Source: Shutterstock
By Bolaji Ojo
“Nvidia stands at a crossroads. Its business is growing fast but it must satisfy hard taskmasters (China and US) with conflicting objectives.” I wrote this late last year, arguing: China Isn’t for Everyone, Maybe Not Even Nvidia.
This week, Nvidia said its fiscal “first quarter results are expected to include up to approximately $5.5 billion of charges associated with H20 products for inventory, purchase commitments, and related reserves.” Those charges pertain primarily to China.
The company’s China fears and nightmare has become its reality.
Why? Because the United States government (USG) on April 9 informed Nvidia that the export of its H20 products now requires “a license for export to China (including Hong Kong and Macau) … and on “April 14, 2025 the USG informed the company that the license requirement will be in effect for the indefinite future.”
Nvidia’s stock price fell 6 percent on the announcement. What’s the surprise here? That the US is in a trade war with China, that the US government does not approve of the sales of artificial intelligence accelerators developed by Nvidia to enterprises in the Communist nation or that Nvidia would somehow avoid getting snagged by the tangled geopolitical mesh the world now finds itself?
The future of Nvidia in China was carved in stone long before the Donald Trump administration came into office. Nobody needs a diviner to figure this out. Nvidia is the world’s biggest supplier of artificial intelligence accelerators, a technology considered groundbreaking and so disruptive that we even now have limited vision of the world that will emerge once it is fully developed.
AI is impacting global economies and promises to upend all areas of human activities, including not just enterprise activities but also major activities governments deem critical to national survival. It is going to reshape – and it is already reshaping – what happens in national politics, defense and military activities, aviation, automotive, banking and finance, defense, insurance, military, medical, industrial, space, transportation … anything involving humanity. In a matter of decades, AI will conceive and give birth to things and structures we’ll struggle to define and imagine.
Nvidia, for now, is the centrifugal force that is giving form to and helping in the shaping of phenomena that its founder and CEO Jensen Huang may not already grasp. But he can and should at least realize and accept that there are concerns about the evolution of this technology and the challenges it poses to existing state structures and systems. Huang should by now be aware of the frictions threatening to disrupt and destroy the economic and political systems upon which the modern world was built.
If he doesn’t get this, his position as CEO at Nvidia will be in jeopardy. Huang may have already entered the danger zone, though. If he doesn’t appreciate or comprehend the catalyst role Nvidia has in today’s geopolitical system and the evolving one, then he will not last long in his position.
Pick a side
We don’t know how the geopolitical state of the world will evolve over the next years. It is clear, however, that the current system is getting a structural test, with scant hopes of returning to the state it was only months ago. Threats to national boundaries are happening, and alliances are getting tested as new ones are explored. Until the shaking ends, enterprises will have to pick a side.
Nvidia will have to cast its lot with either China or the United States. The choice is that stark. But it is also quite simple. For now, most of its customers belong in the now fraying “Western” bloc. But the glue holding the “West” together is drying and flaking off. For example the old Canada-US relationship is “over,” prime minister Mark Carney said after Trump claimed Canada country as the 51st US state.
A similar vein has popped in Europe. Ursula von der Leyen, president of the European Union, in an interview with reporters this week said the “West as we knew it no longer exists.” Von der Leyen referred to the US’ anchor role in EU-US relationship and how the imposition of tariffs by America has challenged the trust fundamentals between the two regions. “We don’t have bros or oligarchs making the rules,” von der Leyen said. “We don’t invade our neighbors, and we don’t punish them.”
Nvidia has a different reality. The US can do what it wants with Nvidia. America, under Trump, has not rejected the notion of invading friendly or neighboring nations. America, under Trump, is ready to punish friends and foes alike. Nvidia is neither a friend nor a foe. It is a US-incorporated enterprise that has developed a technology America sees as fundamental to its dominance of global commerce, defense and politics. America does not want to share this with China and not even with countries that previous administrations considered US allies.
Nvidia must choose a side even if its options are limited.
The company’s supply chain is lopsided in its geographical spread. It is “mainly concentrated in the Asia-Pacific region,” as Nvidia acknowledged in its fiscal 2025 annual report filed with the US Securities and Exchange Commission. Taiwan-based TSMC manufactures Nvidia silicon along with South Korea’s Samsung Electronics. Another South Korean company, SK Hynix, supplies a chunk of its high bandwidth memory (HBM) with Boise, Idaho-based Micron Technology chipping in.
Nvidia’s AI platforms are assembled by “independent subcontractors and contract manufacturers such as Hon Hai Precision Industry Co. Ltd., Wistron Corporation, and Fabrinet,” the company said, in the same 10-K filing. Both Hon Hai and Wistron are Taiwanese. Fabrinet is based in Thailand. Nvidia is the poster child for the vulnerable and Asia-dependent “Western” technology company.
Inevitable Split
Nvidia is aware of the implications of its lopsided supply chain. The Joe Biden administration had imposed restrictions on what the company could export to China, forcing it at one point to develop separate and less-advanced products for China and a group of countries that included Saudi Arabia, the United Arab Emirates and Vietnam.
The US government later made shipment of AI accelerators to these and other countries contingent upon the approval of export licenses. The licensing requirement covered the sale of the company’s A100, A800, H100, H800 and L40S products, Nvidia said, noting in the fiscal 2025 annual filing that it had “to date, not received licenses to ship these restricted products to China.”
“Our competitive position has been harmed by the existing export controls, and our competitive position and future results may be further harmed … if there are further changes in the USG’s (United States Government’s) export controls,” Nvidia said, in its filing. “Given the increasing strategic importance of AI and rising geopolitical tensions, the USG has changed and may again change the export control rules at any time and further subject a wider range of our products to export restrictions and licensing requirements, negatively impacting our business and financial results.”
Nvidia didn’t need a crystal ball to come to this conclusion,/ Months ago, there were sufficient signs that the financial charge of $5.5 billion for the fiscal 2026 quarter ending April 27, 2025, was inevitable. In the fiscal year ended January 26, 2025, China, including Hong Kong, contributed $17.1 billion of Nvidia’s $130.5 billion revenue, representing 13 percent of its sales. The comparable figures for fiscal 2024 and fiscal 2023 were 17 percent and 22 percent, respectively.
The trend is obviously down. Does anybody still question the advice that Nvidia should wash its hands off China?
Here’s the rub. Nvidia will either voluntarily quit the Chinese market or its presence in the country will be erased by forces beyond its control. By moving out now, it can at least orchestrate an orderly withdrawal from a battleground where it is getting hogtied by two irreconcilable combatants.
Related articles:
China Isn’t for Everyone, Maybe Not Even Nvidia
Prefix AI, Nvidia and TSMC will dominate a slow 2025
Synopsys Talks and Walks AI with Nvidia
Nvidia: Not ‘Just a Chip Company’ Anymore
Bolaji Ojo is publisher and managing editor of the Ojo-Yoshida Report. He can be reached at [email protected].
Copyright permission/reprint service of a full Ojo-Yoshida Report story is available for promotional use on your website, marketing materials and social media promotions. Please send us an email at [email protected] for details.