Remote detection of data centres

How it works

Remote detection infers the existence, size and status of data centres without entering them ¹. Electro-optical satellite imagery shows these features ¹:

• large data halls;
• substations and switchyards;
• on-site gas turbines and backup generators;
• cooling towers and chillers;
• construction progress.

Analysts combine imagery with permits, utility filings, company announcements and other open sources ¹. Free imagery comes from archives such as the ESA Copernicus Sentinel and NASA Landsat missions ¹. Paid commercial imagery reaches sub-metre resolution, from providers such as Planet Labs, Airbus and Vantor ¹.

Capacity can be estimated from cooling equipment ^{1 2}. Epoch AI's Frontier Data Centers Hub finds chillers and cooling towers in satellite images and checks them against permits and public disclosures ². From these it infers each facility's power capacity, and then the compute installed, in H100-equivalents ².

Halstead and Larsen treat waste heat as the main obstacle to hiding a facility ⁴. They note that "each megawatt of electricity going into a datacenter must be matched with a megawatt going out in some form (mostly as heat)" ⁴. They list other signals ⁴:

• Infrared satellites can see exhaust air and cooling plumes.
• Radar interferometry can reveal ground movement from excavation.
• Ground moving-target radar can track logistics vehicles.

Ansari also notes that infrared remote sensing can detect undeclared data centres through their heat signatures ⁶.

Verification frameworks usually give these signals a supporting role ⁵. Baker et al. list satellite or aerial images, open-source intelligence, information from data-centre suppliers and financial audits as "less robust mechanisms" that could supplement the main ones ⁵. One of their main layers is national intelligence activities, which can draw on human, cyber and signals intelligence ⁵. Remote detection complements chip accounting (Chip registries and manufacturing records) and workload classification (Workload classification from telemetry and side channels).

What it establishes

From outside, analysts can see construction progress, site layout and power and cooling infrastructure, and can estimate power capacity approximately ^{1 2}. In one of Krawec's case studies, Epoch AI's capacity estimates from cooling equipment ranged from about 200 to 500 MW, against a published expected capacity of 300 MW ¹. Imagery can also reveal gaps between announced and observed construction ¹.

Remote detection has five gaps:

• Electro-optical imagery "can only view the outside of buildings", so it gives no chip counts, chip types or actual power consumption ¹.
• Capacity figures from cooling equipment are approximations, not measurements of energy use ¹.
• Power figures alone cannot distinguish AI training from other high-performance computing ⁶.
• Heim and Pilz wrote in 2024 that AI data centres had no visual features that set them apart from data centres hosting other compute, and that large companies often place AI compute on existing campuses ⁷.
• Observing data centres does not help find chips that have been produced but not yet installed ⁴.

Threat model

Remote detection assumes that large facilities have physical signatures that are hard to remove, such as buildings, substations, cooling equipment and waste heat ^{1 4}. It uses imagery and public records, so it needs no cooperation from the operator ¹.

The adversary may try to hide a facility ⁴. Halstead and Larsen consider disguising sites as industrial facilities, and underground designs whose cooling avoids visible plumes ⁴.

The approach also assumes access to recent imagery and skilled analysts. Krawec lists cloud cover, smoke and weather, the cost of recent high-resolution imagery and the need for trained analysts as constraints ¹.

Limitations

• Facilities can be disguised as other industry or placed underground ⁴.
• A sufficiently small covert project could not be ruled out with confidence ⁴. Small data centres in existing buildings may lack distinctive signatures ¹.
• Without a known location, telling data centres apart from other industrial facilities is hard, and automated detection is mainly conceptual ¹.
• Chip counts, chip types and actual energy use cannot be observed ¹.
• Cooling and power features vary across companies and countries, and not every site generates its own power ¹.