The archive · Developer & Business Tools · Product decision · 2024–2026
Spice AI bets open-source data acceleration, not ETL, will feed AI apps
Spice AI's Apache-2.0 Rust engine materializes and accelerates SQL from any source; the 2024 Show HN drew 177 points and the repo now shows 3.1k stars.
Spice AI
What the business is
Spice AI makes Spice, a portable Apache-2.0 Rust engine that materializes, accelerates and serves SQL query, search and LLM inference from 30+ existing data sources at localhost latency, with managed Spice Cloud and on-prem Spice.ai Enterprise sold on top.
How it started
On 2024-03-28 Spice AI's open-source runtime reached Hacker News as 'Show HN: Spice.ai – materialize, accelerate, and query SQL data from any source'. The project's stated goal — build data-grounded apps in minutes, no pipelines, no glue — framed the problem as one of location and latency: data sits in sources and warehouses far from the application, so teams stitch ETL and query layers together while the app waits. The answer was an engine that materializes a working set of the source's data and serves it locally at millisecond speed.
What happened
Over the following two years the runtime expanded from SQL materialization into search and inference: 30+ connectors with query pushdown, hybrid vector/full-text search, text-to-SQL, hosted and local model serving, Iceberg read/write and an MCP server, all under Apache-2.0. Spice 2.0 added real-time analytics replicas of operational Postgres, MySQL and MongoDB databases via native CDC — the README claims ~170x faster ingest, ~2-second freshness and 1,046 analytical QPH at SF1000 under a 266,000+ tpmC load — plus a cluster-sidecar architecture for localhost latency at any scale. Luke Kim spoke at Carnegie Mellon's database group, Spice AI presented an AWS re:Invent 2025 session on S3 data lakes for AI, and the README names Twilio and Barracuda as production users.
How it ended up
Still running and expanding as of the 2026-09-05 snapshot: no shutdown, acquisition or pivot is disclosed; the repository kept growing (7,448 commits), Spice 2.0 shipped, and the README advertises managed Spice Cloud and Spice.ai Enterprise for production workloads.
Background
Spice AI sells Spice, a portable Rust engine that materializes, accelerates and serves SQL query, search and LLM inference from existing sources (Postgres, S3, Snowflake, Iceberg and 30+ others) at localhost latency. Its bet, presented in a 2024-03-28 Show HN, was that builders of data-grounded apps would run an open-source engine giving fast local access to their own data instead of assembling ETL pipelines and glue.
Spice works as an active cache: the runtime connects to any declared dataset, keeps a sandboxed, accelerated working set locally, and answers queries there instead of hitting the source — the README's FAQ contrasts this with a cache that only fetches on miss. Teams start with one dataset and no pipeline rewrite; acceleration later widened into hybrid vector and full-text search, text-to-SQL, OpenAI-compatible model serving and an MCP server. The Apache-2.0 core sits under paid managed Spice Cloud and on-prem Spice.ai Enterprise.
Between the launch and the 2026-09-05 snapshot the repository grew to 7,448 commits and 3.1k stars. Spice 2.0 added real-time analytics replicas of operational Postgres, MySQL and MongoDB databases over native CDC — no ETL, Debezium or Kafka — with the README claiming ~170x faster ingest, ~2-second freshness and 1,046 analytical QPH at SF1000 under a 266,000+ tpmC load. Luke Kim spoke at Carnegie Mellon's database group, Spice AI ran an AWS re:Invent 2025 session on S3 data lakes for AI, and Twilio and Barracuda are named as production users.
What has to be true
- Federated SQL with an accelerated local copy removes the ETL tax every team wiring a warehouse or lake into an app feels; the pitch needs no education.
- Apache-2.0 licensing and acceptance of existing sources, one dataset at a time, cut the cost of trying Spice from a migration to minutes.
- Materializing data locally attacks query latency that remote sources cannot fix, so the wedge is a performance win, not just convenience.
- A portable single-binary runtime leaves an upgrade path: free OSS in production first, then managed Spice Cloud and enterprise for teams that will not run it themselves.
What can be applied
Attacking a hated integration step (ETL glue) gives an open-source project its wedge, but the durable business is the managed layer developers pay so they do not run it themselves.
Aftermath
As of the 2026-09-05 snapshot Spice is an active, expanding project: 3.1k stars, 226 forks and 7,448 commits on GitHub, with Spice 2.0 shipped (native-CDC analytics replicas, cluster-sidecar architecture), enterprise hardening (mTLS, Vault and Azure Key Vault secret stores, read-only API keys) and Twilio and Barracuda named in the README as production users. Spice AI still offers the Apache-2.0 runtime beside managed Spice Cloud and Spice.ai Enterprise. No revenue, funding or customer-count figures appear in the reviewed sources.
Sources
- Show HN: Spice.ai – materialize, accelerate, and query SQL data from any source
- spiceai/spiceai — portable accelerated SQL query, search and LLM-inference engine (2026-09-05 snapshot)
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card