Nvidia’s China Exports Tighten Even as RTX Spark Prepares a Windows Offensive

Nvidia clarifies China export compliance after BIS rule, posts $81.6B quarterly revenue, and unveils RTX Spark superchip for consumer laptops.

The same week Washington closed a loophole that could have let Nvidia’s most advanced chips reach China through offshore subsidiaries, the company unveiled plans to put petascale AI compute inside consumer laptops. The juxtaposition underscores how deeply regulatory risk and product ambition are now entwined for the world’s most valuable semiconductor firm.

Export rule targets Chinese parent companies abroad

On May 31, the Bureau of Industry and Security (BIS) clarified that any sale of advanced computing components to a company whose parent is based in China or Macau requires an export license — even if the buyer itself operates outside those jurisdictions. The clarification is not a new rule; the BIS noted the licensing requirement has been in effect since November 17, 2023. The update simply makes explicit that offshore structures of Chinese conglomerates fall under the same restrictions.

Reuters reported that the move was aimed at a potential gap through which top-tier processors such as Nvidia’s Blackwell line could reach Chinese affiliates located beyond the mainland. Nvidia itself said little had changed operationally. A company representative told Reuters the Commerce Department had already imposed licensing obligations on Nvidia via a separate letter. In a statement to Al Jazeera, Nvidia noted that its sales and vetting processes already align with the clarified rules, adding that licenses for controlled products to China-based entities have long been standard practice.

Financial foundation contains the shock

The clarification landed hard on the heels of Nvidia’s fiscal first-quarter results, which already baked in minimal China exposure. For the quarter ended April 27 (the first quarter of fiscal 2027), Nvidia posted revenue of $81.6 billion, up 85 percent year over year. The data-center segment alone generated $75.2 billion, a 92 percent surge. GAAP gross margin came in at 74.9 percent, and diluted earnings per share reached $2.39.

For the current quarter, Nvidia guided for $91.0 billion in revenue, plus or minus 2 percent. Crucially, the projection assumes zero data-center revenue from China. That cushion means the BIS clarification does not force an immediate downward revision. The regulator also made clear that existing data-center systems do not need to be shut down; the directive applies only to future transactions and compliance. The central question for investors is whether the rule change alters the volume or timing of shipments to China-linked customers outside the mainland — and whether those volumes are material enough to matter against a $91 billion quarterly target.

PC ambitions shift focus to the desktop

While regulators tightened the screws on China sales, Nvidia’s Computex presentations pointed in the opposite direction: toward a mass-market PC push. The newly introduced RTX Spark superchip marries a CPU with up to 20 cores and a Blackwell GPU with 6,144 cores, sharing up to 128 GB of unified memory. Nvidia claims the chip delivers a petaflop of AI performance, allowing laptops and desktops to run large language models, generative-AI video, and ultra-large 3D scenes locally — tasks that previously required data-center firepower.

Jensen Huang framed the chip as turning the PC from a tool into a “teammate” for complex workloads. Microsoft is a close collaborator: the two companies are tailoring Windows for personal AI agents that can plan, create content, and operate across applications. Adobe is already adapting Photoshop and Premiere to the new platform, underscoring Nvidia’s reliance on a software ecosystem as much as hardware.

First machines with RTX Spark are slated for autumn, with Dell, Lenovo, Asus, HP, MSI, Acer, and Gigabyte among the launch partners. Microsoft’s own Surface Laptop Ultra will also use the chip. For Intel and AMD, the move turns a longstanding rival in graphics into a direct competitor for the CPU socket.

Robot brains and autonomous models expand the AI footprint

Parallel to the PC offensive, Nvidia continues to extend its AI reach into physical systems. Cosmos 3, an open foundation model for physical AI, is designed for world simulation and action generation. The Isaac GR00T platform provides an open reference for humanoid robotics, while Alpamayo 2 Super, a vision-language-action model with 32 billion parameters, targets Level 4 robotaxi development.

These products, alongside RTX Spark, signal a strategy that pushes AI onto devices used by millions every day — not just into data centers. The autumn launch of the PC chips will serve as the first real-world test of that broader ambition.

Market takes the mixed signals in stride

Investors have largely shrugged off the BIS headlines. The stock closed at €196.98 on June 2, up 2.12 percent on the day, and had gained 6.68 percent over the prior week. By Tuesday of the following week it had eased to €193.64, a 0.38 percent daily gain but still showing a 14.05 percent monthly advance. The shares sit 3.69 percent below their 52-week high. Year to date, Nvidia has climbed 22.27 percent, with a market capitalization of roughly $5.59 trillion.

The measured price action suggests the market views the export clarification as a known variable — one Nvidia’s own guidance has already priced in. The real story may be whether the RTX Spark and its ecosystem can open a new revenue stream large enough to offset whatever China-related headwinds eventually materialize.