AI accelerator chip: Use case
Use cases
Each use, its share of demand and volume, how critical the product is to it and why, and the price paid
| Use case | Volume | Unit | Criticality | Reason | Year | Source |
|---|---|---|---|---|---|---|
| Training large AI models in data centres | 3.76 | million NVIDIA data-centre GPUs shipped in 2023 (training and inference together) | Essential | Training a large model needs thousands of accelerators working together; general-purpose processors are far too slow. | 2023 | [45] |
| Serving trained models to users (inference) in cloud data centres | – | – | Important | Small models can run on ordinary processors, but large ones need accelerators; operational time for inference is assumed at 40% against 80% for training. | 2024 | [22] |
| Public AI compute for research and start-ups, India | 38,000 | GPUs onboarded on the IndiaAI compute portal | Important | The state pools GPUs to rent out at an affordable rate, because buying chips is beyond most start-ups. | 2026 | [23] |
as of 2026-11-16. No source gives demand shares by use. TechInsights counts only NVIDIA units and the 2023 figure includes all uses [45].
Industries that buy it
The industries and kinds of firm that buy it, for which use case, and how critical it is to them
Used by 4- Hyperscale cloud providers Training and serving models in cloud data centres · criticality essential NVIDIA names Amazon Web Services, Google, Meta, Microsoft, Oracle and others among the organisations expected to adopt Blackwell [14]; Google also builds its own TPUs [39].
- Foundation model developers Training and running frontier models · criticality essential Anthropic has announced plans to use up to a million Google TPUs [39]; OpenAI is among Blackwell adopters [14].
- Hyperscale clouds and GPU cloud operators Renting GPU time to firms that cannot buy chips · criticality important GPU cloud operators buy the same chips and resell their time; no share found.
- Government IT and digital public infrastructure providers Pooled public AI compute (India) · criticality important The IndiaAI Mission has onboarded over 38,000 GPUs on a common compute portal [23].
as of 2026-11-16
What it competes with
The products and technologies that compete with it in each use case
Competes with- Processor chip Inference and on-device AI General-purpose processors, and the neural-processing blocks built into them, can run small models, at lower speed and efficiency. No source opened measures how far they can replace accelerators.
- Nm Neuromorphic computing Low-power inference at the edge An emerging approach that copies the brain's way of computing. No source opened compares its cost or speed with accelerators.
as of 2026-11-16
How it is used
Lifetime, use per year and utilisation, by variant
| Measure | Value | Unit | Variant | Place | Year | Source |
|---|---|---|---|---|---|---|
| Lifespan assumed for AI machines in a Google life-cycle study | 6 | years | Custom accelerator | World | 2025 | [3] |
| Average lifespan assumed for the A100 GPU in a cradle-to-grave study | 3 | years | Data-centre GPU | World | 2025 | [17] |
| Operational time of servers doing training (LBNL model assumption) | 80 | % of the year | Data-centre GPU | United States of America | 2024 | [22] |
| Operational time of servers doing inference (LBNL model assumption) | 40 | % of the year | Data-centre GPU | United States of America | 2024 | [22] |
| Average power of an 8-H100 node in saturated training, as a share of rated power | 78 | % (about) | Data-centre GPU | United States of America | 2024 | [22] |
| Average power usage effectiveness (PUE) of US data centres | 1.4 | ratio of facility to IT energy | Data-centre GPU | United States of America | 2023 | [22] |
| Share of average power that Google's TPU machines draw when idle | 60 | % | Custom accelerator | World | 2025 | [3] |
| Ratio of TDP to actual average power for TPUs (TDP overstates real use by this factor) | 2 to 6 | times | Custom accelerator | World | 2025 | [3] |
as of 2026-11-16. Hours in a year are 8,760. Estimates of utilisation and lifespan differ by study; the energy figure on the Footprint tab uses the LBNL operational time for training and the LBNL PUE.
Prices
Unit prices by market or use, with year and source
| Market | Value | Unit | Year | Source |
|---|---|---|---|---|
| NVIDIA H100, low end of analysts' estimates | 25,000 | USD per chip | 2024 | [46] |
| NVIDIA H100, high end of analysts' estimates | 40,000 | USD per chip | 2024 | [46] |
| NVIDIA Blackwell, low end of the chief executive's statement | 30,000 | USD per chip | 2024 | [46] |
| NVIDIA Blackwell, high end of the chief executive's statement | 40,000 | USD per chip | 2024 | [46] |
| Cost to make the GPUs of one DGX H100, per GPU (USD 26,600 divided by eight, calculated; not a price) | 3325 | USD per GPU | 2023 | [47] |
as of 2026-11-16. Unit values from trade data are calculated by the site and are not researched here.
Sources
- Life-cycle emissions of AI hardware: a cradle-to-grave approach and generational trends Schneider et al. (Google), arXiv preprint 2502.01671, 2025
- NVIDIA Blackwell Platform Arrives to Power a New Era of Computing NVIDIA investor relations, 2024
- More than Carbon: Cradle-to-Grave environmental impacts of GenAI training on the Nvidia A100 GPU (arXiv:2509.00093v3) arXiv preprint, 2025
- 2024 United States Data Center Energy Usage Report (LBNL-2001637) Lawrence Berkeley National Laboratory, 2024
- IndiaAI Mission expands AI ecosystem with affordable compute and startup support Digital India, Government of India, 2026
- Google's Ironwood TPUs represent a bigger threat than Nvidia would have you believe The Register, 2025
- Nvidia GPU shipments totaled 3.76m in 2023, equating to a 98% market share: report (TechInsights) Data Center Dynamics
- Nvidia's latest AI chip will cost more than $30,000, CEO says CNBC via NBC San Diego, 2024
- AI Server System BOM & Rack Cost: Historical Data & Chart (newest print March 2026) Silicon Analysts, 2026