Observed Signal · Jan 22, 2026 · Product & Technology Development · Source: OpenAI Blog · Impact: 3/5 · Sentiment: Positive
Powering 800 Million ChatGPT Users with PostgreSQL
OpenAI published an engineering post describing how it scaled PostgreSQL to support millions of queries per second and serve 800 million ChatGPT users. The team retained a single-primary Azure PostgreSQL flexible server for writes and operated nearly 50 geo-distributed read replicas, while migrating shardable, write-heavy workloads to sharded systems such as Azure Cosmos DB. Key techniques included aggressive query optimization, workload isolation, PgBouncer connection pooling (reducing average connection time from 50ms to 5ms), cache locking to prevent cache-miss storms, rate limiting, strict schema-change controls, and testing cascading replication with Azure to scale replicas. OpenAI reports low p99 read latency, five-nines availability, and only one SEV-0 Postgres incident in the past 12 months.
OpenAI demonstrates that PostgreSQL can be scaled to very large read-heavy workloads using cloud-managed Postgres, read-replica strategies, connection pooling, and migration of writes to sharded systems. The techniques and operational lessons are relevant to large-scale platform and data infrastructure teams.
Track OpenAI Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your Polaris7 workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- OpenAI’s PostgreSQL load grew more than 10x over the past year.
- OpenAI uses a single-primary Azure PostgreSQL flexible server instance with nearly 50 read replicas across regions.
- OpenAI migrated shardable, write-heavy workloads to sharded systems such as Azure Cosmos DB and disallows adding new tables to the PostgreSQL deployment.
- OpenAI deployed PgBouncer for connection pooling, reducing average connection time from ~50 ms to ~5 ms in benchmarks.
- OpenAI is testing cascading replication with Azure PostgreSQL to scale beyond ~50 replicas while managing WAL shipping and failover complexity.
Connected Companies & Entities
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
OpenAI Scales Habitat Storage to 1 Billion Users
OpenAI has shared insights on scaling its internal storage platform, Habitat, to support over 1 billion weekly users. Habitat now handles more than 70 million requests per second and serves over 500 petabytes of data across nearly 40 regions. Initially a Python client library, it evolved into a standalone service, later rewritten in Rust for improved efficiency. Key challenges included asyncio scheduling delays, LIFO connection pooling, and thundering herd problems, addressed with Envoy for connection management. The platform relies on Azure Cosmos DB and a distributed lookaside cache, reducing global read latency by 12 times (p50 from 120 ms to 10 ms). They also implemented hot key throttling, quality-of-service tiering, and a new serverless mode. OpenAI is now building a custom exabyte-scale storage system using RocksDB Cloud and Raft replication for future growth.
OpenAI Launches Data Agent in ChatGPT Work, Halts Pro Subscriptions on Astra Demand
OpenAI has introduced a new Data Agent feature within ChatGPT Work, allowing users to query company data using natural language, perform root-cause analysis, and create interactive dashboards without needing SQL skills. The agent connects to various data sources including Amazon Redshift, Google BigQuery, Databricks, MongoDB, Snowflake, ClickHouse, and Datadog, as well as files from Google Drive and SharePoint. It respects existing access controls and can incorporate internal definitions and relationships from tools like dbt, Snowflake Horizon, and Databricks Genie Ontology. Additionally, OpenAI is halting new Pro subscriptions due to unprecedented demand for its GPT-6 Astra model, which was launched in early September 2026. The company cites the high computational load from Pro users as the reason. Existing Pro users are unaffected, and other tiers remain available.
OpenAI publishes new engineering and security blog posts
New posts: 'Rapidly scaling online storage to serve over 1 billion ChatGPT users (Part II)' and 'Securing our frontier model training environment'.
Receive all future signals for OpenAI & competitors in your briefing
Never miss material M&A deals, leadership changes, account wins, or product rollouts. Polaris7 monitors 20,000+ primary sources 24/7, delivering verified telemetry directly into your executive briefing.
