Expo In-person

AWS re:Invent 2026

📅 Monday, 30 November 2026 → Friday, 4 December 2026 in 137 days

📍 Las Vegas, United States

AWS re:Invent 2026 (30 Nov-4 Dec, Las Vegas) is the world's largest cloud conference, central to AI infrastructure, Bedrock, agents and enterprise generative AI.

AWS re:Invent 2026 is Amazon Web Services' flagship conference, held 30 November to 4 December 2026 across multiple Las Vegas venues including the Venetian, Caesars Forum, MGM Grand, Mandalay Bay and Wynn, with a pre-day on Sunday 29 November for workshops, certification exams and partner sessions. It is the largest cloud computing event in the world, drawing tens of thousands of builders in person, and as the dominant cloud provider AWS uses re:Invent to set its technical agenda for the year ahead — an agenda that in 2026 is overwhelmingly about artificial intelligence.

What's on the 2026 programme

re:Invent pairs marquee keynotes with an unmatched technical programme — over 2,200 sessions, the majority of them interactive workshops, chalk talks and hands-on labs — across the full AWS stack. The AI and machine-learning content dominates, spanning:

  • Amazon Bedrock — the managed model platform, the Amazon Nova model family, and Bedrock AgentCore for building, securing and governing AI agents in production.
  • Amazon SageMaker — the machine-learning platform for training, fine-tuning and deploying models, including large-scale training infrastructure.
  • AI infrastructure and custom silicon — AWS Trainium and Inferentia chips, networking, and the data-centre capacity that determines the cost and speed of AI workloads.
  • Agents and applied generative AI — the agent frameworks and orchestration now central to AWS's enterprise pitch, plus vector and data services that ground models in business data.
  • Security and core cloud — with the standalone re:Inforce security conference merging into re:Invent for 2026, security content is expanded across the week.

Keynotes and who should attend

The week is anchored by keynotes from AWS leadership, including CEO Matt Garman, where the major launches — new models, silicon, and platform capabilities — are unveiled and customer deployments showcased. re:Invent is built for builders: software engineers, cloud and solutions architects, ML practitioners, DevOps and platform teams, and the enterprise technology leaders deciding how to deploy AI at scale. The sheer breadth of the technical programme means attendees typically come with a specific track in mind — building agentic applications on Bedrock, optimising inference cost on custom silicon, or modernising data platforms to feed AI workloads.

What's new for 2026

The clearest shift for 2026 is the mainstreaming of autonomous AI agents. AWS has positioned Bedrock AgentCore — with runtime, gateway, memory and policy controls for trustworthy agents — as the backbone of enterprise agentic AI, alongside continued investment in the Amazon Nova models and next-generation Trainium silicon aimed at driving down the cost of training and inference. The merger of re:Inforce into re:Invent also signals AWS treating AI security and governance as inseparable from deployment, reflecting enterprise demand for agents that are not just capable but auditable and controllable.

Why re:Invent matters

re:Invent is the definitive view of how AI is being industrialised at cloud scale. AWS supplies the infrastructure a huge share of the AI industry runs on, so its choices — which models it hosts, how it prices inference, what its custom chips can do — shape the economics of building with AI for everyone downstream. The 2026 focus mirrors the wider industry shift toward agents and applied generative AI in the enterprise, combined with the relentless drive to make inference cheaper and faster. For engineers, architects and technology leaders, what AWS announces here sets the practical floor for what is economical to build on the cloud over the following year.

How the week is structured

Beyond the headline keynotes, re:Invent runs on a distinctive mix of session formats that reward planning: deep-dive breakouts, interactive chalk talks, multi-hour hands-on workshops in live AWS accounts, small-group builders’ sessions, and gamified jams. Leadership keynotes are spread across the week — an infrastructure and compute keynote, a dedicated AI and machine-learning keynote, and a developer-focused session — each carrying its own launches. Because popular workshops fill within minutes of the catalogue opening, experienced attendees reserve seats early and build a schedule around the one or two tracks most relevant to their work.

Registering and attending

Registration runs through the official re:Invent site, with a paid conference pass covering the full week across all venues; the major keynotes are typically livestreamed for those who cannot travel. Las Vegas hotel capacity around the event books up early, and the multi-venue layout rewards planning a session schedule in advance. Builders who want the hands-on labs, certification exams and in-person networking should register and reserve sessions as soon as the catalogue opens. For practitioners standardising on AWS, re:Invent 2026 is the single most important week of the year for keeping pace with a cloud AI stack that changes fast.

Related events