Vercel

Ask which AI model is best and the answer changes before the ink dries. Vercel's AI Gateway sees tens of trillions of tokens across real production workloads. The picture looks nothing like the benchmarks. • Spend and volume crown different winners. • Agentic requests carry 59% of all token volume, up from 32% six months ago. • Teams at scale average 35 models. Full breakdown in the article. ↓

Vercel Sandbox firewall can now proxy and filter outbound requests. ▪︎ Control all agent traffic from your own proxy ▪︎ More granular credential injection ▪︎ Log and audit outgoing traffic https://lnkd.in/e_A_7sR2

Use tags to easily manage thousands of Vercel Sandboxes. • Tag by environment, team, or customer • Filter to surface exactly what you need • Built for agents and multi-tenant platforms https://lnkd.in/e3uTHuxB

Ship 26 tickets just dropped. London, Berlin, New York, Sydney, San Francisco. Hear from customers shipping AI agents and apps to production, with talks and workshops designed to help your team do the same. Request your ticket: vercel.com/ship

This year's AI Accelerator winners shipped agents for finance teams, application security, and real estate. They built with $8M in credits and hands-on mentorship from Vercel, Amazon Web Services (AWS), OpenAI, Anthropic, and other partners. https://lnkd.in/eaUKRY2t

In collaboration with GitHub, Microsoft, npm, Inc., and Socket, our security team has confirmed that no npm packages published by Vercel have been compromised. There is no evidence of tampering, and we believe the supply chain remains safe. Additional security bulletin updates include: - Clarification that account and project deletion do not eliminate environment variable risk - Guidance on multi-factor authentication - Product updates to help you strengthen your security posture https://lnkd.in/eHqsp8AR

We’ve identified a security incident that involved unauthorized access to certain internal Vercel systems, impacting a limited subset of customers. Please see our security bulletin: https://lnkd.in/eHqsp8AR

Orchestrating 50 concurrent image models exposes the limits of standard compute. Timeouts happen. State gets lost. When FLORA built their creative AI agent FAUNA, they hit this wall. To run thousands of parallel AI jobs reliably, they needed durable execution without managing custom background queues. On April 21, FLORA CTO Charlton Roberts & Vercel, Head of Workflows Pranay Prakash will break down exactly how they achieved this. Register to see how they built it: https://lnkd.in/gWAn8RZc

Vercel Workflows is GA. Your code is the orchestrator. Ship agents, backends, or any long-running process without managing queues, retries, or workers. https://lnkd.in/ep3fWc9A

Companies are buying less software and generating more of it. A few years ago, this sounded "crazy". Now it's a normal day at Vercel and for our customers, as Guillermo Rauch shared on the HumanX stage: → Sales teams building their own lead-qualifying agents → A data analyst agent anyone in the company can query in Slack → 90%+ of support requests automated Generative software needs secure infrastructure built for speed and scale. Read more about Agentic Infrastructure → https://lnkd.in/gxAVXPSK