Jensen Huang has thrown in the towel on China—at least as far as Nvidia's financial forecasts go.
It's a stunning admission from the typically bullish CEO. Imagine telling your investors, "Hey, we're just gonna pretend that massive market over there doesn't exist in our projections." Bold move. Necessary? Probably. A little alarming? You bet.
The backstory here is painfully familiar to anyone watching the tech cold war unfold. Nvidia custom-designed its H20 AI chips specifically to comply with U.S. export restrictions, only to have regulators effectively move the goalposts in April. The result? A $2.5 billion hit to revenue that, frankly, the company barely seemed to notice given its otherwise astronomical growth.
I've covered semiconductor politics since the Trump administration first took aim at Chinese tech, and this particular episode feels different. There's something uniquely frustrating about watching a company play by the rules only to have the rulebook rewritten mid-game.
"It's like baking a cake with the exact ingredients the recipe called for, then being told your cake is illegal anyway," one industry analyst told me over coffee last week. He requested anonymity to speak freely about government policy. "What message does that send to American companies trying to navigate these restrictions?"
The timing couldn't be more ironic. While U.S. officials are busy blocking even watered-down AI chips from reaching Chinese data centers, Huawei—the very company these policies were originally aimed at crippling—is stepping into the void.
Look, there's a fundamental disconnect here between policy goals and outcomes. Huang himself pointed this out (though in more diplomatic language) when he noted that "the goals of the export controls are not being achieved."
He's right.
History has this annoying habit of repeating itself. When the U.S. blocked supercomputer exports to China back in 2015, Beijing didn't exactly throw up its hands in defeat. By the following year, China had built the world's fastest supercomputer using—surprise!—entirely domestic processors.
The lesson? Technology restrictions don't prevent advancement; they redirect it.
(And they cost American companies billions in the process.)
What's Nvidia doing about all this? Pivoting hard toward Europe, for one thing. The company just announced plans for the world's first cloud computing platform dedicated to industrial AI applications there. Smart move... if you can't sell to one massive market, find another.
The whole situation reminds me of that old saying about closing barn doors after horses escape, except in this case, we're bolting the door while simultaneously teaching the horse how to pick locks.
Dan Ives of Wedbush didn't mince words when he described the H20 ban as essentially "handing a good portion of Nvidia's business directly to Huawei on a silver platter." Having interviewed Ives several times over the years, I can confirm he's not prone to hyperbole.
For investors watching this unfold, the million-dollar question is whether this China forecasting blackout represents a temporary adjustment or the beginning of a permanent technological divorce between East and West. The implications are vastly different depending on which it turns out to be.
Nvidia's strategy makes perfect sense in the meantime: don't promise China revenue, then deliver a pleasant surprise if restrictions ease. It's Corporate Management 101—under-promise, over-deliver.
But there's something profoundly weird about a leading American tech company having to exclude the world's second-largest economy from its financial planning. It's like a global restaurant chain pretending an entire continent doesn't exist on its balance sheet.
The semiconductor industry has always been global by nature. These new fragmented supply chains and segregated markets run counter to decades of established business practices. They're creating inefficiencies that benefit... well, that's the trillion-dollar question, isn't it?
Wall Street doesn't seem particularly bothered for now. When you're posting 69% year-over-year growth, investors tend to forgive a lot. But the longer-term implications of this technological splintering deserve more attention than they're getting.
In this high-stakes game of technological chess between superpowers, the pawns are corporate balance sheets and innovation timelines.
And I'm not convinced anyone has thought more than three moves ahead.
