Mountain View, CA, USA

Alphabet

Google designs the TPU line and operates it at pod and multi-pod scale. TPU FLOPS are almost entirely absent from public rankings.

Systems this company appears in

Systems: 15
Aggregate Rmax: none published
Aggregate power: 0 kW
First appearance: 2020
Most recent: 2026-04
Concentration: -

Aggregate Rmax is the sum of the systems this company supplies any part of. It is not revenue and not attributable share: an integrator and a memory vendor in the same machine both count its full FLOPS here. Attribution weighted by role is Phase 4.

Role breakdown

Role	Systems	Rmax in those systems
Integrator	7	none published
Accelerator	14	none published
CPU	2	none published
Interconnect	8	none published

Appearances

System	Supplied as	Shipped under	Rmax	Rank
Google TPU v6e (Trillium) pod USA · 2024-12	Accelerator, Interconnect	Alphabet	-	unranked
Google Council Bluffs East TPU Campus USA · 2024-11	Accelerator, Integrator	Alphabet	-	unranked
Google Pryor North TPU Campus USA · 2024-07	Accelerator, Integrator	Alphabet	-	unranked
Google Columbus TPU Campus USA · 2024-04	Accelerator, Integrator	Alphabet	-	unranked
Google New Albany TPU Campus USA · 2024-04	Accelerator, Integrator	Alphabet	-	unranked
Google TPU v4 ML Hub (Oklahoma) USA · 2022-05	Accelerator, Integrator	Alphabet	-	unranked
Google TPU v4 supercomputer USA · 2020	Accelerator, Interconnect	Alphabet	-	unranked
Google TPU v5p pod USA ·	Accelerator, Interconnect	Alphabet	-	unranked
Google TPU7x (Ironwood) pod USA ·	Accelerator, Interconnect	Alphabet	-	unranked
Google TPU (Ironwood) capacity for Anthropic USA ·	Accelerator, Integrator	Alphabet	-	unranked
Google India AI Hub (Vizag) India ·	Integrator	Alphabet	-	unranked
Google TPU v5e pod USA ·	Accelerator, Interconnect	Alphabet	-	unranked
Google TPU v3 pod USA ·	Accelerator, Interconnect	Alphabet	-	unranked
Google TPU 8t superpod USA ·	Accelerator, Interconnect, CPU	Alphabet	-	unranked
Google TPU 8i pod USA ·	Accelerator, Interconnect, CPU	Alphabet	-	unranked

Parts supplied

Part	Category	Generation	Released
Google TPU 8t	Accelerators	TPU 8	2026
Google TPU 8i	Accelerators	TPU 8	2026
Google TPU7x (Ironwood)	Accelerators	TPU7x	2025
Google TPU v6e (Trillium)	Accelerators	v6e	2024
Google TPU v5p	Accelerators	TPU v5	2023
Google TPU v5e	Accelerators	v5e	2023
Google TPU v4	Accelerators	TPU v4	2020
Google TPU inter-chip interconnect (3D torus)	Interconnect	ICI	-

AI datacenters operated

System	Status	Accelerators	Stated power (MW)
Google TPU v5p pod	operational	8,960	-
Google TPU7x (Ironwood) pod	operational	9,216	-
Google TPU v4 supercomputer	operational	4,096	-
Google TPU (Ironwood) capacity for Anthropic	under construction	1,000,000	1,000
Google TPU v4 ML Hub (Oklahoma)	operational	-	-
Google Pryor North TPU Campus	operational	101,400	368
Google Columbus TPU Campus	operational	465,700	303
Google New Albany TPU Campus	operational	576,400	333
Google Council Bluffs East TPU Campus	operational	305,600	237
Google TPU v3 pod	operational	1,024	-
Google TPU 8t superpod	planned	9,600	-
Google TPU 8i pod	planned	1,024	-

Figures come from the AI datacenter metrics; blank means none recorded.

Compute deals

Date	Role	Counterparty	Type	MW	Value ($bn)	Source
2025-10-23	Seller	Anthropic	Cloud contract	1,000	-	anthropic.com nasdaq.com
2025-08-21	Seller	Meta Platforms	Cloud contract	-	10	cnbc.com datacentrereview.com

Headline values are as announced and are not additive. See all compute deals.

Quantum computers operated

Google Sycamore (Superconducting, operational, 53 qubits)
Google Willow (Superconducting, operational, 105 qubits)

Sites operated

Site	Location	Cooling	Commissioned
Google Cloud (undisclosed location)	USA	-	2023
Google Mayes County Data Center	Mayes County, OK, USA	-	-
Visakhapatnam Google AI Hub Campus	Visakhapatnam, Andhra Pradesh, India	-	-
Google Columbus Data Center	Columbus, OH, USA	-	-
Google New Albany Data Center	New Albany, OH, USA	-	-
Google Council Bluffs East Data Center	Council Bluffs, IA, USA	-	-

Related news

2026-09-16 Emerald AI, Google and NVIDIA Launch Alliance to Advance Flexible AI Data Centers · NVIDIA Blog
NVIDIA, Google and startup Emerald AI founded the AI Energy Management Alliance to set technology-neutral standards letting AI data centers shift load, discharge storage or curtail draw on signal from the grid, instead of connecting as fixed always-on demand. It matters here because it is the two largest buyers of AI accelerators publicly treating grid power, not chip supply, as the binding constraint on how fast their own compute buildouts can scale, and because flexibility commitments like these are what utilities are starting to condition faster interconnection on.
2026-09-10 Google and Blackstone's TPU venture launches as Crux AI with a stated 2GW capacity target · Data Center Dynamics
TPUs sit in this graph as parts inside Google's own infrastructure; a separately capitalised operator selling TPU capacity puts the same accelerator on the far side of a company boundary, which is exactly the case entity resolution here has to get right rather than collapse into its parent. The capacity numbers are the company's own forward plans and an executive hire, not built floor space, so they establish intent and nothing more.
2025-11-06 Google's Ironwood (TPU v7) becomes generally available, compared directly to El Capitan · Google
Google's own launch materials put a full Ironwood pod at 42.5 exaflops of FP8 compute, described as more than 24x El Capitan's 1.7 exaflops figure, though the two numbers measure different things, dense FP8 inference throughput versus double precision HPL, and are not a like for like comparison. It is nonetheless the most direct public benchmark claim yet tying a hyperscaler's own accelerator generation to a specific system already in this dataset.

All news →