OpenEvidence says shared compute cut serverless spending by 90%

OpenEvidence reduced its serverless spending by 90% while maintaining performance after enabling Vercel's Fluid compute, its engineering leadership said in a Vercel case study. The medical search service serves 650,000 doctors. Fluid compute lets multiple requests share a running function instance, allowing other work to proceed while one request waits for a database or AI service. Under Active CPU pricing, that waiting time does not incur CPU charges, although memory remains billed while requests are in progress. OpenEvidence says Vercel accounts for less than 5% of its overall infrastructure spending.