Data product · dated public source claims

IndiaAI compute: dated public snapshots

A source-linked record of IndiaAI compute quantities and dated project allocations. Each total shows its source-stated stage, snapshot date, unit, count basis, and definition; unstated fields remain unknown.

Last updated12 Oct 2026

Observation unitGPU counts and mixed accelerator-category entries

Evidence framingDated public-source claims

Recorded observations · 10

These are dated claims, not a live inventory. Announcement does not establish installation, deployment, access, or scheduling.

Mission target

more than 10,000

more-than

India-wide programme

not specified by source

Source/report date: 7 Mar 2024
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: Not a ratio; overall common-compute target. Equipment inclusion criteria are not stated.
Source definition: A policy target of more than 10,000 GPUs; it is not a delivered-capacity count.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

Approved common-compute target; not evidence of delivered hardware.

primary source

Cabinet approves the IndiaAI Mission · Press Information Bureau, Government of India

Initial facility reported

about 10,000

about

India-wide programme

not specified by source

Source/report date: 30 Jan 2025
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: Not a ratio; an India-wide initial-facility count. The source gives no model or operational breakdown.
Source definition: The release describes about 10,000 GPUs in the initial facility; live availability is not defined.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

The release describes an initial facility but does not establish live service or free capacity.

primary source

With robust and high-end common computing facility in place, India all set to launch its own AI model · Press Information Bureau, Government of India

Additional capacity announced

8,693 additional GPUs

exact

India-wide programme

not specified by source

Source/report date: 30 Jan 2025
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: Not a ratio; planned additional tranche. Its relationship to later totals is not stated.
Source definition: The release says 8,693 GPUs were to be added; completion is not established.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

The release says these GPUs were to be added; it does not verify completion.

primary source

With robust and high-end common computing facility in place, India all set to launch its own AI model · Press Information Bureau, Government of India

Empanelled + additional

34,333 total

exact

India-wide programme

not specified by source

Source/report date: 30 May 2025
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: 18,417 empanelled + 15,916 additional GPUs, the two source-stated components; operational inclusion criteria are unknown.
Source definition: The release defines the arithmetic components of 34,333, but does not define the total as installed, live, or schedulable.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

The release combines 18,417 empanelled and 15,916 additional GPUs without defining live or schedulable status.

primary source

India’s Common Compute Capacity Crosses 34,000 GPUs · Press Information Bureau, Government of India

Installed (reported by ET)

a little over 17,300 installed

more-than

India-wide programme

not specified by source

Source/report date: 24 Jun 2025
Source-stated snapshot date: 16 Jun 2025
Count unit: GPUs
Denominator / count basis: Provider-review count; the cited report does not specify a complete provider denominator.
Source definition: The report uses “installed” for a little over 17,300 GPUs in a 16 June provider review; the source does not publish a status crosswalk.

Installed: reported
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

Secondary reporting attributes this early-rollout figure to a 16 June provider review; it is not a current inventory.

independent source

IndiaAI plan moves forward with over 17,000 GPUs successfully installed · ETtech, The Economic Times

Onboarded

38,231 onboarded

exact

India-wide programme

not specified by source · 14 providers

Source/report date: 25 Mar 2026
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: 38,231 GPUs across 14 empanelled service providers, as stated by the source.
Source definition: The release reports the GPUs as “onboarded” but does not define onboarding as installation, live service, or availability.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

The government reports onboarding from 14 providers; no per-provider live inventory is supplied.

primary source

Government supporting organisations and consortia to develop sovereign foundational models · Press Information Bureau, Government of India

Onboarded (rounded)

nearly 38,000 onboarded

nearly

India-wide programme

not specified by source

Source/report date: 29 Jul 2026
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: Aggregate provider count; the release gives no exact count or provider denominator.
Source definition: The release says “nearly 38,000” were onboarded; its exact count and relationship to other stages are unknown.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

A later PIB release repeats a rounded onboarded figure without a new exact count, provider list, or status crosswalk.

primary source

Innovation, Startups, MSMEs and Quality Manufacturing are Pillars of India’s Future Growth · Press Information Bureau, Ministry of Commerce and Industry

Shared capacity

45,000+ shared

more-than

India-wide programme

not specified by source

Source/report date: 13 Aug 2026
Source-stated snapshot date: June 2026
Count unit: GPUs
Denominator / count basis: IndiaAI shared-capacity total; which hardware states or categories are included is not stated.
Source definition: The factsheet reports more than 45,000 as “shared”; it does not define that stage or reconcile it with onboarding.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

The factsheet reports shared capacity as of June; it does not reconcile the figure with onboarding or define currently free devices.

primary source

Semiconductor and AI Revolution · Press Information Bureau, Government of India

Access reported by ETtech

about 30,000 accessible

about

India-wide programme

not specified by source

Source/report date: 1 Oct 2026
Source-stated snapshot date: Not stated by source
Count unit: GPUs
Denominator / count basis: About 30,000 accessible against a 45,000 commitment; the basis of the accessible count is not defined.
Source definition: The article reports access to about 30,000 GPUs; it is not a provider-level inventory or a schedulable count.

Installed: not established
Deployed: not established
Accessible: reported
Schedulable: not established

Scope limitation

A secondary report compares access with a 45,000 commitment; it is not a provider-level or live availability feed.

independent source

IndiaAI Mission likely to rewire after GPU glitch · ETtech, The Economic Times

Provider / accelerator-category table total

45,081 combined entries

exact

India-wide programme

Mixed GPU and non-GPU accelerator categories; category-level counts are shown in the source.

Source/report date: Not stated by source
Source-stated snapshot date: Not stated by source
Document metadata date (not snapshot date): 2 Sept 2026
Count unit: mixed accelerator-category entries
Denominator / count basis: Sum of 14 bidder or consortium rows across NVIDIA/AMD GPU and Intel, AWS, and Google accelerator categories; no common operational denominator is stated.
Source definition: Unknown — the PDF does not say whether the listed quantities are offered, empanelled, installed, or serviceable. Its 2 September 2026 file-creation date is not a snapshot date.

Installed: not established
Deployed: not established
Accessible: not established
Schedulable: not established

Scope limitation

The provider table combines GPU and non-GPU accelerator categories and gives no publication/as-of date, status definition, or live inventory.

primary source

IndiaAI Compute Portal: provider and accelerator-category table · IndiaAI Compute Portal

Methodology

One observation per public compute-count claim. Preserve source report date, source-stated snapshot date, stage wording, unit, denominator/count basis, and definition separately; unknown source fields are not inferred. Treat the provider/category PDF as a mixed-category total, and cite portal allocation rows as project-level records rather than programme-wide hardware totals. Do not add unlike stages or reinterpret claims as installed, deployed, accessible, or schedulable supply.

Limitations

This is not a provider- or facility-level hardware inventory.
The portal allocation register records project-level assignments but does not establish delivery, installation, active reservation, or current availability.
The reviewed sources do not establish a current count of installed, deployed, or schedulable GPUs.
The reported GPU-hours are not defined by the source as metered consumption.
Accelerator models and geographic distribution are not supplied consistently.

Update history

3 Oct 2026Initial publication from dated public claims; unresolved hardware states remain explicitly unestablished.

12 Oct 2026Added source-by-source snapshot dates, stage labels, count units and bases, definitions, and the mixed-category provider table total. Unstated dates and definitions remain unknown; no live capacity crosswalk was found.