AWS re:Invent 2025: The 15 Biggest Products, AI And News Unveiled

From new AI agents and models to Graviton5 and Trainium3 UltraSevers, here are the 15 biggest product launches and news from AWS re:Invent 2025 that you need to know about.

Amazon Web Services launched a slew of new AI models and agents, as well as processors and servers, at AWS re:Invent 2025 that every partner, investor and customer should know about.

From its new Graviton5 processors and Trainium3 UltraSevers to its new wave of frontier agents and Nova 2 models, the $132 billion cloud and AI giant’s innovation engine roared in Las Vegas last week.

AWS also launched new Transform capabilities that can accelerate full-stack Microsoft Windows modernization and eliminate up to 70 percent of Windows maintenance and licensing costs.

[WATCH NOW: The Biggest Products And Innovations For Partners At AWS re:Invent 2025]

Additionally, the Seattle-based world leader in cloud computing launched a new AI Competency for partners, revamped the AWS Marketplace and unveiled massive partner changes slated for 2026.

CRN breaks down the 15 biggest product launches, AI news and partner program updates stemming from AWS re:Invent 2025 last week.

AWS Transform Cuts 70 Percent Of Microsoft Windows License And Maintenance Costs

AWS Transform can now accelerate full-stack Windows modernization by up to 5X across all layers, including .NET apps, SQL Server, UI frameworks, and deployment layers into open source, cloud-native solutions.

Thanks to new capabilities, AWS said its agentic AI service Transform can now eliminate up to 70 percent of Microsoft Windows customers’ maintenance and licensing costs.

The AWS Transform agents start by analyzing a customers’ complete Windows stack and proposing coordinated modernization plans across all layers.

[Related: AWS Trainium3 AI Is ‘The Best Inference Platform In The World,’ CEO Says]

The agent then transforms the application, UI framework, database, and operating system, while providing updates and transformation summaries. The new capabilities enable the rapid modernization of full-stack Windows and SQL Server systems to open-source alternatives.

Click through to read the other 14 biggest AWS product launches and news from re:Invent 2025.

Trainium3 UltraServers

In one of the biggest launches at AWS re:Invent 2025, AWS unveiled its new Trainium3 UltraServers with up to 4.4X more compute performance, 4X greater energy efficiency, and nearly 4X more memory bandwidth compared to Trainium2 UltraServers.

“Trainium3 is actually going to be the best inference platform in the world,” AWS CEO Matt Garman recently told CRN. “It will be the most efficient, most effective, the best cost performance, the lowest latency and the best throughput.”

AWS sees Bedrock with Trainium3 as the world’s leading inference engine and a business that will one day possibly be as large as Amazon EC2.

Powered by AWS’ new Trainium3 chip built on 3nm technology, Tranium3 UltraServers enable organizations of all sizes to train larger AI models faster and serve more users at lower cost.

Tranium3 UltraServers pack up to 144 Trainium3 chips into a single integrated system.

AWS said Trainium3 delivers energy savings of around 40 percent better energy efficiency compared to previous generations.

New Frontier AI Agent: Kiro Autonomous Agent

AWS’ three new Frontier agents represent a new class of AI agents that are autonomous, scalable, and work for hours or days without constant intervention.

Dubbed the frontier agent for software development, the Kiro autonomous agent maintains persistent context across sessions and continuously learns users pull requests and feedback.

It can handle a range of tasks—from triaging bugs to improving code coverage—with a single change spanning multiple repositories. Users can ask it questions, describe a task, and assign tasks in their backlog directly from GitHub.

The agent will then independently figure out how to get the work done, sharing changes as proposed edits and pull requests.

For IT teams, the Kiro autonomous agent is a shared resource that works alongside the team, building a collective understanding of codebases, products, and standards. It connects to a team’s repos, pipelines, and tools—such as Jira, GitHub, and Slack—to maintain context as work progresses and adapt to changes or updates.

Nova Forge

Nova Forge is AWS’ new service for building frontier AI models.

Nova Forge lets customers build their own optimized variants of Nova by blending their proprietary data with Nova’s frontier capabilities.

AWS said the service pioneers “open training” that gives exclusive access to pre-trained, mid-trained, and post-trained Nova model checkpoints so customers can mix their proprietary data with Amazon Nova-curated datasets at every stage of model training.

The result is a customized model that combines Nova's full knowledge and reasoning power with understanding of each organization's specific business.

Nova Forge offers three additional capabilities: the ability to train AI using your own environments; the option to create smaller, faster models that maintain their intelligence at a lower cost; and access to a responsible AI toolkit that allows user to implement safety controls.

Customers can create their own frontier model with Nova Forge, then deploy it on Amazon Bedrock.

Graviton5

AWS introduced its new Graviton5 processors, which is the company’s most advanced custom chip for a broad set of cloud workloads.

New Graviton5-based Amazon EC2 M9g instances deliver up to 25% higher performance than the previous generation, with 192 cores per chip and 5x larger cache.

Each Graviton5 core has access to 2.6x more L3 cache than Graviton4, which translates to fewer delays waiting for data and faster application response times.

Graviton5 instances leverage sixth-generation Nitro Cards to offload virtualization, storage, and networking functions to dedicated hardware.

Graviton5 introduces the Nitro Isolation Engine as an enhancement to the Nitro System, harnessing formal verification to provide mathematical certainty that your workloads are isolated from each other and AWS operators.

New Frontier AI Agent: AWS Security Agent

AWS’ new frontier Security Agent helps users build applications that are secure from the start across AWS, multicloud, and hybrid environments.

Security Agent embeds security expertise throughout the development lifecycle, proactively reviewing design documents and scanning pull requests against organizational security requirements and common vulnerabilities.

Once a customer defines its security standards, the AWS Security Agent automatically validates them across their applications during its review.

The agent also transforms penetration testing from a manual process into an on-demand capability. Users can expand penetration testing across their entire application portfolio. The agent returns validated findings with remediation code to fix the issues it finds.

By continually validating security from design to deployment, the agent helps prevent vulnerabilities early, so you can focus on what matters most.

Nova Act

AWS launched its new service for building and managing AI agents for UI-based workflows.

Powered by a custom Nova 2 Lite model, Nova Act provides the fastest and easiest path to build and manage fleets of agents that automate web browser-based tasks.

Nova Act achieves reliability by training a custom Nova 2 Lite model through reinforcement learning, running thousands of tasks on hundreds of simulated web environments.

AWS said this style of training allows Nova Act to excel on UI-based workflows such as updating data in a customer relationship management (CRM) system, testing website functionality, or submitting health insurance claims.

AWS AI Factories

AWS is providing customers with dedicated AWS AI infrastructure deployed in their own data centers with AI Factories.

AWS AI Factories combines Nvidia GPUs, AWS Trainium chips, AWS networking, and AI services like Amazon Bedrock and SageMaker AI.

This approach lets customers leverage existing data center space and power while meeting data sovereignty and regulatory requirements.

With a focus on enterprises and governments, AWS AI Factories operates like a private AWS Region that gives secure, low-latency access to compute, storage, database, and AI services.

New Frontier AI Agent: AWS DevOps Agent

Dubbed the frontier AI agent for operational excellence, AWS’ new DevOps Agent is on call when incidents happen. It can instantly respond to issues and use its knowledge of applications and the relationship between components to find the root cause of the problem.

Within parent company Amazon, AWS said the DevOps Agent has handled thousands of escalations, with an estimated root cause identification rate of over 86 percent.

AWS DevOps Agent delivers fewer alerts and recommendations for how to continuously improve the reliability and performance of a customers’ applications across AWS, multicloud, and hybrid environments.

It learns about a client’s resources and relationships spanning from observability tools to runbooks, code repositories, and continuous integration and continuous delivery (CI/CD) pipelines.

The DevOps Agent maps a user’s application resources and correlates telemetry, code, and deployment data to precisely pinpoint root causes and reduce mean time to resolution.

Four New Nova 2 Models

AWS launched four new AI models inside its Nova portfolio with the goal of taking AI training, price performance and AI customer opportunity to the next level.

AWS four new Nova 2 AI models are: Lite, Pro, Sonic and Omni.

“Nova 2 Lite compares really favorably in industry benchmarks to models like Claude Haiku 4.5, GPT-5 Mini, and Gemini Flash 2.5,” said CEO Garman at AWS re:Invent. “It’s our new fast and cost-effective reasoning model suitable for broad set of workloads.”

AWS’ new Nova 2 Pro AI model is dubbed Amazon’s most intelligent reasoning model that can process text, images, video, and speech to generate text.

Nova 2 Sonic is Amazon’s new speech-to-speech model that unifies text and speech understanding and generation for real-time, human-like conversational AI.

Nova 2 Omni is a unified multimodal reasoning and generation model that can process text, images, video, and speech inputs while generating both text and images. Omni handles up to 750,000 words, hours of audio, long videos, and hundred-page documents, simultaneously analyzing entire product catalogs, testimonials, brand guidelines, and video libraries at once.

AWS Transform Gets Agentic AI Boost

AWS added new agentic AI capabilities to Transform to accelerate organization-wide code and application modernization across any code, API, framework, runtime, architecture, language, and company-specific programming languages and frameworks.

AWS Transform’s new custom capability makes large-scale modernizations possible for all legacy systems across any software, code, library, and framework.

With pre-built transformations for common patterns and custom transformations for specific tasks, a specialized agent executes consistent, repeatable, and high-quality transformations.

For a typical customer, AWS said Transform custom can scale modernization across hundreds or thousands of applications, achieving transformation up to 5X faster than when done manually. The transformation agent automatically captures feedback and continues to improve over time, so each subsequent transformation becomes more efficient.

AWS Marketplace Innovation

The AWS Marketplace is the company’s online store where billions of dollars are transacted annually, offering thousands of AWS solution offerings from channel partners and ISVs to distributors.

At re:Invent 2025, AWS launched new AI features, pricing models and product packages to accelerate customer deals on Marketplace.

AWS launched its own agentic technology inside Marketplace to drive better search and discovery results for customers. Agent Mode is a new conversational discovery experience that’s purpose-built for software procurement.

Additionally, AWS Marketplace’s new Multi-Product Solutions capability lets AWS partners combine their own software and services with ISV and Amazon products—as well as with other partner offerings—in a single, packaged offer. Partners can sell prepackaged offerings that combine multiple products and services into a solution tailored to specific industries and customer use cases.

Lastly, AWS Marketplace now offers variable payment models for professional services—allowing billing for time and materials, milestones, or outcome-driven pricing that aligns costs with delivered value. For AI agents and tools, partners can now offer contract-based and usage-based pricing models, allowing them to serve longer-term client engagements.

Amazon Bedrock Adds 18 New Open Weight Model Options

AWS added 18 new open weight models to Amazon Bedrock, the company’s flagship AI platform for building generative AI applications and agents.

New open weight models available in Bedrock are from Google, OpenAI, MiniMax AI, Mistral AI, Moonshot AI, Nvidia, and Qwen.

This includes Google Gemma 3, MiniMax M2, Nvidia Nemotron, OpenAI GPT OSS Safeguard, as well as Mistral AI’s two new Mistral Large 3 and Ministral 3 models.

Amazon Bedrock now provides nearly 100 serverless models.

With access to top models and the flexibility to swap them without rewriting code, AWS said Bedrock makes it fast and easy for customers to evaluate, test, and adopt new models, so they can find the best option for their use case.

New AI Competency For Partners

AWS has unveiled a new AI Competency that focuses on agentic AI, offering rich benefits to partners that achieve the new competency, including funding and AWS resources.

The New AI Competency includes three agentic AI categories that a partner can obtain: Agentic AI Applications, Agentic AI Tools and Agentic AI Consulting Services.

These new categories recognize AWS partners that excel in developing and implementing sophisticated AI systems capable of perceiving environments, reasoning over goals, and taking purposeful actions with minimal human intervention.

Partners that achieve an AI Competency not only receive AWS’ stamp of approval that proves to customers they can successfully complete AI projects, but new funding, resources and marketing as well.

MSP Partner Program Boosts

AWS is giving MSPs more margins via new financial incentives to reward partners that drive key services like AI and cybersecurity, as well as government deals and cloud adoption.

AWS has three new incentives inside its MSP partner program, including an Incentive for Strategic Services that rewards partners that help customers innovate and enhance their cloud capabilities through specific AWS services such as generative AI, cybersecurity and modernization.

Additionally, the company’s new MSP Incentive for Customer Management recognizes MSPs that help customers adopt and use cloud solutions.

The third new incentive is AWS’ MSP Government Practice Benefit, which financially supports partners that serve government customers through their digital transformation journey.

The new incentives will be implemented on Jan. 1, 2026.

AWS has huge partner program changes in store for 2026. For more information about all of AWS’ partner changes, read CRN’s December cover story, ‘AWS Partners ‘Are The Lifeblood,’ Says CEO Garman Ahead Of Channel Program Evolution.’