P007 Ruling out secret chip production

Can secret chip production be ruled out, whether at a secret fab or as extra chips from a known one?

Whether secret chip production can be ruled out, at a covert fab or as extra chips from a known one, including by bounding a known fab's output from lithography-tool and wafer-start counts, floor space, power use, trackable materials and census reconciliation against shipment records.

Status

Analyzed A desk estimate conditional on the AI 2040 Plan A regime, in which chip production is monitored at a small number of controlled fabs and upstream suppliers' purchases, inventory and cumulative production are declared and cross-checked layer by layer; concealed domestic production is given a 25% feasibility that the authors call an intuitive guess. Shavit (2023) proposes production-line monitoring on the nuclear-fuel model and tracking the lithography supply chain to detect new fabs, without a precision. No fab's output has been bounded by measurement and no covert-fab detection method has been tested.

Concretely, we should be able to keep diversion to approximately zero and have very high confidence that total diversion is less than 100k H100e.
Halstead & Larsen (2026), Covert AI Projects (AI 2040 Plan A supplement) (2026-07 (no date printed; accessed 2026-10-15)), L89

Raised in: Harack et al. (2025), Verification for International AI Governance (Oxford AIGI); Scher & Thiergart (2024), Mechanisms to Verify International Agreements About AI Development (MIRI); Aarne & Petrie (2025), International Security Applications of Flexible Hardware-Enabled Guarantees (flexHEG Part 3); Shavit (2023), What Does It Take to Catch a Chinchilla? Verifying Rules on Large-Scale Neural Network Training via Compute Monitoring; Koopmanschap & Barten (2026), How to Catch a GPU: A Taxonomy of Verification and Enforcement Mechanisms for International AI Agreements; Wasil, Reed, Miller & Barnett (2024), Verification Methods for International AI Agreements; Proofworks, Living Verification Documents.

Sub-questions

IDSub-questionWhat the sources report
P007.1Can production-line monitoring or lithography-tool / wafer-start counts bound a known advanced fab's accelerator-die output to a stated precision?partly answered
P007.2What is the minimum detectable covert leading-edge fab output, and the minimum floor area and power of a fab producing accelerators at a stated performance level?implied by a source, not stated as open
P007.3Can lithography-tool shipment tracking account for every advanced-node scanner, ruling out secret fabs (and extra chips at known fabs)?open
P007.4What coverage (% of shipped units, to a stated precision of installed compute) can a chip census reach when reconciled against vendor shipment records?partly answered

Sub-questions not yet stated precisely enough to check

IDSub-question
P007.5Are any raw materials AI-specific enough to track?

Possibly relevant work found by automated search

Found by AI agents searching for this problem's sub-questions, mostly by keyword and citation in the OpenAlex scholarly database and on the web; a second AI pass reviewed each one, and we spot-checked some by hand. They seemed reasonably related, as far as one can tell from outside each paper's field. Some come from neighbouring fields, such as arms control and financial audit, and bear on the problem by analogy. Not reviewed against the status above; a paper listed here does not change that status. 928 of 1052 quotes checked word for word against the paper's text.

For P007.1: Can production-line monitoring or lithography-tool / wafer-start counts bound a known advanced fab's accelerator-die output to a stated precision?

For P007.2: What is the minimum detectable covert leading-edge fab output, and the minimum floor area and power of a fab producing accelerators at a stated performance level?

For P007.3: Can lithography-tool shipment tracking account for every advanced-node scanner, ruling out secret fabs (and extra chips at known fabs)?

For P007.4: What coverage (% of shipped units, to a stated precision of installed compute) can a chip census reach when reconciled against vendor shipment records?

* Full text not reviewed; more likely to contain errors.

Data