Quick Read
Nvidia’s data center surged 92% to $75 billion while AMD’s grew 107%, pushing hyperscalers toward dual sourcing of AI chips.
Nvidia’s Vera Rubin ships Q3 promising 35x higher inference throughput as AMD’s Helios launches backed by a 2-gigawatt Anthropic commitment.
Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.
NVIDIA (NASDAQ:NVDA) and AMD (NASDAQ:AMD) closed AI-heavy quarters that reset the champion-challenger story. Nvidia’s Q1 FY2027 landed at $81.61 billion in revenue. AMD’s Q2 2026 hit $11.54 billion. The debate: is Nvidia running an infrastructural monopoly, or is AMD forcing hyperscalers into value-driven dual sourcing?
24/7 Wall St Data Center Carries Both, But for Different Reasons
Nvidia’s data center segment reached $75.25 billion, up 92% year-over-year, with networking alone up 199% as InfiniBand and Spectrum-X pulled NVLink rack-scale deployments through. Jensen Huang called this buildout “the largest infrastructure expansion in human history”. AWS plans to add more than 1 million Blackwell and Rubin GPUs beginning this year. All of that hardware still has to be powered, cooled, and networked by somebody, which is why we pulled together seven non-chipmaker suppliers riding the same buildout in a free report.
AMD’s data center story differs. Revenue hit $6.72 billion, up 107%, now representing 58% of total revenue. Instinct sales more than doubled on the MI350 ramp, and EPYC posted its fifth consecutive quarter of record server CPU revenue. Gaming fell 31% on soft semi-custom demand, a reminder that AMD still carries meaningful non-AI exposure.
Infrastructural Monopoly Meets Value-Driven Dual Sourcing
Nvidia’s pitch is vertical integration you cannot easily unbundle. Huang put it plainly: “Customers do not buy GPUs. They build AI factories.” The GB300 delivered a 2.7x increase in throughput and a 60% reduction in cost per token versus six months earlier. Management flagged visibility to $1 trillion in Blackwell and Rubin revenue through calendar 2027.
Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.
AMD sells the opposite thesis: open ecosystem, comparable performance, better unit economics. Lisa Su said Helios delivers “up to 15% more throughput at the same rack power, and up to 30% more tokens per dollar than the competition.” Anthropic committed to up to two gigawatts of MI450 series GPUs in Helios, and Microsoft plans to run Helios at scale on Azure.
Lens
Nvidia
AMD
Core Bet
Full-stack CUDA platform
Open Helios rackscale
Non-GAAP Gross Margin
75.0%
56%
Anchor AI Customers
Every hyperscaler, OpenAI
Anthropic, OpenAI, Meta
Rubin Ramp Meets Helios Ramp in the Same Quarter
Both companies push new platforms into the same window. Nvidia begins production shipments of Vera Rubin in the second half of this year, starting in Q3, with claims of up to 35x higher inference throughput versus Blackwell. AMD’s Helios begins initial shipments this quarter and ramps through Q4 2026 and into 2027. The key question: do hyperscalers genuinely split orders or trial AMD at the margins?
Nvidia guided Q2 to $91.0 billion. AMD guided Q3 to approximately $13 billion, up roughly 41% year-over-year.
Why Nvidia Leads for Scale, AMD for Optionality
This quarter validates both theses. Nvidia at a 33x trailing P/E against a 63% profit margin looks reasonable if the Rubin roadmap holds. For investors wanting the compounding infrastructure story with dividends and an $80 billion buyback behind it, Nvidia is the cleaner expression. AMD at a 120x trailing P/E only works if Helios converts pilot deployments into gigawatt orders on schedule. That is a turnaround-style setup for investors comfortable with variance. AMD’s setup weakens if Gaming stays soft and HBM allocation slips. Neither looks compelling if hyperscaler capex growth cools before Rubin ships in volume.
Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now.
Contact editorial@247wallst.com for any questions or corrections.