A local cloud sandbox changes where cloud testing happens. For years, teams validated code by shipping it to a non-production cloud environment. That approach brings cloud bills, permission headaches and misconfiguration risk. It also makes production conditions hard to recreate for debugging. Now AI agents are writing more code, and the pressure on shared staging keeps growing.
LocalStack Director of Marketing Colin Neagle examines that shift with Alan Shimel at KubeCon 2026 in Salt Lake City. The two go back years, to when Neagle edited Shimel's early blog at Network World.
Emulating the Cloud on a Laptop
Neagle explains that LocalStack began as an open source developer tool. Its founders asked a simple question. What if the AWS cloud could run on a developer's laptop? There would be no cloud account to manage and no bill to pay. Because everything stays isolated, there is far less security risk.
In October, LocalStack released a public preview of LocalStack for Azure. Many of its AWS customers also build on Microsoft Azure. The preview covers services such as AKS, and general availability is planned for early 2027.
A Digital Twin for Every Developer
Shimel asks whether the emulator works like a digital twin. Neagle says that is how many customers describe it. Giving each of 1,000 developers a cloud staging environment would be costly and hard to manage. Locally, every developer, AI agent and CI/CD pipeline can run its own instance on demand. Nobody waits in line for staging or files a DevOps ticket.
LocalStack also supports Amazon Bedrock. That lets teams test AI agents inside existing workflows without burning tokens.
Security Testing Before Production
The conversation turns to security. Neagle points to cloud security posture checks like public storage buckets, failed IAM policies and leaked credentials. Catching those issues in a local cloud sandbox keeps them out of production. Enterprise customers in banking, defense and healthcare can also run LocalStack fully offline in air-gapped deployments.
Developer Experience Still Matters
Neagle closes on what struck him at KubeCon. Many sessions still focus on developer experience, even as agents take on more coding work. Whether a human or an agent pushes the code, the goal is safe experimentation and faster validation.
Up Next in KubeCon 2026 Salt Lake City
-
NATS Graduates as a Messaging Backbon...
NATS has graduated from the CNCF after roughly 15 years of development. The messaging project quietly underpins thousands of distributed systems. Many developers still have never heard of it.
Synadia's Andrew Connolly and Waldemar Quevedo examine that milestone with Alan Shimel at KubeCon + Clou...
-
Nix Brings Reproducible Software to t...
Reproducible software used to be a niche concern for build engineers. AI agents have changed that. Teams now generate more code than ever, and every package must run the same way everywhere.
Flox CEO and co-founder Ron Efroni examines that shift with Alan Shimel at KubeCon + CloudNativeCon in Sa...
-
Open Source Security Starts With the ...
Open source security is getting harder to ignore. AI code generators pull in open source libraries and packages at a pace no team can review by hand. Vulnerabilities keep climbing, and attackers increasingly target package managers. Yet many organizations still treat open source as free software ...