LinkedIn cuts OpenAI costs by up to 67% with in-house AI models
LinkedIn, a Microsoft subsidiary, reduced AI computing costs by up to 67% for its Hiring Assistant service by replacing OpenAI models with smaller, in-house models. The service, used by over 25,000 companies as of late October 2026—up 25% from September—automates tasks like candidate searches and outreach drafting. LinkedIn’s CTO, Erran Berger, said GPU spending will remain flat through June…
Key points
- LinkedIn replaced OpenAI models with in-house distilled models, slashing costs by up to 67% for tasks like candidate searches
- Hiring Assistant now serves 25,000+ companies (up 25% from September 2026) with flat GPU spending through June 2027
- OpenAI’s models still guide new feature development, but routine tasks use cheaper, specialized models
The approach leverages distillation, where smaller models mimic OpenAI’s performance for routine tasks like candidate ranking and messaging. While LinkedIn’s cost savings are tied to its Microsoft partnership (avoiding OpenAI API fees), the strategy highlights broader implications: companies may prioritize operational efficiency over raw model performance. OpenAI’s own price cuts and new models since 2025 complicate direct comparisons, but LinkedIn’s case suggests a future where AI adoption hinges on balancing cost, quality, and scalability—not just model capabilities.
OpenAI is too expensive—How LinkedIn cut costs by up to 67% by switching to 'homemade' AI
note.com · 4 November 2026
Loading the full article…
This text was published by note.com and written by 構造で読むテック. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Business & Funding
All →- Amazon stock rebounds nearly 3% after earnings as AWS growth offsets AI spending concerns · 11 src
- Anthropic may debut on Nasdaq as soon as Nov. 16, reports say · 4 src
- Trefis analysis: AI market may not have a single winner like search · 2 src
- AMD takes $1.07 billion charge for OpenAI and Meta chip warrants · 1 src
- Jeffrey Gundlach calls Anthropic’s investment-grade bond rating hopes absurd as IPO looms · 2 src
Comments
via GitHub Discussions