Nvidia has delivered its initial H200 chips to China, but anticipates that it will not generate any revenue from data centers as a result.
Nvidia has begun shipping its H200 processors to China as part of a new US licensing arrangement, although these shipments made up less than 1% of the $89 billion in quarterly data center revenue. The company's forecast of $108 billion for the current quarter does not anticipate any revenue from data center computing in China.
The expectation from Nvidia regarding these H200 processors is clearly defined: investors should not expect significant contributions from them. The processors were dispatched in the quarter ending July 26 under the new licensing scheme and represented a minimal fraction of the total data center revenue.
Notably, Nvidia's forecast warrants careful attention, as the company expects to reach $108 billion for the current quarter while assuming no revenue from Chinese data center computing.
Nvidia's CEO, Jensen Huang, remarked that even that figure is constrained, suggesting that actual demand could exceed the 70% growth the company anticipates for the following year.
In light of this, the figures offer a different perspective. A single supplier can exclude the world’s second-largest economy from its projections and still raise its guidance by nearly 20%.
In contrast, Europe is responding with plans for seven AI gigafactories. The European Commission has started the bidding process for these facilities, estimated at €30 billion, with approximately €10 billion coming from public funds and €20 billion from private investments.
However, only about €1 billion of the public funding is currently secured; the rest hinges on future long-term budgets, which a Commission representative refrained from prematurely discussing.
When comparing these figures, it's important to note that Nvidia generates around a billion dollars in data center revenue daily, meaning the committed European funding for the seven gigafactories amounts to roughly a day’s worth of Nvidia's sales.
Each facility is expected to accommodate at least 100,000 advanced chips that will be sourced from the open market, specifically from the company that projects $108 billion in revenue.
For buyers, the price dynamics are unfavorable. Nvidia has informed its customers that server prices are set to increase by over 15% starting early next year.
Furthermore, the timeline does not favor quick action; construction is slated to commence in early 2027, with initial operations beginning by mid-2028.
The Next Web has previously pointed out that viewing this matter through the lens of sovereignty is somewhat superficial. Acquiring the same silicon from a European source offers an illusion of independence rather than genuine self-sufficiency.
Ultimately, the narrative surrounding the H200 processors is about leverage. Nvidia can forfeit a market as significant as China and continue to grow, while Europe is working towards establishing facilities without the necessary funding and relying on chips that it does not manufacture.
Other articles
Nvidia has delivered its initial H200 chips to China, but anticipates that it will not generate any revenue from data centers as a result.
Nvidia has sent its initial H200 shipments to China but continues to anticipate no data center revenue from that market. In response, Europe plans to invest €30 billion in gigafactories, which are largely without funding.
