Nvidia’s China Exit and $1 Trillion Order Book Set the Stage for a Make-or-Break Earnings Day

Despite US export controls wiping out its China business, Nvidia sees $1 trillion in orders through 2027 as hyperscalers ramp up AI infrastructure spending.

The AI chipmaker that once commanded more than 90% of China’s accelerator market now holds precisely zero. CEO Jensen Huang confirmed the complete evaporation of that business during an investor call on Wednesday, attributing the collapse to Washington’s tightening export controls. The US government, determined to preserve its technological edge, has effectively barred Beijing from accessing Nvidia’s latest chip generations — and Huang made clear he is falling in line with that mandate.

Yet the loss of what was once the company’s most important growth market has done little to rattle investors. The reason is simple: demand from everywhere else is so ferocious that supply still cannot keep pace.

The world’s largest cloud operators are throwing open their wallets at a pace never seen before. Google, Amazon, Microsoft and Meta collectively plan to spend roughly $725 billion on capital investments in 2026 — a 77% surge from the prior year. Amazon alone is earmarking around $200 billion, while Microsoft has committed about $190 billion, well above the $154 billion analysts had penciled in. Meta has lifted the ceiling on its investment forecast to $145 billion.

This is not merely an arms race for chatbot capacity. The hyperscalers are building infrastructure for autonomous AI agents, robotics and AI-powered search — applications that consume multiples more computing power than today’s large language models. And every dollar they spend on that infrastructure funnels directly into Nvidia’s ecosystem.

Huang, speaking at the GTC conference in March, put a staggering number on the table: he now sees purchase orders worth over $1 trillion for the Blackwell and Vera-Rubin platforms stretching through at least 2027. That is double the $500 billion pipeline he cited just a year earlier. The Vera-Rubin platform, scheduled for volume shipments in the second half of this year, promises to train AI models with 75% fewer GPUs than Blackwell and slash inference costs by 90%.

The company’s product roadmap remains aggressive despite the China void. Nvidia is now refreshing its architectures on an annual cadence. The current Blackwell systems will be followed by the Rubin platform in 2026, a Rubin Ultra variant in 2027, and the Feynman architecture in 2028.

All of this sets the stage for the earnings report on May 20, when Nvidia will release results for the first quarter of fiscal 2027. Management has guided for revenue of $78 billion — a 77% year-over-year increase. But the analyst consensus already sits at 79% growth, and many on Wall Street consider 80% or more the minimum threshold for a positive stock reaction. For the second quarter, the Street is looking for $86.6 billion.

The stock currently trades at €167.96, roughly 8% below the 52-week high of €182.26 set in late April. The year-to-date gain stands at about 68%. The relative strength index hovers near 50, while annualized volatility remains elevated at 34%.

Analyst price targets paint a bullish picture. DBS lifted its target to $250 at the end of April. Bernstein and Cantor Fitzgerald each see $300, while Rosenblatt is calling for $325.

Nvidia enters the reporting day with net liquidity of $51 billion and a gross margin above 71%. The data-center segment, which powers the bulk of the story, posted revenue of nearly $194 billion in the most recent period — up 75% year over year, driven by Hopper and Blackwell system sales.

The China exit is a stark reminder of geopolitical risk, but the math tells a different story. With hyperscaler spending accelerating, a $1 trillion order pipeline, and a product cycle that shows no signs of slowing, the May 20 print will test whether even the loftiest expectations are high enough.