Given that there are numerous companies – particularly in the East – that are building out national AI infrastructure with government support, the concentration of revenue sources is a concern for the company going forward. These companies are highly disincentivized from purchasing products of an American company (which Nvidia ostensibly is) towards the buildout of critical infrastructure. This is particularly true under “Trump Season 2”, i.e. Donald Trump’s second term in office.

Among the Eastern countries building out AI infrastructure, China and India lead the pack. While the administration expected to receive 15% of the revenue generated from licensed H20 sales, Nvidia reports that China’s purchases of H20 chips ground to a halt in its Q2. Instead, an “unrestricted” customer outside of China purchased $650 million of Nvidia’s H20 inventory. Furthermore, the company stated that the U.S. government hasn’t yet published a regulation codifying the 15% payout.

The idea of a 15% payout is, in itself, possibly indefensible. The original intent of the export control regime was to restrict China’s access to advanced technology. A 15% “tax” isn’t the same as a “restriction” ostensibly brought in to prevent the loss of intellectual property that was the basis behind the control regime. Considering that “Trump Season 1” initiated these restrictions and the subsequent Biden administration maintained it, the “tax” demand would be a reversal of stance that had bipartisan support in the legislature. Also, at a 15% level, the company likely wouldn’t make a profit (at best) or (perhaps) incur a loss – which means that Nvidia might have to exit the China market.

U.S. Commerce Secretary Howard Lutnick publicly stated³ that China would only get Nvidia’s “fourth best” chip and the plan was to “sell the Chinese enough that they get addicted to the American technology stack.” As it stands, the H20 is reportedly already surpassed by other chip design companies and the Chinese government, subsequent to Lutnick’s comments, told its tech companies to stop using H20 chips. Moreover, as DeepSeek’s early success despite the lack of top-of-the-line chips suggests, there is more than one way to crunch AI data: while Nvidia’s leading products offers a means to do it the hardware route, software- and code-driven optimizations are another demonstrable means of doing so with “older” and “less advanced” chips. The key determinant in choice is hardware availability and cost.

Next, India was referenced in a number of past earnings releases wherein the company stated that it was making inroads with local companies. With foreign-based computation resources deemed prohibitively expensive, the Indian government had recently allocated⁴ $1 billion in funding to kickstart native AI development ecosystems predominantly among small businesses and startups while including the buildout of a “national AI compute grid” within a country that already has 16% of the world’s AI workforce. A significant presence in this key market has been put paid by Trump’s tariff war on the country.

Now, the tariff hikes per se don’t presently⁵ impact India’s predominant exports to the U.S. – namely goods made of iron, steel, aluminum and copper; automobiles and related components; pharmaceuticals; petroleum products and electronics. However, the language employed by the President and his administration has created substantial rancour – which is especially true in the East wherein words are chosen and employed with care. It also tarnishes nearly 25 years of diplomatic and strategic outreach that began with the George W. Bush administration and subsequently raised to much higher levels with substantial bipartisan consensus. It is quite likely that India’s body politic, industries and populace don’t consider the U.S. to be a reliable trading partner anymore, with American companies also likely to face a distinct cooling off in its fostered relations going forward.

As of on the 29th of August, the Indian Prime Minister is visiting another of India’s historic and esteemed trading partners: Japan. A number of agreements in the areas of semiconductor manufacturing and AI⁶ along with a “talent-exchange” program that will see 500,000 professionals⁷ being granted work opportunities in each other’s countries over the next five years.

These moves currently in play by two major Eastern powers that don’t see eye to eye on all matters – with India, in particular, historically being a strategic partner for the U.S. against China – are being executed independently and in alignment with their own interests. These have long-term consequences for American companies that need a global market base with no substantial incentive to have them replaced. Through the current administration’s behavior, there is now an incentive.

Overall, these externalities effectively impair the long-term growth outlook for Nvidia despite its best efforts and through no fault of its own. However, “tech” is effectively considered the next best thing to U.S. Treasuries (and for some investors, even preferable to) in a market with increasingly low breadth. While Nvidia won’t expect a precipitous market drop, it is increasingly less a “global” force and more an “American” one. Over the next FY or so, growth can be expected to progressively lower.

With time, being “American” wouldn’t necessarily mean a “world leader” and this would be particularly true in strategic industries such as the one Nvidia inhabits. Bouts of bearishness in the stock can be expected going forward, although low market breadth could possibly create some support.