Pulumi

Every platform team fields the same questions. What is running? What breaks if we change this? What can we safely delete? The answers exist. They're just scattered across state files, cloud consoles, and the memories of whoever set things up. Pulumi already holds most of the picture: the resources your programs manage, the dependencies between them, how stacks consume each other's outputs, and the resources discovered in your cloud accounts outside IaC entirely. Today we launched the Context API, which connects those views into a single graph you can query. We designed it agent-first. The API serves its own primer, so any agent, from Neo to Claude Code to your own automation, can learn the whole query language in one request. Validation errors are specific enough that agents correct their own queries without help. Which stacks break if we upgrade this provider? How much of our infrastructure lives outside IaC? What has no dependents and can be retired? Each question can be answered with a single query. https://lnkd.in/gtcybY_3

Your platform team spent a decade building production Terraform. Embracing the agentic future of infrastructure shouldn't mean throwing that away — and it doesn't have to. Tomorrow — Pulumi for All Your IaC: Including Terraform and HCL, a free hour where Daniel Perlovsky (Principal Product Manager at Pulumi) and Engin Diri demo three ways your existing estate comes along, with no migration project and no rewrite: • Point an existing Terraform Enterprise workload at Pulumi Cloud — remote plans, approvals, and deployments keep working exactly as they did before. • Reuse your existing Terraform modules inside a brand-new Pulumi program. • Write infrastructure in HCL as a first-class, 100% OpenTofu-compatible Pulumi language, with access to the full Pulumi provider ecosystem. Plus the guardrails Terraform alone doesn't give you, via Neo AI code review. 🗓️ Wednesday, August 26 · 10 AM ET / 4 PM CEST · 60 minutes · virtual & free (EMEA-friendly time) Register at the link below. https://lnkd.in/gZEZ_GvQ

Waldemar Kindler plans these migrations for a living, and he says the question has changed. It used to be about cost. Now the companies coming to him, German manufacturers and large enterprises, are asking whether they could run on European infrastructure in parallel and switch over if they ever needed to. That's a different engineering problem than making the bill smaller. It's portability as a requirement, with NIS2 widening what counts as critical infrastructure underneath it. On September 3 he joins the panel to say what that portability actually costs, since he's the one doing the work. Sarbjeet Johal will push back on how many companies genuinely need it. Adam Gordon Bell moderates, with live audience Q&A. 🗓️ Thursday, September 3 · 12 PM ET / 9 AM PT / 6 PM CET · 60 minutes https://lnkd.in/g5TznvH8

Jim Dowling moved Hopsworks to OVHcloud and cut his infrastructure bill 62%. The number gets the attention, but the more useful detail is what he didn't do. He didn't leave the cloud. He changed providers, and he deliberately kept the serverless piece in North America because latency mattered more there than consistency did. That's an architecture decision with tradeoffs, not a statement. On September 3 he walks through what moved, what stayed, and what he'd do differently. He's on the panel alongside Sarbjeet Johal, who argues that for most companies at scale the economics still favour staying exactly where they are. Adam Gordon Bell moderates, with live audience Q&A. 🗓️ Thursday, September 3 · 12 PM ET / 9 AM PT / 6 PM CET · 60 minutes https://lnkd.in/g5TznvH8

If you squint hard enough, Terraform and Pulumi do the same thing: turn desired state into provider steps. So Pulumi HCL makes a simple promise: a program that works for tofu apply also works for pulumi up. Our tfcompat tests assert exactly that. A test never says what Pulumi HCL should do — it's just an HCL program and its providers. If OpenTofu and Pulumi HCL don't make the same provider calls, it fails. Tests like that are very hard to make fail for a bad reason. So we let LLMs write them, and every failure they find is a real divergence. We wrote up how it works, prompt included. https://lnkd.in/gW39DdKZ

European regulations and rising cloud bills have some CTOs asking a question that was unthinkable five years ago: do we need an exit plan from our cloud provider? Cloud sovereignty tends to get discussed as either a slogan or a compliance headache. For a growing number of European teams it's neither. It's an architecture decision, with real numbers on both sides of it. On September 3 we're hosting a live panel with three people who don't agree about the answer: • Waldemar Kindler (CEO, Think Ahead Technologies) plans these migrations for German manufacturers and enterprises • Jim Dowling (CEO, Hopsworks) has actually done one, and will walk through what moved and what stayed put • Sarbjeet Johal (cloud economist, theCUBE) makes the case that at scale the public cloud math usually still wins The discussion covers what the EU regulatory push actually requires, which is narrower and stranger than the headlines suggest, and the Data Act's January 2027 rules on switching providers. Adam Gordon Bell moderates, with live audience Q&A. 🗓️ Thursday, September 3 · 12 PM ET / 9 AM PT / 6 PM CET · 60 minutes https://lnkd.in/g5TznvH8

You may have noticed it too, I certainly did: geopolitics echoes into our work here in Europe much more than it used to. For a long time the choice was simple. You picked one of the big cloud providers because they were convenient, offered every service you could think of, and the price was right. That default got challenged over the last couple of years. June's Claude export-control episode made a lot of people here realize that access to services can become a political question. Add the CLOUD Act, rising cloud bills, and regulations like NIS2 and DORA, and "just pick a hyperscaler" stops being the automatic answer. Sovereign cloud is not a new topic, but these events gave it real momentum. Providers like STACKIT, Exoscale, OVHcloud, or Hetzner are winning workloads now that would have gone to a hyperscaler without a second thought a few years ago. STACKIT is a good example: it belongs to the Schwarz Gruppe, the company behind Lidl and Kaufland, and a retailer that competes with Amazon was never going to run on AWS. So they built their own cloud and put more than 13,000 stores and warehouses on it before selling a single VM to outside customers. Their data centers are all in Germany and Austria, and there is no US parent company for a CLOUD Act request to land on. In April the European Commission picked them as one of four providers in a €180M cloud tender. If you want to go deeper, Robin Hermann at STACKIT and Bert Hubert are the two voices I'd follow on this. So I'm pretty excited about the panel we're hosting at Pulumi on September 3rd, where Adam Gordon Bell will dig into exactly this with: - Waldemar Kindler, CEO and Co-Founder of Think Ahead Technologies GmbH - Jim Dowling, Co-Founder and CEO of Hopsworks - Sarbjeet Johal, Cloud Economist and Founder of Stackpane Waldemar I know personally from my time at the Schwarz Gruppe. He lives and breathes cloud and European sovereignty like few people I know. September 3rd, one hour, online. I want to hear them argue about whether European providers can actually close the service gap, and what it costs to try. Link in the first comment as usual!

You may have noticed it too, I certainly did: geopolitics echoes into our work here in Europe much more than it used to. For a long time the choice was simple. You picked one of the big cloud providers because they were convenient, offered every service you could think of, and the price was right. That default got challenged over the last couple of years. June's Claude export-control episode made a lot of people here realize that access to services can become a political question. Add the CLOUD Act, rising cloud bills, and regulations like NIS2 and DORA, and "just pick a hyperscaler" stops being the automatic answer. Sovereign cloud is not a new topic, but these events gave it real momentum. Providers like STACKIT, Exoscale, OVHcloud, or Hetzner are winning workloads now that would have gone to a hyperscaler without a second thought a few years ago. STACKIT is a good example: it belongs to the Schwarz Gruppe, the company behind Lidl and Kaufland, and a retailer that competes with Amazon was never going to run on AWS. So they built their own cloud and put more than 13,000 stores and warehouses on it before selling a single VM to outside customers. Their data centers are all in Germany and Austria, and there is no US parent company for a CLOUD Act request to land on. In April the European Commission picked them as one of four providers in a €180M cloud tender. If you want to go deeper, Robin Hermann at STACKIT and Bert Hubert are the two voices I'd follow on this. So I'm pretty excited about the panel we're hosting at Pulumi on September 3rd, where Adam Gordon Bell will dig into exactly this with: - Waldemar Kindler, CEO and Co-Founder of Think Ahead Technologies GmbH - Jim Dowling, Co-Founder and CEO of Hopsworks - Sarbjeet Johal, Cloud Economist and Founder of Stackpane Waldemar I know personally from my time at the Schwarz Gruppe. He lives and breathes cloud and European sovereignty like few people I know. September 3rd, one hour, online. I want to hear them argue about whether European providers can actually close the service gap, and what it costs to try. Link in the first comment as usual!

If your coding agent writes infrastructure code that looks right but takes a few iterations before it actually deploys cleanly, the fix is in the setup, not better prompts or a bigger model. Tomorrow — Getting Started with DevOps AI Skills, a free hour where we configure Claude Code and opencode from scratch (settings, LSP, hooks, the official Pulumi skills), then put them to work on a real DevOps task, live, so you can see what a properly configured agent does differently. You walk out with an agent that produces Pulumi that's correct and consistent. 🗓️ Wednesday, August 12 · 12 PM ET / 9 AM PT · 60 minutes · virtual & free There's still time to register 👇 https://lnkd.in/gmZJTB3K

You hit an unexpected Pulumi error. To debug it, someone needs the logs, but logging wasn't on, so now you're re-running the command trying to reproduce it. Sometimes it won't reproduce. Sometimes the state Pulumi was in is already gone. And even when you do capture a log, there's no safe way to hand it over. These files still contain secrets. Pulumi v3.254.0 fixes both. Every operation is now logged to an encrypted file automatically, and a new command re-encrypts and redacts it so you can share it over any channel, even a public GitHub issue. We wrote up how it works. https://lnkd.in/g9CKGXMw