to zero. That’s a direct quote from CEO Jensen Huang, and it’s a number that changes the entire story. Last week we discussed Nvidia outpacing expectations, but this week we have to talk about where those expectations just hit a concrete wall. The record five-point-four trillion dollar market cap reached on May thirteenth is the headline everyone sees. The China news is the detail that will matter in a decade. Let's look at the numbers everyone is waiting for. The company reports Q1 earnings on May twentieth.

The street expects seventy-eight-point-seven-five billion dollars in revenue, which is roughly eighty percent year-over-year growth. Citi is even more bullish, projecting eighty billion, citing a stronger-than-expected ramp of the new B300 GPU. The data center segment is expected to deliver almost seventy-three billion of that. These numbers are staggering. They reflect a company operating at the absolute peak of its power, with demand for its Blackwell and Rubin systems already exceeding five hundred billion dollars. Jensen Huang himself now projects that opportunity will hit at least one trillion dollars through 2027.

So with the stock up seventy-four percent in the last twelve months, and earnings looking this strong, the bull case seems airtight. Here’s the problem. That five-point-four trillion dollar valuation is a bet that this momentum is permanent. And two things happened this week that suggest it is not. First, as I said, is China. Huang’s statement wasn’t an estimate. It was a final verdict. "We had, you know, call it 90-some-odd percent of the world’s market share. Today, in China, we have now dropped to zero." A geopolitical wall has been built, and Nvidia is on the outside.

In 2045, the loss of the world's second-largest economy as a customer will have mattered far more than one quarter's earnings beat. Second, the competition is finally changing the game, not just playing it. AMD’s CEO Lisa Su laid out a different vision for the future of AI. She noted the server CPU market is set to grow over thirty-five percent annually, hitting one hundred and twenty billion dollars by 2030. Why? Because the next phase of AI is less about training massive models on GPUs, and more about running inference workloads—agentic software—which rely heavily on CPUs.

AMD’s own data center revenue was up fifty-seven percent. This isn't a direct attack on Nvidia's GPU dominance. It's an end-run around it. The market is optimizing for the world Nvidia built, where the GPU is king. But the long-term trajectory is shifting. The fight for AI infrastructure is expanding from a one-front war over training chips to a two-front war that includes inference. Nvidia’s valuation prices it as the winner of a war that is already over. The risk is that the next war has already begun, and the battlefield is much larger than Wall Street currently believes.

This brings us to our recommendation. Nvidia’s technology is undeniable. Its ecosystem is a powerful moat. But the stock is priced for a level of perfection that the world no longer offers. The loss of China is a permanent impairment to its total addressable market. The rise of CPU-driven inference is a structural threat to its GPU-centric dominance. Therefore, our recommendation is to HOLD. We are setting a twelve-month price target of six hundred and fifty dollars per share. The current momentum could certainly carry the stock higher post-earnings.

But we are investing for the next decade, not the next quarter. Before we can upgrade to a buy, we need to see Nvidia’s strategy for a world where China is a competitor, not a customer, and where the CPU is re-emerging as a critical piece of the AI puzzle. Momentum can build a five-trillion-dollar company. A competitive moat is what lets you keep it.