Vercel
Written by the Bee, read it before you send it
KPIMore first meetings turn into real deals, on less preparation time
selling on-prem storage
Vercel just closed the Stakpak acquisition on Jul 15, 2026, expanding its AI Cloud footprint at exactly the moment its underlying storage layer is under compliance and cost strain.
Private-cloud deployment options with GDPR-compliant, non-training data handling medium
Vercel today
- Industry
- Cloud infrastructure / developer tools (AI Cloud platform)
- Size
- ~900 employees (as of mid-2026)
- HQ
- San Francisco, CA (also reported as Covina, CA in some filings)
- Footprint
- San Francisco HQ · Global customer base across enterprises and startups · 20+ regions for Blob storage infrastructure
Vercel is a cloud infrastructure and developer tools company, maintainer of the open-source Next.js framework, providing serverless hosting, CDN, and its "AI Cloud" platform for building and deploying web and AI-native applications. It serves customers ranging from solo developers to large enterprises including PayPal, Supreme, and Under Armour.
- Merger/acquisition of Stakpak closed Jul 15, 2026
- Private Blobs with scoped access went GA, per Jun 30, 2026 changelog update
- Closed $300M Series F at $9.3B post-money valuation, Sep 30, 2025, with a parallel $300M tender offer
- Appointed HashiCorp co-founder Mitchell Hashimoto to its board of directors around the Series F announcement
- Vercel Blob does not support HIPAA Business Associate Agreements, pushing regulated workloads to AWS S3 or alternatives
- Blob storage pricing is positioned as roughly 10x raw R2 pricing, creating cost pressure at scale for large media/file workloads
- v0 stores millions of generated files, and Blob already serves over 400 million files, implying growing internal storage scale demands
- Committed Series F proceeds to scale the AI Cloud platform and v0, its natural-language full-stack app generation agent
- Publicly framed itself as building 'secure, scalable foundation' for enterprise AI agents per CEO Guillermo Rauch's Sep 2025 statement
- Investing in built-in security for running untrusted AI-generated code as 'vibecoding' security concerns rose in 2025
Positioning
We give enterprises a governed data layer so regulated and sensitive workloads don't have to sit on cost-inflated or compliance-gapped storage just to stay usable — your data stays in your governed environment and is never used to train models.
Conversation arc
“You're scaling AI Cloud fast post-Stakpak, and I want to talk specifically about where your storage layer is creating compliance and cost friction as that scales.”
“Walk me through how you're currently handling workloads that need HIPAA-level compliance since Blob doesn't support a BAA today.”
“Our private-cloud deployment keeps data in a governed, non-training environment, which is the exact gap pushing regulated workloads off Blob and onto S3 or elsewhere.”
“I hear that you're already routing regulated workloads to S3 as a workaround — let's talk about whether that workaround scales as Blob crosses 400 million files.”
“Can we get 30 minutes with whoever owns Blob's compliance roadmap to map this against your regulated-workload volume?”
Feasibility
What they appear to run
- Vercel Blob docs state it "leverages Amazon S3 as its underlying storage infrastructure" for durability
- Vercel Storage Marketplace lists Postgres/KV/NoSQL integrations from Neon, Upstash, and Supabase
- Vercel Blob changelog shows Private Blobs with scoped access shipped GA on Jun 30, 2026