Skip to main content
South Korea
AIMenta
Regulation EU

EU AI Act GPAI obligations enter the one-year transition period

August 2 marked the start of the one-year transition period for general-purpose AI obligations under the EU AI Act, with full enforcement from August 2026.

AE By AIMenta Editorial Team ·
AIMenta editorial take

The clock is now real. APAC providers serving EU users should treat GPAI compliance as an active 2026 program, not a 2027 problem.

The General-Purpose AI (GPAI) provisions of the EU AI Act entered their one-year transition period, setting August 2026 as the compliance deadline for providers of foundation models and API-accessible AI services used by enterprises in EU member states. For APAC enterprises supplying AI products or services to European customers, or using EU-based AI service providers in their own products, this transition period is a working deadline — not a future planning item.

**What the GPAI provisions require.** Foundation model providers must produce and maintain model cards disclosing training data provenance, benchmark performance, known limitations, and intended use cases. API providers must maintain technical documentation supporting downstream enterprise obligations — specifically, the ability for enterprise customers to assess whether the AI system constitutes a 'high-risk AI system' under EU AI Act Annex III. The documentation requirements are binding on the model provider, but enterprise customers bear responsibility for assessing their own deployment against the high-risk categories.

**Which APAC enterprises are in scope.** The EU AI Act applies extraterritorially: APAC companies are in scope if their AI system's output is used in the EU, if they are established in the EU (through a subsidiary, branch, or legal representative), or if they supply an AI system to a deployer who operates in the EU. For APAC mid-market enterprises, the most common in-scope scenario is supplying B2B software, analytics services, or AI-enabled products to European enterprise customers.

**The high-risk category assessment.** The GPAI provisions are separate from the high-risk AI system requirements in Annex III. However, APAC enterprises deploying AI in employment screening, credit assessment, education, healthcare, or critical infrastructure management in Europe face additional obligations under the high-risk categories that require conformity assessment, CE marking, and registration in the EU database. These obligations begin for existing systems after the full compliance deadline.

**AIMenta's editorial read.** For APAC enterprises with European exposure, the August 2026 deadline is real. The one-year transition period should be used to: (1) assess whether your AI deployments are in scope, (2) obtain GPAI documentation from your foundation model providers, (3) complete a high-risk category screening for your specific use cases. Leaving this to Q3 2026 is the most common and most avoidable mistake.

How AIMenta helps clients act on this

Where this story lands in our practice — explore the relevant service line and market.

Beyond this story

Cross-reference our practice depth.

News pieces sit on top of working capability. Browse the service pillars, industry verticals, and Asian markets where AIMenta turns these stories into engagements.

Tagged
#ai-act #gpai #eu #compliance

Related stories