Nvidia Reports $96.2 Billion Quarter as Potential AI Server Tariffs Loom

Nvidia Corp. delivered record fiscal second-quarter financial results, generating $96.2 billion in total revenue—a 106% jump compared to the same period last year and an 18% increase from the prior quarter, according to financial reporting by Yahoo Finance. The semiconductor giant posted GAAP earnings of $2.46 per diluted share while maintaining a 75% gross margin, reinforcing investor optimism around the momentum of global artificial intelligence infrastructure spending.

The company's Data Center business unit generated $89 billion in sales during the quarter, representing a 117% year-over-year surge and accounting for roughly 92.5% of Nvidia’s overall quarterly revenue. That figure nearly doubles the $46.7 billion in total quarterly revenue Nvidia generated across all business units just one year earlier. Chief Executive Officer Jensen Huang noted that order momentum continues to gain speed as corporate enterprises, AI research laboratories, and early-stage startups expand their computing footprints.

Despite the strong performance, new trade policy considerations could introduce headwinds for the broader hardware ecosystem. Federal policymakers are considering a wider framework of semiconductor tariffs, as reported by Reuters and detailed by Yahoo Finance. Unlike previous narrow measures, the prospective rules under discussion could assess duties on completed technology hardware containing imported chips, including data center servers, consumer laptops, and video game consoles.

The proposed trade adjustments build on actions taken in January, when the White House imposed a 25% tariff on specific high-end computing processors following a Department of Commerce finding that U.S. reliance on overseas semiconductor production threatened national security. That order explicitly identified chips such as Nvidia's H200 and Advanced Micro Devices' MI325X, though it featured key exemptions for chips destined for domestic data centers, startup infrastructure, academic research, and selected industrial applications.

Because Nvidia functions as a fabless chip designer, it relies on overseas foundries to manufacture its most sophisticated accelerators rather than building and operating its own domestic fabrication facilities. Competitors with established U.S. manufacturing bases, such as Intel Corp., have poured capital into expanding domestic chip foundries. While complex global supply chains mean no single company is entirely isolated from trade friction, policy frameworks that favor domestic production could alter relative cost structures within the semiconductor sector.

The central uncertainty for enterprise technology buyers centers on whether future trade actions will target fully assembled computing systems. Modern AI infrastructure requires integrated racks combining GPUs, central processing units, high-bandwidth memory, high-speed networking equipment, power systems, and specialized liquid cooling—components sourced from multiple global suppliers. As Nvidia shifts from selling individual GPUs to complete computing platforms, system-level tariffs could affect total deployment costs for large-scale data facilities.

MarketWatch noted that potential regulatory changes could modify or eliminate the data-center exemptions established under the January tariff framework. A duty structure that retains data-center carve-outs would preserve current price levels for U.S. cloud infrastructure buildouts. Conversely, removing those exemptions would apply additional cost pressure to the hardware stacks powering modern hyperscale data centers.

While major cloud vendors, technology firms, and device manufacturers have demonstrated significant capital commitment by allocating tens of billions of dollars to AI hardware each quarter, industry observers are watching how elastic that capital spending will remain if regulatory policy drives up equipment prices across the supply chain.