Nvidia's Order Book Meets Its Limits: China Supply Gap, DOJ Scrutiny, and a November Reckoning

Chinese artificial-intelligence and cloud providers are bracing for a severe shortfall on Nvidia hardware, with Bloomberg reporting that they expect to receive only tens of thousands of the RTX Pro 5500 chips they have ordered over the next four months. Against roughly 6 million units on order, that works out to about 1% of the requested volume — a gap the report attributes to constrained supply.

The figure captures the distance between demand and delivery, though it reflects the expectations of Chinese buyers rather than any shipment forecast confirmed by Nvidia. Nor does it translate directly into a specific revenue loss. Because the horizon spans four months, the issue is not a brief delay on individual consignments but availability across a longer planning window — which is why any sound assessment has to separate robust demand from the company's ability to actually fill it.

Capacity Plans Cut Both Ways

Nvidia has not stood still on the supply side. On 21 October it unveiled plans with supply-chain partners to expand global manufacturing capacity, covering accelerated computing and its full platform for AI factories. That announcement speaks to the underlying tension between demand and available capacity, but it does not demonstrate that relief has already reached the orders Bloomberg described. A planned build-out is not the same as additional product shipped, and the global expansion and the expected deliveries to Chinese customers sit on different levels. Nothing in the capacity announcement alone implies that these buyers' supply situation improves within their own timeframe.

A $12.93 Billion Deal Draws Washington's Eye

Competitive questions have meanwhile surfaced around Nvidia's proposed acquisition of Hugging Face. Reuters reported on 28 October that the US Department of Justice has sought information on the deal's antitrust implications, putting the execution of a multibillion-dollar purchase alongside the infrastructure build-out on investors' agenda. The transaction carries a volume of $12.93 billion, with closing anticipated in the first half of 2027 — meaning it remains a proposal, not a completed acquisition. The inquiries concern competitive effects and are distinct from the operational expansion plans: announced partnerships do not prove the deal has closed, and the department's attention does not amount to a decision against it.

Growth ambitions stretch well beyond Hugging Face. Roughly a week before the Reuters report, Nvidia and Oracle said they would scale Oracle Cloud Infrastructure superclusters to several gigawatts, deploying Nvidia's Vera Rubin NVL72 systems, with those systems and Vera processors also declared generally available on OCI. Ties with Meta are widening too: on 27 October the two companies announced an expanded collaboration for Meta's Muse model family, and Meta plans to deploy Nvidia infrastructure on a larger scale. Those moves concern partnerships and infrastructure use — not the regulatory status of the Hugging Face purchase.

Possible restrictions on cloud partners are also in the air. The Financial Times reported that Nvidia has instructed customers to sell at least 51% of their services to non-Chinese companies; Reuters said it could not immediately verify the report. The claimed requirement should therefore not be treated as confirmed business policy.

17 November Is the Date That Counts

The next confirmed financial marker is the earnings call scheduled for 17 November 2026, announced on 28 October for the third quarter of fiscal 2027, a period that ended on 25 October. For investors, the results are the next hard checkpoint — and the key question is not simply how large demand is, but how much of it Nvidia can genuinely serve. The Bloomberg report points squarely at that execution gap. The quarterly figures will also be the next concrete anchor for weighing the growth plans, though they will not substitute for a completed acquisition or a Justice Department ruling.