Nvidia's Blowout Earnings Just Lit Up The Next AI Hardware Trade

$NVIDIA (NVDA.US)$ 's latest earnings did more than confirm AI demand. They gave investors a new map for the semiconductor trade.

The company reported Q1 FY2027 revenue of $81.6 billion, up 85% year over year, with Data Center revenue of $75.2 billion, up 92%. Q2 guidance was even stronger at $91.0 billion, plus or minus 2%, and the outlook again assumes no Data Center compute revenue from China.

That means the AI infrastructure trade is broadening. Nvidia is still the anchor, but the market is starting to reprice the full hardware stack around it: CPUs, DRAM, NAND, HDDs, foundry, equipment, substrates, sockets, controllers and networking.

CPU is back in the AI story

The biggest new signal from the call was Vera CPU. $NVIDIA (NVDA.US)$ said Vera opens a new $200 billion TAM and that it has visibility to nearly $20 billion of total CPU revenue this year. Management later clarified that this figure refers to standalone Vera CPU revenue, separate from Vera bundled inside Vera Rubin systems.

This matters for the CPU basket, including $NVIDIA (NVDA.US)$ , $Intel (INTC.US)$ , $Advanced Micro Devices (AMD.US)$ , $Arm Holdings (ARM.US)$ and now $Qualcomm (QCOM.US)$ . The old AI trade was mostly about GPUs doing the heavy thinking. The new AI agent trade also needs CPUs for orchestration, memory management, tool use, storage, security and confidential computing.

Memory and storage become bottleneck assets

AI factories are not just accelerator farms. They need memory and storage to feed models, maintain context, store data and support inference at scale.

This is why memory and storage names can be repriced after Nvidia's report. The stronger the AI factory buildout, the more investors will look for the next scarcity layer beyond GPUs. HBM remains the most obvious memory bottleneck, but DRAM, NAND and high-capacity HDDs also become more important as AI workloads move from training into inference and agentic use cases.

Networking is becoming the second engine

Nvidia's Data Center networking revenue reached $14.8 billion in Q1, up 199% year over year and 35% sequentially. Annualized, that is close to a $60 billion revenue run rate. Management also said Spectrum-X is now larger than all Ethernet network peers combined, while InfiniBand grew more than 4x year over year.

That changes the way investors should think about AI infrastructure. GPUs are still the engine, but networking is the nervous system. The bigger AI clusters become, the more valuable NVLink, Ethernet, InfiniBand, DPUs, switches, optics, PCB materials, substrates and controller chips become. This is why the AI trade is spreading into networking and interconnect names, not just accelerator names.

Foundry and equipment remain the toll roads

The foundry and equipment buckets, including $Taiwan Semiconductor (TSM.US)$ , $ASML Holding (ASML.US)$ , $Applied Materials (AMAT.US)$ , and $KLA Corp (KLAC.US)$ , remain the toll roads of the AI buildout. Nvidia's Q2 guide of $91 billion signals that Blackwell demand is still scaling, while Vera Rubin production shipments are on track to start in the second half of this year, beginning in Q3.

This supports demand visibility for advanced logic, CoWoS-style packaging, HBM capacity, inspection, deposition, etch, lithography and test. The key point is not that every equipment stock moves the same way. It is that Nvidia's growth curve keeps validating a multi-year semiconductor capacity cycle.

The next catalyst

The next catalyst is the June AI hardware calendar.

– First, Jensen Huang will give the GTC Taipei keynote at COMPUTEX on June 1. Investors will watch for updates on Rubin, AI factories, Taiwan supply chain partners, networking and robotics.

– Second, $Advanced Micro Devices (AMD.US)$ remains the key challenger to Nvidia's AI stack. Its next major AI event is in July, but its June conference appearance could still offer clues on MI400/MI450, Helios racks and customer demand.

– Third, $Qualcomm (QCOM.US)$ Investor Day on June 24 could be the next CPU catalyst. The company says it will discuss AI opportunities across gigawatt-scale data centers, industrial AI, physical AI, personal AI for agentic workloads and 6G.

Summary

Nvidia's earnings did not simply say AI demand is strong. They showed where the next AI infrastructure dollars may flow.

GPUs remain the center, but the next repricing wave is likely to focus on CPUs, memory, storage, networking, substrates, foundry and equipment. For investors, the better question is no longer just "who sells the GPU?" It is "who owns the next bottleneck in the AI factory?"

312K Views