NVIDIA (NASDAQ: NVDA) and AMD (NASDAQ: AMD) both reported earnings in May, and the results paint a striking picture of where the AI hardware market stands today.
NVIDIA posted a massive $81.6 billion quarter driven by its Blackwell platform, while AMD reported $10.25 billion, with a landmark Meta partnership beginning to reshape its data center narrative in meaningful ways.
NVIDIA’s Data Center segment alone reached $75.25 billion, representing 92% year-over-year growth, with networking revenue hitting $14.8 billion, a staggering 199% increase from the prior year period.
That networking revenue figure is larger than AMD’s entire data center business, underscoring just how dominant NVIDIA’s infrastructure position has become across the global AI buildout.
CEO Jensen Huang framed the moment directly: “The buildout of AI factories, the largest infrastructure expansion in human history, is accelerating at extraordinary speed.”
AMD’s quarter was smaller but showed genuine acceleration, with Data Center revenue reaching $5.78 billion, up 57% year over year, as customer engagement around its next-generation products continued to build.
CEO Lisa Su told investors that “Customer engagement around MI450 Series and Helios is strengthening, with leading customer forecasts exceeding our initial expectations,” signaling growing confidence in AMD’s accelerator roadmap.
A 6 gigawatt Meta deployment, starting with a custom MI450 design, gives AMD something it has long lacked: a flagship hyperscaler willing to make a serious long-term bet on its hardware.
The valuation contrast between the two companies is striking, with AMD trading at 68 times forward earnings despite 38% revenue growth, while NVIDIA carries a forward multiple of just 24 times earnings despite growing at 85% with 75% gross margins.
NVIDIA’s Q2 guidance points to $91 billion in revenue, compared to AMD’s guidance of approximately $11.2 billion, and NVIDIA’s non-GAAP gross margin of 75% dwarfs AMD’s 55% figure by a considerable margin.
NVIDIA has also locked in $119 billion in supply commitments, giving it a significant manufacturing security advantage, while AMD is still in negotiations with Samsung over HBM4 memory capacity for the MI455X.
NVIDIA’s competitive platform continues to widen, with Vera Rubin pairing a custom CPU with Rubin GPUs, Dynamo 1.0 reportedly lifting Blackwell inference performance by up to 7 times, and key partnerships with Marvell, Corning, Lumentum, and Coherent securing the optics layer.
NVIDIA shares have pulled back 8.2% since its May 20 earnings report, while AMD has surged 37.5% over the same period, making NVIDIA’s near-term setup look comparatively more attractive on a risk-adjusted basis.
NVIDIA also authorized an $80 billion fresh share buyback and raised its dividend by 25 times, reflecting extraordinary confidence in the durability of its cash generation and long-term earnings power.
Key variables for the second half include whether AMD’s MI450 customer forecasts convert into firm orders visible in Q3 results, and whether any shift in U.S.-China trade policy unlocks data center compute revenue that NVIDIA’s current Q2 guidance explicitly excludes.
