The Pulse
AWS and Cognition Sign Multi-Year Devin Deployment Agreement
Amazon Web Services and Cognition have signed a multi-year agreement to help enterprises deploy Devin autonomous software engineers on AWS. The collaboration targets legacy-system modernization, security backlogs and parallel engineering wo
AI.info Team ·
“With Devin, enterprises can pair their engineers with autonomous agents on infrastructure they already trust, and take on the work they’ve been deferring for years, leaving them time to build.”
Art Levy, VP of Global Partnerships at Cognition
Cognition and Amazon Web Services have signed a multi-year Strategic Collaboration Agreement to put Devin into more enterprise production environments, the companies announced September 15, 2026. The deal focuses on software modernization, security remediation and engineering maintenance performed on AWS infrastructure.
Customers can already buy Devin through AWS Marketplace and use the Agent Toolkit for AWS directly inside Devin. The companies say they are also exploring deeper integrations that would make the coding agent easier to operate within existing AWS environments.
Devin Targets the Work Enterprises Keep Postponing
The agreement is aimed at a familiar problem for large engineering organizations: migration projects, framework upgrades, security fixes and maintenance work compete with new product development. Cognition says Devin can inspect a codebase, plan changes, write and test code, then remediate problems it identifies, while engineers decide what should be built and review the resulting work.
Under the arrangement, Cognition and AWS will work with customers on modernization projects that move older workloads onto AWS. Devin can analyze the behavior of an existing system, write replacement code and test the replacement against known outputs. The companies describe that process as a way to reduce modernization timelines from years to months, although the announcement does not provide a general benchmark for those projects.
Customers can also connect Devin to a dedicated AWS virtual private cloud and run several sessions at once. That setup is designed for parallel migrations, security work and other backlogged tasks within an organization’s existing cloud environment.
Single-Tenant Deployments Put Devin Inside AWS Controls
The deployment model gives each customer a single-tenant Devin environment on AWS, with the customer choosing the region where it runs. Cognition says organizational data and knowledge remain inside a dedicated VPC, supporting internal policies and data-residency requirements.
Each Devin session runs in its own virtual machine. The machine is created for the assigned task and removed when the work ends. Devin’s audit-log APIs record user actions and can feed the monitoring and compliance systems that platform teams already use across their AWS environments.
Anasuya Strasner, AWS director of North America ISVs, said the arrangement is intended to help customers modernize faster and put AI into production on infrastructure they already use.
“This collaboration brings autonomous software engineering to enterprises on the secure, global infrastructure they already rely on, so they can accelerate migrations, clear technical debt, and focus their teams on what moves the business forward.”
Anasuya Strasner, Director of North America ISVs at AWS
Mercedes-Benz Case Study Shows the Intended Use
The announcement points to Mercedes-Benz and Echo Global Logistics as examples of joint customers using Devin. Echo says its teams use Devin for codebase insights through DeepWiki and for closing gaps in legacy systems.
Cognition says Mercedes-Benz used Devin to analyze more than 200,000 lines of COBOL, reducing an estimated eight-month modernization project to eight days. The release also describes another global automaker moving a 25,000-line COBOL workflow from a mainframe to AWS Lambda at an estimated 73% lower cost. The announcement does not identify that automaker.
Those examples place Devin’s AWS deployment strategy squarely in the work that companies often struggle to staff: understanding old systems, translating their behavior into new code and validating that replacements still perform the required tasks.
AWS Marketplace Gives the Deal a Procurement Path
Making Devin available through AWS Marketplace gives existing AWS customers a familiar route for purchasing and deploying the service. The arrangement also ties Cognition’s enterprise sales motion to AWS’s customer relationships and cloud operating controls, while giving AWS another software partner focused on AI-assisted engineering workloads.
The agreement does not announce a new Devin model or a change to the agent’s basic function. Its significance is operational: Cognition is packaging Devin for use inside the security, networking, regional and audit structures that large AWS customers already maintain.
Organizations seeking larger deployments can purchase Devin through AWS Marketplace or contact Cognition’s enterprise team. The companies’ initial agreement centers on putting autonomous software engineering into existing AWS environments, with the first named results concentrated in COBOL modernization and legacy-system work.