Amazon Web Services

Confidence 0.85 · 6 sources · last confirmed 2026-06-15

Amazon Web Services (AWS) is the cloud-computing subsidiary of Amazon. Launched 2006. The dominant public-cloud provider by revenue alongside Microsoft Azure and Google Cloud — a major hyperscaler hosting much of the world’s enterprise AI workload.

Role in the wiki

Two distinct angles on AWS surface across the ingested sources:

1. As an AI infrastructure provider

AI Index 2025 §4.1 timeline: Anthropic’s partnership with AWS expanded by an additional $4 billion investment from Amazon in November 2024, bringing the total to $8 billion. Significant signal of AWS’s commitment to frontier model partnerships.

2. As a thought-leadership platform

AWS runs an “Executives in Residence” program advising Fortune 500 leadership teams. Jana Werner and Phil Le-Brun (authors of “Become an Octopus Organization” in HBR Nov–Dec 2025) hold these positions. The Octopus framework is implicitly AWS-aligned: customer-centricity, distributed systems, microservices, decoupling — all align with AWS’s core selling points (cloud microservices, serverless, decoupled architectures).

The AWS Events keynote series (“AWS Events” YouTube channel)

The AWS Enterprise Strategy team (former C-level executives — Jonathan Allen, Steven Brovich, Tom Godden, et al.) delivers a recurring executive-forum keynote on agentic-AI leadership, published on the AWS Events channel (an alias of this entity). Three are in the wiki, all sharing the same four-question framework (economics / talent / structure / governance) and USE/COMPOSE/BUILD doctrine:

Discount the Octopus framing slightly for AWS’s commercial interest, but the underlying argument is well-supported by examples external to AWS (Netflix, Google, Coca-Cola, U.S. Army).

3. As an AI-stack vendor — Jassy’s three-layer framing (Jassy 2025)

Andy Jassy (Amazon CEO, AWS founder) named the wiki’s earliest first-party three-layer AI-stack framing in May 2025:

LayerAWS productPurpose
Bottom (infrastructure)TrainiumCustom AI training chip — “going to help people save a lot of money relative to what the cost has been to date”
Bottom (infrastructure)SageMaker”The standard way for people building their own models — to get the data in, build a model, experiment, deploy into production”
Middle (orchestration)BedrockFrontier-model marketplace + guardrails / RAG / agentic capabilities features — “take a bunch of automated actions in succession”
Top (applications)QAmazon’s AI coding assistant — “the best AI-powered coding assistant” (Jassy’s framing)

Jassy’s complaint in this section: “there was a lot of attention early on on ChatGPT … I think people have slept a little bit on the other layers of the stack” — direct marketing message positioning AWS as the stack-vendor versus OpenAI / Anthropic as application-vendors. The structural decomposition is compatible with Chase’s later (April 2026) four-sub-layer Build split (Frameworks / Runtimes / Harnesses / No-code) at a coarser granularity. Bedrock’s agentic-capabilities feature is the wiki’s earliest first-party-CEO mention of a productised harness primitive — eight months before Anthropic’s Managed Agents announcement.

4. As an AI-stack vendor 12 months on — the USE / COMPOSE / BUILD frame ( AWS London Exec Forum 2026)

A year after Jassy’s three-layer framing, Jonathan Allen (AWS Executive in Residence, the third holder of that title to appear in the wiki after Werner and Le-Brun) gives the agentic-AI-era restatement. The USE / COMPOSE / BUILD economic-decision framework is the agentic-system-economics overlay on Jassy’s stack:

  • USE = third-party agentic vendor solutions (the SaaS layer Jassy’s top-of-stack applications).
  • COMPOSE = frontier-model APIs (Haiku / Sonnet / Opus 4.7 from Anthropic on Bedrock + cross-vendor models). Allen reports ~80% of his AWS customers currently land here.
  • BUILD = train your own model. SageMaker / Amazon Nova / the Forge are the tooling; justified only when fine-tuning improves inference economics over the frontier model.

Allen’s keynote also names the Amazon Bedrock AgentCore productised harness/runtime layer (announced at AWS Summit by Francesca Vasquez earlier the same day) — the AWS agentic-runtime equivalent of Anthropic Managed Agents, for containment, simplicity, security, observability. Brooklyn Solutions’ customer testimony ratifies AgentCore + Bedrock Guardrails as the AWS-vendor harness composition in regulated-customer contexts (MoD, Danske Bank).

The wiki’s longitudinal-AWS-doctrine reading: the three-layer stack has stabilised at 12-month delta; what’s new in May 2026 is the explicit USE/COMPOSE/BUILD economic-decision overlay on the middle (orchestration) layer. The middle layer is where 80% of customer activity lives.

Notable people (mentioned in this wiki)

  • Jana Werner — Executive in Residence, AWS; co-author of the Octopus Org book and HBR article.
  • Phil Le-Brun — Executive in Residence, AWS; co-author of the Octopus Org book and HBR article.
  • Jonathan Allen — Executive in Residence, AWS; headline speaker at the 2026 London Executive Forum (Dangling — single-source so far, deferred entity-page promotion per second-source rule).
  • Anthony Liguori — VP Distinguished Engineer, AWS; Project Mantle lead in Allen’s keynote (76-day Bedrock-substrate-rebuild case) (Dangling).
  • Matt Garman — AWS CEO (since June 2024); quoted by Allen on the junior-hiring-crisis moral framing (Dangling).
  • Francesca Vasquez — AWS executive who introduced AgentCore / security-agent / DevOps-agent at the London Forum earlier the same day (Dangling; first-name-only in the transcript, presumed Francesca Vasquez).

AWS’s positioning vs. competitors (note for future ingests)

In the AI infrastructure layer:

  • AWS — cloud + Anthropic partnership ($8B); SageMaker; Bedrock for foundation-model access
  • Microsoft Azure — OpenAI partnership; Copilot integration
  • Google Cloud — Gemini family; deep ML expertise via Google DeepMind

The wiki’s foundation-models page lists notable models from each parent company. The AWS-Anthropic partnership likely deepens further; future ingests should track.

Open questions

  • AWS’s strategic AI roadmap beyond the Anthropic partnership — Bedrock model selection, Q business agent, Trainium/Inferentia silicon.
  • The AWS Executives in Residence program — how many residents, how selected, what outputs.
  • Conflict-of-interest framing: how to evaluate AWS-affiliated authors writing about org design / customer-centricity / distributed systems.