- Nvidia posted Q1 revenue of $81.6 billion — up 85% year-over-year and beating consensus by 3.4% — with net income of $58.3 billion, more than three times the year-ago result and 36.5% above estimates.
- Networking hardware revenue tripled year-over-year to a record $14.8 billion; CEO Jensen Huang declared “the era of agentic AI is here” and said demand has “gone parabolic.”
- Nvidia announced an $80 billion share buyback and hiked its quarterly dividend 25x — from $0.01 to $0.25 — committing to return 50% of free cash flow to shareholders this year.
- China remains zero revenue for data-center chips as Beijing has not approved H200 sales; Huang sees China as a potential $50 billion annual opportunity, and his last-minute inclusion on Trump’s Beijing summit flight raised speculation the topic was raised.
What Happened?
Nvidia reported first-quarter results that were extraordinary across every metric. Revenue of $81.6 billion grew 85% from a year ago, powered by data-center GPUs, networking, and inference systems as hyperscalers and enterprise customers raced to secure AI compute. Net income of $58.3 billion crushed the $42.9 billion consensus estimate. Networking hardware alone tripled to $14.8 billion, reflecting the explosion of large-scale AI clusters requiring massive high-speed interconnects. Huang framed the moment on the analyst call: agentic AI has arrived — autonomous AI systems operating on behalf of users — and demand for the compute required to run them is outpacing supply. The company announced an $80 billion buyback and a quarterly dividend hike from $0.01 to $0.25, signaling confidence in sustained cash generation at this scale.
Why It Matters?
Nvidia’s results are a real-time referendum on whether the AI capex supercycle is real. The answer this quarter: unambiguously real, and still accelerating. With Nvidia now the largest publicly listed company in the world at nearly $5.5 trillion market cap, investors have priced in extraordinary durability — and this quarter validated that bet. The competitive landscape is shifting, with Google’s TPU partnership with Blackstone, AWS custom silicon at ~$50 billion annualized, and Cerebras’s $5.6 billion IPO all signaling rivals are finally investing at scale to challenge Nvidia. But none are close to displacing it. Huang also flagged that the enterprise and neo-cloud segment is expected to grow faster than hyperscalers going forward, which would further broaden Nvidia’s customer base and reduce revenue concentration.
What’s Next?
The Vera Rubin GPU platform is expected to begin shipping later this year, and Huang said he expects it to outperform Grace Blackwell — a high bar given how ubiquitous that generation has already become. China is the swing factor: Huang pegs the market at $50 billion annually, and zero revenue today means any Beijing approval of H200 sales would be an instant material upside. Watch whether Trump-Xi diplomacy — and Huang’s presence on that flight — produces regulatory movement. The next earnings report in August will test whether agentic AI demand sustains its parabolic trajectory or shows first signs of digestion.
Source: The Wall Street Journal
