NVIDIA's AI Dominance Outshines Regulatory Hurdles: A Buy at These Levels?

NVIDIA's fiscal Q1 2025 earnings underscore a paradox: record revenue growth in its data center business—up 427% year-over-year—has been overshadowed by geopolitical headwinds. While U.S. export restrictions to China cost the company $2.5 billion in lost revenue and forced a $4.5 billion inventory write-down for unsellable H20 chips, NVIDIA's strategic pivot to U.S.-based AI manufacturing and its unshaken AI leadership position the stock as a compelling buy. Here's why investors should look past near-term noise and focus on the $500 billion AI infrastructure opportunity ahead.

The Data Center Tsunami: Revenue Growth Defies Headwinds
NVIDIA's data center division now accounts for 88% of total revenue, soaring to $22.6 billion in Q1. This is not merely growth—it's a seismic shift in the AI infrastructure landscape. Major cloud providers like Microsoft, Amazon, and Google are racing to scale AI capabilities, and NVIDIA's GPUs remain the gold standard. Even with the H20 chip sidelined in China, demand for its Blackwell chips has exploded: cloud customers have ordered 3.6 million units in 2025 alone, a 170% jump over prior-gen H100 shipments.
The key metric here is non-GAAP gross margins, which held steady in the mid-70% range despite the inventory charge. This signals that NVIDIA's pricing power and scale are intact. The $4.5 billion write-down was a one-time hit, not a recurring drag. CEO Jensen Huang's emphasis on “AI factories”—specialized data centers optimized for advanced reasoning tasks—suggests that margins will expand as these facilities come online in the U.S. over the next 18 months.
The China Dilemma: A Speedbump, Not a Roadblock
The $8 billion in projected Q2 China revenue losses from H20 chip restrictions are painful but not terminal. NVIDIA's CEO framed this as a policy misstep that risks accelerating China's AI self-reliance—but the company is already countering this by reshoring manufacturing. By 2026, its Arizona and Texas facilities will produce AI chips and supercomputers domestically, insulating NVIDIA from export bans. Partnerships with TSMC, Foxconn, and Wistron are creating “gigawatt AI factories” capable of delivering 15 exaFLOPs of compute power (14x more than current H100 chips). This isn't just about replacing lost China sales; it's about owning the next-gen AI infrastructure stack.
Margin Resilience: A Test of NVIDIA's Pricing Power
Analysts initially feared that inventory charges and lower H20 sales would crater margins. Instead, NVIDIA's adjusted EPS rose 461% year-over-year to $6.12, proving its ability to offset losses with soaring data center sales. The $500 billion AI infrastructure pipeline—fueled by hyperscalers and governments racing to build AI supercomputers—will further bolster margins. By 2028, each gigawatt data center could generate $40–50 billion in recurring revenue for NVIDIA, a testament to its software-driven ecosystem.
Why the Stock Is Undervalued Now
NVIDIA's stock has dipped 23% from its 2025 peak, but this creates a buying opportunity. Analysts project 51% earnings growth in fiscal 2026, with a median price target of $175—52% above current levels. The ten-for-one stock split and dividend hike signal confidence in long-term cash flows. Meanwhile, competitors like Huawei are years behind NVIDIA's AI roadmap (Blackwell Ultra, Rubin series chips), and U.S. trade policies are now favoring reshoring, not restriction.
The Bottom Line: Buy the Dip
NVIDIA's Q1 results are a temporary stumble in a marathon. The company is doubling down on AI's future with U.S.-based factories, software ecosystems (Blackwell, NIM), and partnerships that no rival can match. While China's market remains closed, the global AI infrastructure boom—projected to hit $1 trillion in data center spending by 2028—is NVIDIA's to dominate. Investors who ignore the noise and focus on the $500 billion AI opportunity ahead will be rewarded. This is a buy.
Comments
No comments yet