Oracle Quietly Halves Free Tier Ampere A1 Compute Limits with No Public Announcement

2026-07-03 · Source: InfoQ · Field: Technology & Digital — Cloud Computing & IT Infrastructure, Software Development & Engineering · Depth: Intermediate, short

Summary

Oracle Cloud Infrastructure (OCI) has quietly reduced its Always Free Ampere A1 Compute allowance, effective June 15, 2026. The limits were halved from 4 OCPUs and 24 GB of RAM to 2 OCPUs and 12 GB of RAM, without any public announcement or customer notification. This change, which reduces the monthly capacity from 3,000 OCPU hours and 18,000 GB hours to 1,500 OCPU hours and 9,000 GB hours, caused confusion among users. Free-only accounts saw instances shut down, while Pay-As-You-Go (PAYG) users worried about unexpected charges. Although Oracle support agents reportedly confirmed PAYG accounts might retain the old limits, official documentation contradicts this, leading to reliance on unofficial communications. Grandfathered instances face a "trap": any termination will lock them to the new, lower limits.

Key takeaway

For DevOps Engineers or self-hosters relying on Oracle Cloud's Always Free tier, immediately audit your Ampere A1 instances. Ensure your total OCPU and RAM usage across all instances does not exceed the new 2 OCPU and 12 GB limits to prevent unexpected shutdowns. If you have a Pay-As-You-Go account, contact Oracle Support to confirm your specific limits and retain their response, as official documentation remains ambiguous. Be aware that even grandfathered instances are vulnerable to future termination and subsequent re-creation at the reduced limits.

Key insights

Oracle silently halved its free cloud compute, eroding user trust through poor communication and ambiguous policy.

Principles

• Cloud providers adjust free tiers.
• Silent policy changes erode user trust.
• Documentation must match support claims.

In practice

• Audit OCI Ampere A1 instances.
• Resize instances to new limits.
• Contact support for PAYG clarity.

Topics

• Oracle Cloud Infrastructure
• Free Tier
• Cloud Compute
• Ampere A1
• Billing Policy
• Cloud Governance
• Developer Trust

Best for: CTO, VP of Engineering/Data, Director of AI/ML, Software Engineer, AI Engineer, DevOps Engineer

Related on AIssential

• When GPU Utilization Lies: The Hidden Systems Problem Slowing Modern AI — GPU utilization alone is insufficient; I/O bottlenecks and resource fragmentation significantly reduce GenAI system productivity and increase costs.
• Redesigning the Artemis account deletion back-end — Asynchronous processing prevents timeouts and improves user experience for data-intensive operations.
• Rate Limiter System Design: Token Bucket, Leaky Bucket, Scaling — Token Bucket algorithm is the industry standard for robust API rate limiting, balancing simplicity and effectiveness.
• Building a high-volume metrics pipeline with OpenTelemetry and vmagent — Migrating high-volume metrics to OpenTelemetry and Prometheus requires careful instrumentation, scalable aggregation, and addressing counter semantics.
• before you switch models, run this 30-minute audit on your openclaw stack — A "token autopsy" method systematically identifies and rectifies AI stack cost inefficiencies.
• Service-Induced Congestion in Memory-Constrained LLM Serving — LLM serving's memory growth causes "service-induced congestion," leading to instability and significant throughput loss, which heterogeneity can mitigate.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by InfoQ.