Vibe coding just got an enterprise security wrapper. Superblocks, a 50-person startup that has raised $60 million as of its Series A announced in May 2025, signed a multi-year joint marketing agreement with AWS. The deal lets Superblocks run inside the private clouds of AWS customers rather than spinning up external services.
What changes for enterprise users
When a business user at an AWS enterprise customer builds an app through Superblocks, that app stays inside the company’s AWS account. Databases spin up as Amazon Aurora instances rather than external Supabase databases. Models run through Amazon Bedrock. Data does not leave the customer’s environment, and the apps automatically fall under the company’s existing IT auditing, encryption, and network controls.
Superblocks co-founder and CEO Brad Menezes framed it directly:
“We’re going to bring it to your data inside your private cloud. The big thing about that is data never leaves. It’s their AWS account and basically secure with all of the auditing, all of the encryption, all of the network controls.”
AWS will also co-sell Superblocks to its enterprise customer base through its Marketplace partner program.

The bigger pattern behind the deal
This is not just a startup landing a distribution partner. It reflects a deliberate strategy from hyperscaler clouds: pull the AI infrastructure layer away from the frontier model providers and anchor it in their own platforms.
Microsoft CEO Satya Nadella has been making the same argument publicly, telling enterprises to use multiple models to cut costs and avoid lock-in, and warning that AI labs should not be trusted with agent orchestration because they might use that business data to study and later compete with their own customers.
Enterprise buyers appear to have already arrived at the same conclusion independently. Menezes noted the shift: enterprises that 60 days ago insisted on a specific model, naming Anthropic by name, have now moved to a multi-model posture. Open-weight models accounted for 29% of all traffic routed through Vercel’s AI gateway last month. Menezes put the stakes plainly: “Any enterprise that is betting on a single model provider, that executive will be fired.”
The operator takeaway
AWS does not yet have its own vibe coding tool aimed at business users. It has Kiro, an AI coding agent for developers, and Amazon Q, an AI assistant for business users, but neither is a direct equivalent of Lovable or Replit for enterprise workflows. That gap is exactly why the Superblocks partnership exists.
For anyone building AI tooling for enterprise buyers, the lesson is clear: data residency, private cloud deployment, and multi-model flexibility are now table stakes. The vendors who win enterprise deals will be the ones who let the customer own the infrastructure, not the ones who require data to leave.
