What is an AI datacenter?

An AI datacenter is a facility purpose-built or retrofitted for GPU and accelerator workloads — typically at rack densities above 30 kW, with liquid-cooling readiness and high-bandwidth fabric (InfiniBand NDR, NVLink, or Spectrum-X). At least 40% of IT load runs AI training or inference on GPUs such as NVIDIA H100, H200, B200, or GB200 NVL72.

Which datacenter has the most GPUs?

As of 2026, xAI Colossus in Memphis, Tennessee holds the title of the world's largest single-site GPU deployment, with more than 200,000 NVIDIA H100 and H200 GPUs in one InfiniBand domain. Meta Hyperion (Louisiana) and OpenAI Stargate Abilene (Texas) are scaling toward gigawatt-class deployments that will overtake it within the next 18 months.

What is the largest AI cluster?

xAI Colossus is the largest contiguous AI training cluster in 2026, exceeding 200,000 NVIDIA H100/H200 GPUs on a single InfiniBand NDR 400G fabric. Meta's Hyperion build and OpenAI's Stargate program are constructing multi-site fabrics that will exceed one million accelerators across coupled campuses by 2027.

Which GPU cloud is cheapest for H100?

On-demand: Together AI ($1.75/hr) and Voltage Park ($1.89/hr) lead. Reserved 1-year: Voltage Park ($1.20/hr) and Crusoe ($1.39/hr) are typically lowest. AWS p5 sits at the top of the price stack at ~$4.10/hr on-demand.

How is Datacenter.computer different from a datacenter directory?

Datacenter.computer is an agent-readable intelligence layer, not a directory. It normalizes pricing, capacity, power, fabric, and latency across 240+ facilities and 40+ operators into a single queryable graph for AI agents, traders, and procurement teams.

[DC.COMPUTER] Layer 5 · Colocation

AirTrunk

Every indexed AirTrunk datacenter — facility-level capacity, GPU fleet, and country coverage.

Datacenter Portfolio

[AI Summary]

AirTrunk runs 57 indexed datacenters across 14 countries, totaling 1,304 MW of IT capacity. The portfolio spans colocation tiers and deploys Mixed, MI300X, A100, H100, H200, L40S, B200.

Facilities

1,304 MW

Total IT

Countries

GPU SKUs

[Facilities · 57]

Facility	Operator	Location	GPU	MW	PUE
AirTrunk SYD1	AirTrunk	Sydney, Australia	Mixed	130	1.20
AirTrunk TYO1	AirTrunk	Inzai, Japan	Mixed	300	1.20
AirTrunk HKG1	AirTrunk	Tseung Kwan O, Hong Kong	Mixed	70	1.32
AirTrunk SINGAPORE52 South	AirTrunk	Singapore, Singapore	—	6	1.42
AirTrunk SINGAPORE56 Central	AirTrunk	Singapore, Singapore	—	30	1.32
AirTrunk SINGAPORE60	AirTrunk	Singapore, Singapore	MI300X	26	1.26
AirTrunk SINGAPORE64 South	AirTrunk	Singapore, Singapore	—	17	1.52
AirTrunk SINGAPORE68 West	AirTrunk	Singapore, Singapore	—	28	1.28
AirTrunk TOKYO1	AirTrunk	Tokyo, Japan	A100	35	1.33
AirTrunk TOKYO18 North	AirTrunk	Tokyo, Japan	—	6	1.38
AirTrunk OSAKA35 West	AirTrunk	Osaka, Japan	H100	25	1.27
AirTrunk HONG23 Park	AirTrunk	Hong Kong, Hong Kong	—	29	1.39
AirTrunk HONG27 Central	AirTrunk	Hong Kong, Hong Kong	—	24	1.28
AirTrunk TAIPEI24 Park	AirTrunk	Taipei, Taiwan	—	9	1.49
AirTrunk SHANGHAI19 Park	AirTrunk	Shanghai, China	—	29	1.26
AirTrunk SHANGHAI28	AirTrunk	Shanghai, China	—	17	1.53
AirTrunk SHANGHAI35 East	AirTrunk	Shanghai, China	—	5	1.41
AirTrunk GUANGZHOU37 Park	AirTrunk	Guangzhou, China	—	11	1.46
AirTrunk HANGZHOU21 Park	AirTrunk	Hangzhou, China	—	9	1.47
AirTrunk HANGZHOU23 Central	AirTrunk	Hangzhou, China	—	17	1.26
AirTrunk HANGZHOU29 East	AirTrunk	Hangzhou, China	—	7	1.44
AirTrunk CHENGDU2 East	AirTrunk	Chengdu, China	—	23	1.35
AirTrunk CHENGDU10 East	AirTrunk	Chengdu, China	—	13	1.53
AirTrunk ULANQAB3 East	AirTrunk	Ulanqab, China	—	18	1.31
AirTrunk ULANQAB5 Annex	AirTrunk	Ulanqab, China	H200	12	1.24
AirTrunk GUIYANG10 Annex	AirTrunk	Guiyang, China	—	16	1.26
AirTrunk GUIYANG13	AirTrunk	Guiyang, China	H200	3	1.41
AirTrunk GUIYANG17	AirTrunk	Guiyang, China	—	13	1.23
AirTrunk MUMBAI1 Annex	AirTrunk	Mumbai, India	—	21	1.25
AirTrunk MUMBAI9 South	AirTrunk	Mumbai, India	L40S	3	1.40
AirTrunk MUMBAI13 West	AirTrunk	Mumbai, India	—	4	1.20
AirTrunk MUMBAI17 North	AirTrunk	Mumbai, India	B200	21	1.22
AirTrunk MUMBAI21 Central	AirTrunk	Mumbai, India	MI300X	14	1.50
AirTrunk CHENNAI35 West	AirTrunk	Chennai, India	—	19	1.23
AirTrunk PUNE12 North	AirTrunk	Pune, India	—	12	1.52
AirTrunk SYDNEY52 Campus	AirTrunk	Sydney, Australia	B200	19	1.23
AirTrunk MELBOURNE4 West	AirTrunk	Melbourne, Australia	—	25	1.32
AirTrunk MELBOURNE10 Central	AirTrunk	Melbourne, Australia	—	17	1.48
AirTrunk AUCKLAND4 Park	AirTrunk	Auckland, New Zealand	—	11	1.52
AirTrunk AUCKLAND6 North	AirTrunk	Auckland, New Zealand	—	14	1.27
AirTrunk AUCKLAND8 South	AirTrunk	Auckland, New Zealand	—	5	1.41
AirTrunk JAKARTA1 East	AirTrunk	Jakarta, Indonesia	L40S	19	1.26
AirTrunk JAKARTA5	AirTrunk	Jakarta, Indonesia	—	24	1.33
AirTrunk KUALA18 Campus	AirTrunk	Kuala Lumpur, Malaysia	—	3	1.40
AirTrunk KUALA23 Annex	AirTrunk	Kuala Lumpur, Malaysia	—	2	1.43
AirTrunk KUALA27 Central	AirTrunk	Kuala Lumpur, Malaysia	—	15	1.23
AirTrunk JOHOR5 North	AirTrunk	Johor Bahru, Malaysia	—	9	1.44
AirTrunk JOHOR9 Central	AirTrunk	Johor Bahru, Malaysia	—	4	1.40
AirTrunk MANILA5 Central	AirTrunk	Manila, Philippines	—	7	1.46
AirTrunk HO11 South	AirTrunk	Ho Chi Minh City, Vietnam	—	19	1.29
AirTrunk HO19 Campus	AirTrunk	Ho Chi Minh City, Vietnam	—	15	1.25
AirTrunk HANOI12 East	AirTrunk	Hanoi, Vietnam	—	8	1.20
AirTrunk HANOI14 North	AirTrunk	Hanoi, Vietnam	—	20	1.33
AirTrunk HANOI16 Annex	AirTrunk	Hanoi, Vietnam	B200	5	1.45
AirTrunk DHAKA1 Campus	AirTrunk	Dhaka, Bangladesh	—	10	1.53
AirTrunk ALMATY4 South	AirTrunk	Almaty, Kazakhstan	—	19	1.38
AirTrunk ALMATY8 Park	AirTrunk	Almaty, Kazakhstan	—	12	1.25

[FAQ]

How many datacenters does AirTrunk operate?

57 indexed facilities totaling 1,304 MW across 14 countries.

Which GPUs does AirTrunk deploy?

Mixed, MI300X, A100, H100, H200, L40S, B200

In which countries does AirTrunk operate?

Australia, Bangladesh, China, Hong Kong, India, Indonesia, Japan, Kazakhstan, Malaysia, New Zealand, Philippines, Singapore, Taiwan, Vietnam

[Knowledge Graph]

Company → Datacenter → GPU → Customer → Industry

Every entity on this site is cross-linked. Follow the graph from operators down to specific facilities, GPU clusters, customers, and sectoral context.