AI Chip GPUs

Data Center GPU/Accelerator Shipments (Estimate)

Sterling estimate: NVIDIA, AMD and Intel do not publish unit shipments, and no tracker prints these counts. Quarterly shipments of AI accelerators: the specialized chips (GPUs, TPUs, custom ASICs) that train and run AI models in data centers. NVIDIA captures over 80% of unit volume; AMD, Intel, and Google's TPUs trail well behind. Figures in thousands of units.

AI Accelerator Architecture Comparison

Side-by-side comparison of AI training and inference accelerators: specs, memory, performance, and pricing

NVIDIA logoH100 SXMHopper
Node
4nm (TSMC)
Transistors
80B
Memory
80GB HBM3
Mem BW
3.35 TB/s
FP16 TFLOPS
1,979
FP8 TFLOPS
3,958
TDP
700W
Launched
2023-Q1
Interconnect
NVLink 4.0 (900 GB/s)
Price Est.
$25,000–$40,000
NVIDIA logoB200Blackwell
Node
4nm (TSMC)
Transistors
208B
Memory
192GB HBM3E
Mem BW
8 TB/s
FP16 TFLOPS
5,000
FP8 TFLOPS
10,000
TDP
1000W
Launched
2024-Q4
Interconnect
NVLink 5.0 (1.8 TB/s)
Price Est.
$30,000–$50,000
NVIDIA logoGB200 NVL72Blackwell
Node
4nm (TSMC)
Transistors
208B (x72)
Memory
13.5TB HBM3E
Mem BW
576 TB/s
FP16 TFLOPS
162K
FP8 TFLOPS
1440K
TDP
120kW
Launched
2025-Q1
Interconnect
NVLink 5.0 (full rack)
Price Est.
$2M–$3M (rack)
AMD logoMI300XCDNA 3
Node
5nm/6nm (TSMC)
Transistors
153B
Memory
192GB HBM3
Mem BW
5.3 TB/s
FP16 TFLOPS
1,307
FP8 TFLOPS
2,614
TDP
750W
Launched
2023-Q4
Interconnect
Infinity Fabric (896 GB/s)
Price Est.
$10,000–$15,000
AMD logoMI325XCDNA 3.5
Node
5nm/6nm (TSMC)
Transistors
153B
Memory
256GB HBM3E
Mem BW
6 TB/s
FP16 TFLOPS
1,307
FP8 TFLOPS
2,614
TDP
750W
Launched
2024-Q4
Interconnect
Infinity Fabric (896 GB/s)
Price Est.
$15,000–$20,000
AMD logoMI350XCDNA 4
Node
3nm (TSMC)
Transistors
185B
Memory
288GB HBM3E
Mem BW
8 TB/s
FP16 TFLOPS
5,000
FP8 TFLOPS
10,000
TDP
1000W
Launched
2025-Q2
Interconnect
UALink / Infinity Fabric 4.0
Price Est.
TBD
Intel logoGaudi 2Gaudi
Node
7nm (TSMC)
Transistors
~25B
Memory
96GB HBM2E
Mem BW
2.46 TB/s
FP16 TFLOPS
432
FP8 TFLOPS
864
TDP
600W
Launched
2023-Q2
Interconnect
24x 100GbE RDMA
Price Est.
$12,000–$15,000
Intel logoGaudi 3Gaudi
Node
5nm (TSMC)
Transistors
~50B
Memory
128GB HBM2E
Mem BW
3.7 TB/s
FP16 TFLOPS
1,835
FP8 TFLOPS
3,670
TDP
900W
Launched
2024-Q4
Interconnect
24x 200GbE RDMA
Price Est.
$15,000–$20,000
Broadcom logoGoogle TPU v5eCustom ASIC
Node
7nm (TSMC)
Transistors
~20B
Memory
16GB HBM2E
Mem BW
820 GB/s
FP16 TFLOPS
197
FP8 TFLOPS
394
TDP
200W
Launched
2023-Q3
Interconnect
ICI (Inter-Chip Interconnect)
Price Est.
Cloud-only
Broadcom logoGoogle TPU v6e (Trillium)Custom ASIC
Node
5nm (TSMC)
Transistors
~35B
Memory
32GB HBM3
Mem BW
1.6 TB/s
FP16 TFLOPS
460
FP8 TFLOPS
920
TDP
300W
Launched
2024-Q4
Interconnect
ICI v2
Price Est.
Cloud-only
ProductArchitectureNodeTransistorsMemoryMem BWFP16 TFLOPSFP8 TFLOPSTDPInterconnectLaunchedPrice Est.
NVIDIA logoH100 SXM
Hopper4nm (TSMC)80B80GB HBM33.35 TB/s1,9793,958700WNVLink 4.0 (900 GB/s)2023-Q1$25,000–$40,000
NVIDIA logoB200
Blackwell4nm (TSMC)208B192GB HBM3E8 TB/s5,00010,0001000WNVLink 5.0 (1.8 TB/s)2024-Q4$30,000–$50,000
NVIDIA logoGB200 NVL72
Blackwell4nm (TSMC)208B (x72)13.5TB HBM3E576 TB/s162K1440K120kWNVLink 5.0 (full rack)2025-Q1$2M–$3M (rack)
AMD logoMI300X
CDNA 35nm/6nm (TSMC)153B192GB HBM35.3 TB/s1,3072,614750WInfinity Fabric (896 GB/s)2023-Q4$10,000–$15,000
AMD logoMI325X
CDNA 3.55nm/6nm (TSMC)153B256GB HBM3E6 TB/s1,3072,614750WInfinity Fabric (896 GB/s)2024-Q4$15,000–$20,000
AMD logoMI350X
CDNA 43nm (TSMC)185B288GB HBM3E8 TB/s5,00010,0001000WUALink / Infinity Fabric 4.02025-Q2TBD
Intel logoGaudi 2
Gaudi7nm (TSMC)~25B96GB HBM2E2.46 TB/s432864600W24x 100GbE RDMA2023-Q2$12,000–$15,000
Intel logoGaudi 3
Gaudi5nm (TSMC)~50B128GB HBM2E3.7 TB/s1,8353,670900W24x 200GbE RDMA2024-Q4$15,000–$20,000
Broadcom logoGoogle TPU v5e
Custom ASIC7nm (TSMC)~20B16GB HBM2E820 GB/s197394200WICI (Inter-Chip Interconnect)2023-Q3Cloud-only
Broadcom logoGoogle TPU v6e (Trillium)
Custom ASIC5nm (TSMC)~35B32GB HBM31.6 TB/s460920300WICI v22024-Q4Cloud-only

AI Chip Product Roadmap

Product availability timeline for AI accelerators across NVIDIA, AMD, Intel, and Broadcom/Google

Legacy
In Production
Announced
Roadmap
NVIDIA logoNVIDIA
  • A100

    7nm

    2020 Q2 to 2023 Q4
  • H100

    4nm

    2022 Q4 to 2025 Q2
  • H200

    4nm

    2024 Q1 to 2026 Q2
  • B100/B200

    4nm

    from 2024 Q4
  • GB200 NVL72

    4nm

    from 2025 Q1
  • Rubin (R100)

    3nm

    from 2026 Q1
  • Feynman

    Node TBD

    from 2028
AMD logoAMD
  • MI250X

    6nm

    2022 Q1 to 2024 Q2
  • MI300X

    5nm

    from 2023 Q4
  • MI325X

    5nm

    from 2024 Q4
  • MI350X

    3nm

    from 2025 Q2
  • MI400

    2nm

    from 2026 H2
Intel logoIntel
  • Gaudi 2

    7nm

    2023 Q2 to 2025 Q1
  • Gaudi 3

    5nm

    from 2024 Q3
  • Jaguar Shores

    Intel 18A

    from 2026 H1
Broadcom logoBroadcom
  • Google TPU v5e

    7nm

    from 2023 Q3
  • Google TPU v6e (Trillium)

    5nm

    from 2024 Q4
  • Google TPU v7 (Ironwood)

    Node not disclosed

    from 2025 Q4
  • Meta MTIA v2

    5nm

    from 2024 Q3

Data Center GPU + Accelerator Revenue by Vendor (Quarterly)

Quarterly data center AI accelerator revenue by vendor (NVDA, AMD, INTC, AVGO, MRVL). NVIDIA's Data Center line is by far the largest.

NVIDIA Data CenterQ2 2026: $89B
Other vendorsBroadcom AIAMDIntelMarvell

NVIDIA Data Center Compute vs Networking

NVIDIA-disclosed Data Center compute versus networking, only for fiscal quarters where both lines were printed. Networking (InfiniBand, Spectrum-X, NVLink fabric) was $14.8B in FY27 Q1, almost 20% of Data Center.

Hyperscaler Custom AI Accelerators

Company-stated custom accelerators (Google TPU, AWS Trainium, Azure Maia, Meta MTIA) plus Intel Foundry external customer revenue. This is a landscape table, not a shipment ranking. Foundry names appear only when the cited page states them.

TPU Ironwood (7th gen)

Owner
Google
What the owner has said
Google unveiled Ironwood, its 7th-generation TPU built for inference and training at scale, on April 9, 2025. TPU7x, the first Ironwood release, entered Preview on Google Cloud on November 24, 2025 and became generally available on March 31, 2026, with 192 GB of HBM per chip
Foundry (only if disclosed)
Not disclosed by Google; TSMC is widely assumed and is not treated as a Google disclosure here

Trainium2 / Trn2

Owner
Amazon Web Services
What the owner has said
AWS documents Trainium2 as the current-generation training chip for Amazon EC2 Trn2 instances and UltraServers, sold as an NVIDIA alternative for AWS-hosted training
Foundry (only if disclosed)
Not stated on the cited AWS product page

Azure Maia 200

Owner
Microsoft
What the owner has said
Microsoft introduced Maia 200, the successor to Maia 100, on January 26, 2026 as an inference accelerator with 216 GB of HBM3e. It is deployed in Azure's US Central region near Des Moines, Iowa, with US West 3 near Phoenix, Arizona, named as next, and serves models including OpenAI's GPT-5.2
Foundry (only if disclosed)
TSMC 3nm (stated by Microsoft)

MTIA

Owner
Meta
What the owner has said
Meta's engineering blog describes MTIA as a family of training and inference accelerators for ranking and recommendation, now in production inside Meta data centers
Foundry (only if disclosed)
Meta has publicly worked with Broadcom on custom silicon; the cited MTIA blog is about the chip, not a foundry share figure

Intel Foundry external AI silicon

Owner
Intel Foundry customers
What the owner has said
Intel's Q2 2026 10-Q: third-party foundry and assembly/test revenue was $293 million in the quarter (vs $22 million a year earlier). That is the disclosed external foundry print, not a named list of AI chips
Foundry (only if disclosed)
Intel Foundry
ProductOwnerWhat the owner has saidFoundry (only if disclosed)
TPU Ironwood (7th gen)GoogleGoogle unveiled Ironwood, its 7th-generation TPU built for inference and training at scale, on April 9, 2025. TPU7x, the first Ironwood release, entered Preview on Google Cloud on November 24, 2025 and became generally available on March 31, 2026, with 192 GB of HBM per chipNot disclosed by Google; TSMC is widely assumed and is not treated as a Google disclosure here
Trainium2 / Trn2Amazon Web ServicesAWS documents Trainium2 as the current-generation training chip for Amazon EC2 Trn2 instances and UltraServers, sold as an NVIDIA alternative for AWS-hosted trainingNot stated on the cited AWS product page
Azure Maia 200MicrosoftMicrosoft introduced Maia 200, the successor to Maia 100, on January 26, 2026 as an inference accelerator with 216 GB of HBM3e. It is deployed in Azure's US Central region near Des Moines, Iowa, with US West 3 near Phoenix, Arizona, named as next, and serves models including OpenAI's GPT-5.2TSMC 3nm (stated by Microsoft)
MTIAMetaMeta's engineering blog describes MTIA as a family of training and inference accelerators for ranking and recommendation, now in production inside Meta data centersMeta has publicly worked with Broadcom on custom silicon; the cited MTIA blog is about the chip, not a foundry share figure
Intel Foundry external AI siliconIntel Foundry customersIntel's Q2 2026 10-Q: third-party foundry and assembly/test revenue was $293 million in the quarter (vs $22 million a year earlier). That is the disclosed external foundry print, not a named list of AI chipsIntel Foundry

Ask Sterling

Register for a premium account to gain access to Sterling AI.

Get Started

Things you can ask Sterling:

Summarize Tesla's latest earnings reportWhy did NVIDIA's margins expand?Compare Apple vs Microsoft's cash flowWhat's driving EV industry growth?
Menu
Favorites
AI Chip GPUs: Market Data | Sterling