The dirty secret of security tooling: the more friction you add, the more your engineers find ways around it. Shadow access, shared credentials, workarounds that make your audit trail meaningless.

Brian Shaw spent years inside large banks where the simplest infrastructure task required navigating layers of process and approval. As SVP of Infrastructure and Core Banking at @UpholdInc, a fintech navigating SOC, PCI, ISO, and multiple regulatory jurisdictions simultaneously, he's built something different — an access model that satisfies the same compliance demands without the friction. Thank you Brian for a session that was as honest as it was practical. #TeleportIdentitySummit

Teleport was named Most Innovative IAM Platform in the @TheHackersNews 2026 Cybersecurity Stars Awards. The reason: a unified identity layer that applies the same cryptographic verification and short-lived access to humans, machines, and AI agents, with no standing privileges and no anonymous actors. → goteleport.com/about/newsroo…

Most teams spend weeks answering this question — pulling logs from per-site bastions, cross-referencing VPN records, stitching together incomplete evidence. Teleport records every session with command-level logging tied to a cryptographic identity, so the answer exists before the question gets asked. Learn how it works: goteleport.com/use-cases/dat…

Audit logs and session recordings were designed for humans. But what happens when AI starts acting as a participant in infrastructure workflows? Abe Ingersoll, CSO at @THGIngenuity, brought that question to Teleport Identity Summit London and worked through it using real signals from THG's own clusters, including early experiments with Claude Code as a terminal interface. Thank you Abe for a session that had the room thinking. #TeleportIdentitySummit

Full EU AI Act enforcement is two months away. The organizations that will struggle aren't ignoring it. They already think they're covered. Kayne McGladrey on where the gaps actually hide. goteleport.com/blog/eu-ai-ac…

SSH keys don't expire. That single operational reality is what drove a decade of innovation in secure remote access. Leon Fong, Senior Security Engineer at Adobe, was there for most of it. goteleport.com/blog/evolutio…

Teleport Identity Summit London brought together security and infrastructure leaders from across Europe for an afternoon of sessions worth the trip. Abe Ingersoll, CSO at @THGIngenuity, explored whether AI can detect AI inside live infrastructure workflows. Anton Tsarev and Angel Ivanov walked through how @DraftKings consolidated database access across a SOX-regulated environment. Brian Shaw, SVP at @UpholdInc, made the case that fintech compliance and engineering speed don't have to be at odds. Mukund Cadambi from @coalfire laid out what an identity layer for AI agents in production actually looks like. And Teresa To from @Wise and Thomas Herbin from @canalplus joined the fireside to talk about scaling secure access across global organizations without slowing delivery. Thank you to every speaker who joined us at #TeleportIdentitySummit.

Auditors: "Show us what that script actually did." Your logs: "Someone ran deploy.sh." That gap is a blind spot— and across multi-site infrastructure, closing it takes weeks. goteleport.com/blog/data-cen…

We've looked through every photo from Teleport Identity Summit London and it's hard to find one where someone isn't smiling. That says everything. London, we'll be back. #TeleportIdentitySummit

By October 2026, every new DoD contract requires CMMC certification. If you have agentic AI touching CUI and can't demonstrate how your controls apply to it, now's the time to figure that out. goteleport.com/blog/nist-800…

Teleport Identity Summit London was everything we hoped it would be and then some. Security and infrastructure leaders from around the world, Shangri-La The Shard, and an afternoon that reminded us why we build these gatherings. Thank you to everyone who made the trip. More moments soon. hashtag#TeleportIdentitySummit

I built a local AI agent with its own identity and thought I was doing security right. Boy, was I wrong. Identity is one layer. It doesn't close the four attack vectors that come with running an agent on your machine. 🧵

Imagine having to secure this: • Multiple nodes • Across different data centers • Letting researchers and engineers access them • While having agents run automated jobs It's impossible to keep all of this secure by sharing static keys. Identity-based access is the solution for this. The way this works is pretty simple: Instead of asking "do you have the key to enter?", identity-based access asks "can you prove who you are?" Basically, you give every human and every service a verified identity that belongs to them, is short-lived, and scoped to the specific job they need. You use your identity as long as it is valid; when it expires, it's done. There are no keys to share or rotate, and you always know who (or what) did what. This is what @goteleport is built for. They are an AI Infrastructure Identity platform designed for exactly this class of modern infrastructure. Here is a full explanation of how this works. It's a pretty interesting reading: goteleport.com/how-it-works/ Thanks to the Teleport team for partnering with me on this post.

An AI agent retrieves CUI from three document repositories. The user's query is logged. The output is stored. The actual data access events? Invisible in most implementations. That's the CMMC audit trail problem — and assessors are going to start asking about it. goteleport.com/blog/cmmc-for…

Thousands of racks. Dozens of engineers. Vendors coming and going. At this scale, every interaction with your infrastructure needs a strong identity behind it — not a shared key that nobody owns. See how Teleport's Infrastructure Identity platform was built for the scale and complexity of modern data centers. goteleport.com/use-cases/dat…

4.5x more incidents when AI agents are over-privileged. The controls exist — come see them at Booth #745. @Identiverse | June 15–18 | Las Vegas | #Identiverse

2AM. Node down. A field tech gets in through a shared VPN, leaving teams with no record of who was there or what they did. Anonymity in your infrastructure is a compliance problem, a customer accountability problem, and an audit problem — all at once. See how Teleport ensures every vendor and engineer session is attributed to a real identity — with short-lived privileges that expire when the work is done. goteleport.com/use-cases/dat…

Agentic AI is already in production on Kubernetes. The standards to secure it are still catching up. The CNCF just published the first real attempt at a baseline — and least privilege is where most teams are falling short. goteleport.com/blog/kubernet…

Everyone is asking how to secure their AI infrastructure. Ev Kontsevoy says that's the wrong first question.

Live from The Shard for Teleport Identity Summit London — incredible energy in the room today 🔐