Vector search just got a gpu transplant. elasticsearch's new cuvs integration indexes 138 million…
vector search just got a gpu transplant. elasticsearch's new cuvs integration indexes 138 million vectors in under 10 minutes, per elastic, with throughput up 7x and p90 latency down 6x.
the database is now the accelerator's client.
Context
Elastic's engineering blog from September 25 describes GPU-accelerated vector indexing in Elasticsearch built with NVIDIA cuVS. It says a 138 million vector index that took about an hour on CPUs now builds in under 10 minutes.
Elastic reports up to 7x indexing throughput and 6x lower p90 search latency while indexing runs, on 8 NVIDIA RTX PRO 6000 GPUs, with the same recall as CPU-built indexes.
The 138 million vectors, under 10 minutes, 7x throughput and 6x p90 latency all match Elastic's own post. All are vendor figures on one hardware setup, as the note says 'per Elastic'.
Elastic says the latency gain is measured under concurrent indexing load, so it is not a general query speedup.
'The database is now the accelerator's client' is the author's opinion.
Related work
- Elastic: vector indexing on GPU ↗The primary post with the benchmark.
- Elastic on LinkedIn ↗Adds the force merge numbers.
Watch next
- Read the benchmark setup to see how it holds on smaller clusters.
Sources
- Elasticsearch Labs, Sep 25, 2026elastic.co
- Elastic on LinkedIn, Sep 25, 2026linkedin.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 10 October 2026 at 00:16 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →