NvidiaNVDA

NASDAQ · USA · Segment 1 of 8 in the AI & digital infrastructure value chain: Chips & accelerators

Nvidia sells the accelerators most large AI models are trained on, from H100 through B200 to Rubin. Harder to replace is CUDA, the software layer developers have standardised on. NVLink and the Mellanox networking business tie whole racks together.

Chips & accelerators: Designs and builds the processors, accelerators and memory that train and run AI.

Key metrics

Market cap

$5.4T

Price

$208.48

Today

−2.9%

Return YTD

+15.1%

Return 1Y

+22.7%

P/E

34.0

P/S

21.4

P/B

27.7

Data as of 2026-08-25 · updates daily after U.S. close

Theme exposure

AI hardwareData center exposureGPU training & inferenceCUDA ecosystemDatacenter accelerators

In the value chain

Products & services

H100/B200/Rubin AI GPUs
CUDA software platform
NVLink & networking (Mellanox)

Competitors

AMD
Broadcom (custom chips)
Google's own TPUs

Risks

High customer concentration among a few hyperscalers
Export restrictions against China
Competition from customers' own chips

Supply chain

Supplies →

MicrosoftMSFT AmazonAMZN AlphabetGOOGL Meta PlatformsMETA OracleORCL CoreWeaveCRWV TeslaTSLA HPEHPE IBMIBM AccentureACN Dell TechnologiesDELL Super MicroSMCI

← Suppliers

MicronMU Arm HoldingsARM SynopsysSNPS Cadence Design SystemsCDNS AmphenolAPH Monolithic PowerMPWR

Full supply chain for NVDA →

Same segment: Chips & accelerators

BroadcomAVGO MicronMU AMDAMD IntelINTC Arm HoldingsARM Texas InstrumentsTXN SeagateSTX MarvellMRVL

Compare NVDA

NVDA vs AVGO NVDA vs MU NVDA vs AMD NVDA vs INTC NVDA vs ARM NVDA vs TXN

See where Nvidia sits in the full AI & digital infrastructure map: every segment from upstream suppliers to end markets, with daily metrics.Open the AI value chain →