Sovereign AI goes mainstream as OpenAI localizes GPT-6

The gist

Sovereign AI just moved from policy wish-list to product reality as OpenAI localizes GPT-6, letting businesses keep advanced model inference strictly inside their own borders.

What to know

  • In November 2026, OpenAI launched Zero Data Retention-compliant, in-country GPT-6 Luna inference in Japan, then extended local processing to India the very next day.
  • Local AI deployment is now big business: McKinsey projects sovereign AI markets to soar from $48B to $320B by 2030, with applications alone expected to hit $180B.
  • Enterprise spend is already exploding—Menlo Ventures found AI deployment budgets hit $177B in September 2026, up 4.7x from 2025 and far outpacing expectations.

Sovereignty Built Into Code

OpenAI’s GPT-6 Luna now enforces data residency and retention at the infrastructure level, turning compliance from a policy into a technical guarantee for enterprises in Japan and India.

What changed in late 2026 was not just more local hosting, but the way providers turned sovereignty demands into enforceable product controls. OpenAI said on November 11 that “OpenAI's API can now run inference for GPT-6 Luna inside Japan,” while also requiring “Zero Data Retention (ZDR) approved by OpenAI,” a structure that separates storage from processing and ties regional use to retention, approvals, and contractual settings so compliance and auditability become technical defaults rather than policy promises.

India shows the same mechanism at enterprise scale: the headline “OpenAI’s GPT‑6 Luna API Inference Residency Launches in India, Led by Nitin Bawankule” signals that buyer requirements are being answered through localized compute, and DQ reported OpenAI “introduced local processing… in India, giving eligible businesses and developers a way to keep model inference within the country,” with the November 12 rollout extending OpenAI’s existing India data residency to where models actually run. That offer is bounded but usable—priced at a 10% premium, backed by a “massive 100MW infrastructure foundation” with Tata Group and TCS HyperVault, and sold as “100% locally processed, sovereign AI infrastructure” for workloads whose processing and computation stay on Indian soil.

Sources

Enterprise Spend Outpaces Projections

Corporate AI budgets are scaling far faster than anticipated, with real-world deployments and spending already validating sovereign AI as a mainstream, high-growth market.

The commercial case for sovereign AI is no longer speculative; McKinsey calls it “one of the largest opportunities in AI,” projecting the combined market for applications, models and supporting tools to grow from roughly $48 billion today to about $320 billion by 2030. That implies a 45% CAGR, and the scale is not confined to infrastructure alone: the applications layer by itself is estimated to expand more than 35% annually, rising from $40 billion to about $180 billion, which suggests buyers are funding real deployments across the stack rather than treating sovereignty as a niche compliance add-on.

That top-down forecast is reinforced by what enterprises are already spending. Menlo Ventures put enterprise AI deployment spend at about $177 billion by September 2026, based on a survey of nearly 400 U.S. buyers responsible for more than $4 billion in AI spend, and said that total was up from $33 billion a year earlier. Just as important, enterprises spent on average 4.7x their 2025 levels against a planned 1.9x, a gap that shows AI budgets are scaling faster than companies themselves expected and that the market is already commercially material.

Sources
Beachman AI Investing WhispersResilient Cyber

Get the stories behind the trends

Deep-dive reporting and the weekly brief, in your inbox.