AI Data Center PartsSearch
Home / Parts / Compute silicon & packaging / Merchant GPUs / accelerators
1.01

Who makes merchant gpus / accelerators for AI data centers

GPUs and other accelerator chips designed by merchant vendors and sold to any buyer, rather than designed in-house by a single hyperscaler; the workhorse silicon for AI training and inference.

Merchant accelerators are the chips that do the actual matrix math in AI training and inference — bought off a price list (allocation permitting) rather than designed for one company's fleet alone. They sit downstream of foundry (the advanced-node foundry step) and advanced packaging, which combine the compute die with HBM stacks onto a substrate, and upstream of the server board and tray assembly in Layer 4, where they're paired with server CPUs, NICs, and cooling. Because merchant GPUs are sold to every buyer from a single enterprise pilot through a gigawatt-class campus, this is the part of the catalog investors search first when asking who benefits from AI buildouts. An investor should watch supplier concentration — a small number of vendors set the pace for each new architecture generation — alongside qualification cycles, since a new generation typically requires re-validation across the software stack, board designs, and cooling before hyperscalers will deploy it at scale. Pricing has moved less on open competition than on allocation: demand has outrun available packaging and wafer capacity for multiple generations running, which has kept lead times elongated and kept the vendors that can secure capacity in a stronger negotiating position than a normal semiconductor upcycle would imply.

AI delta: AI training and inference workloads are the primary demand driver for merchant accelerators, which is why unit volumes and generational upgrade cadence have both accelerated well beyond a typical chip replacement cycle.

Per rack: 72 GPUs per NVL72 rack (Nvidia rack specification, via taxonomy v0.1 Appendix A).

Appears at:Enterprise / small coloLarge colo / hyperscale buildingAI campusGW-class campus· cooling:any — Same die designs ship into an enterprise pilot rack and a gigawatt campus alike; only unit count and generation mix change with scale.

Companies on this part (3)

CompanyRoleOwnershipTickerPrice1dConfidenceNote
AMDmanufacturerpublicAMD477 USD-1.7%inferredInstinct MI350/MI400
IntelmanufacturerpublicINTC92.09 USD+2.4%inferredGaudi 3
NvidiamanufacturerpublicNVDA228 USD+3.8%inferredBlackwell Ultra

Who competes here

The same table read the other way: every company above competes for merchant gpus / accelerators sockets. Confidence tags matter — an inferred placement is a lead, not a confirmed rivalry.

Nearby parts

Watch

Understanding NVIDIA GPU Hardware as a CUDA C Programmer | Episode 2: GPU Compute Architecture
Tushar Gautam — walks through the compute architecture inside a merchant Nvidia GPU (SMs, warps, memory hierarchy)
Click to load from YouTube · open on YouTube ↗

Independent videos — no affiliation; nothing loads from YouTube until you click.

Latest feed

Sources: placements without a source URL come from the taxonomy seed and are tagged accordingly. How this data is builtMarkdown · JSON