Superblocks, a vibe-coding startup, has signed a multiyear joint marketing deal with Amazon Web Services that will allow the platform to be embedded directly within AWS customers' private cloud environments, according to a report published by TechCrunch on August 3, 2026. Under the arrangement, enterprises subscribing to Superblocks on AWS can provide vibe coding capabilities to business users while ensuring that applications never transmit data or information externally to model providers or databases. The partnership represents a broader shift in how hyperscaler cloud providers position themselves against frontier AI labs, urging enterprise customers to separate AI models from the surrounding infrastructure needed to operate enterprise AI systems.
The agreement enables Superblocks applications to create Amazon Aurora databases within a company's private cloud infrastructure rather than generating external databases like Supabase, which is commonly used in vibe-coding environments, the report states. These applications will also connect with Amazon Bedrock, the cloud giant's platform for AI app development, AI gateway, and inference services. Superblocks, which employs 50 people and raised a total of $60 million through its Series A round announced in May 2025, counts Spark Capital, Kleiner Perkins, Meritech Capital, and Greenoaks among its backers. Open models represented 29% of all traffic routed through Vercel's AI gateway last month, a tool widely used by enterprises to manage multi-model AI deployments.
"We're going to bring it to your data inside your private cloud," Superblocks co-founder and CEO Brad Menezes told TechCrunch, explaining that data never leaves the customer's AWS account and remains protected by all auditing, encryption, and network controls. Menezes also noted a dramatic reversal in enterprise preferences, observing that "60 days ago they were like, I want a specific model. It's called Anthropic," but now enterprises have shifted to adopting multiple models, particularly frontier Chinese open-weight options. The report highlights that AWS does not yet offer its own vibe-coding agent aimed at business users, though it has Kiro, an AI coding agent for developers, and Quick, an AI assistant for business users that functions more like Claude Cowork or Microsoft Copilot.
The partnership reflects a growing trend among hyperscaler cloud providers pushing enterprise customers to acquire AI harnesses (also known as agentic apps), AI orchestration, security tools, and related infrastructure from cloud platforms rather than from frontier AI providers, according to the report. Microsoft CEO Satya Nadella has been advocating this approach in recent weeks, encouraging enterprise customers to use multiple models to cut costs and avoid lock-in, while warning that AI labs may not be trustworthy enough for agent orchestration or app-level harnesses because they could use that data to study a business and later become competitors. Menezes predicted that "any enterprise that is betting on a single model provider, that executive will be fired," emphasizing that a multi-model strategy across major frontier labs like OpenAI and Anthropic, plus open source options including Chinese and U.S. variants, is now essential for chief information officers. An AWS spokesperson described vibe coding as "an emerging category with real momentum, and exactly the kind of innovation we support," adding that the company supports partners where it sees strong customer demand and alignment with how customers want to build.
The report positions this development as a potential second wave following the introduction of AI coding agents for enterprise developers, now extending to vibe coding for business users within private, secure cloud environments. AWS will also assist in selling Superblocks to enterprises as it does for many of its Marketplace partners. The movement toward multi-model strategies means that enterprises can't tie all their AI scaffolding to a single provider, with companies now seeking model choice for use cases spanning coding, customer service, human resources, and sales automation. If hyperscalers succeed in capturing the infrastructure layer around AI models, they stand to reshape the competitive landscape between cloud platforms and AI labs, potentially limiting the labs' role to model provision while cloud providers control access, security, and orchestration—a shift that could redefine enterprise AI architecture for the next decade.