The Board Room
OpenAI is now on AWS Bedrock, the Microsoft exclusivity is dissolved
A reasonable skeptic will call this a procurement footnote. The footnote is that AWS now hosts both frontier providers while three open-weight alternatives land within 5-8 points of GPT-5.5 on the Intelligence Index at 3.2x better cost economics. Buyer leverage is real for roughly ninety days.
Cloud AI Vendor Landscape Restructured Overnight
OpenAI on AWS Bedrock + exclusivity unwind means Azure's AI moat evaporated. Open-weight models (DeepSeek V4 Pro, Kimi K2.6) score 52-54 vs GPT-5.5's 60 with radically better cost structures. Hugging Face forecasts proprietary API share dropping from 99% to 5%. The window to renegotiate cloud AI contracts is open now and closing fast.
Update: Pentagon Formalizes Government AI Oligopoly
Escalation from Thursday: Anthropic was formally labeled a 'supply chain risk' and excluded from classified networks (IL6/IL7). Seven vendors named: OpenAI, Google, Microsoft, Amazon, Nvidia, xAI, Reflection. Safety guardrails are now a disqualifying factor in defense procurement. Any Anthropic-dependent stack touching government work carries explicit counterparty risk as of this week.
$725B Capex Meets a 4-Year Depreciation Clock
Big Four committed $725B in combined AI capex. GPUs depreciate on 4-year cycles, not 20-year fiber schedules — meaning the revenue ramp must be far steeper than any prior infrastructure buildout. OpenAI's revenue miss against this backdrop isn't one bad quarter; it's the first data point on a monetization curve that needs to service an unprecedented capital commitment.
Platform Beats Model: Revenue Proof Arrives
Replit grew from $2.8M to ~$1B ARR in 18 months (350x) with 300% NRR and positive margins. Codex doubled API revenue in 7 days — not on model quality but CI integration, tooling, and developer experience. The moat has migrated from intelligence to platform stickiness, and revenue data now proves it.
Physical Infrastructure Under Kinetic Threat
Amazon data centers suffered drone strikes requiring months of repair. Coatue launched a 'tens of billions' fund to acquire land for AI data centers, reflecting physical scarcity. Combined with Strait of Hormuz closure and Spirit Airlines' liquidation from energy price shock, digital infrastructure's physical dependencies are repricing simultaneously.
The 90-Day Cloud AI Renegotiation Window — and Why It Closes Before You Think
The Exclusivity That Defined the Market Is Over
OpenAI landed on AWS Bedrock this week with GPT-5.5 and a joint agent platform. The AGI clause in the Microsoft partnership has dissolved. In one news cycle, the model that defined Azure's AI competitive advantage now runs on Azure's primary competitor. AWS now hosts both Anthropic and OpenAI, which makes it the default multi-model cloud whether it wanted the title or not.
Any enterprise that chose Azure primarily for OpenAI access just lost its switching cost. The switching cost has not increased elsewhere. It has evaporated at the origin.
A reasonable skeptic would argue that nothing material has changed for a customer already committed to Azure. The reasonable skeptic is half right. Nothing forces a migration. What changed is that the negotiating leverage shifted toward the buyer for the first time in three years, and the window will not stay open long. Any procurement team that built a cloud AI thesis around Azure-OpenAI exclusivity should treat the thesis as expired.
The Open-Weight Gap Collapsed to a Rounding Error
Three open-weight MoE systems — DeepSeek V4 Pro, Kimi K2.6, and MiMo V2.5 Pro — scored 52-54 on the Intelligence Index against GPT-5.5's 60. A year ago, that gap was the investment thesis for closed-model APIs. Today it is 5-8 points, and it is closing faster than most procurement cycles can absorb.
The economic gap is larger than the quality gap. DeepSeek's disk-based KV cache, which persists for hours against the industry-standard five minutes, delivered a published 3.2x effective cost reduction on $1,050 in spend against $3,351 in cache savings. For agent workloads with repeated long-context calls, that is a structural cost advantage per-token pricing comparisons do not capture. Hugging Face's CEO predicts proprietary API share drops from 99% to 5%, and Google's TPU 8 announcement (170-180% training cost-performance gain) shortens the timeline further by letting non-hyperscaler labs train competitive models.
The Moat Migrated, and Revenue Data Proves It
If the model is commoditizing, the value has to go somewhere. Revenue data this week points to where. OpenAI's Codex doubled API revenue in seven days, not on model quality, where Opus 4.7 scores within 3 points, but on CI integration, device toolbar, migration tooling, and developer experience. Nadella, on the same beat, reframed AI as a "usage business" targeting "intense users", which is the closest admission yet that mass-market adoption is running behind what the capex assumed.
The moat has migrated from model intelligence, which three open-weight competitors now approximate for free, to the things the weights do not give you: agent infrastructure, cache economics, and developer experience. Any thesis built on betting on whoever has the smartest model had a shelf life, and the shelf life expired this quarter.
What This Means for Your Stack
The decision this quarter is not which cloud to choose. It is whether the surrounding architecture is abstracted enough to swap providers in 18 months without a replatform. Teams that spend the next two quarters building a thin abstraction over both a frontier vendor and an open-weight deployment will pay a modest engineering tax and keep their pricing leverage. Teams that sign a three-year commitment with the incumbent will discover in month nine that they locked in the wrong floor.
- Update: Musk v. OpenAI seeks structural remedies — asset transfers and leadership removal, not damages — making this a governance decapitation attempt that puts OpenAI's corporate form at risk, not just its wallet
- CopyFail Linux vulnerability enables root access across most distributions via a single exploit script — escalate to CISO with a 48-hour patching mandate for all production systems
- Meta acquired Assured Robot Intelligence and is running Meta Robotics Studio — an 'Android of humanoids' platform bet with whole-body control models, e-Flesh tactile sensing, and on-device model compression
- Nearly 40% of new podcast feeds in a nine-day window were AI-generated — a structural threat to any business model dependent on distinguishing human from machine content
- Hugging Face expects agent users to surpass human users by late 2026 — already rebuilding platform with CLIs, agents.md files, and token-efficient APIs for machine consumers
- Google TPU 8 delivers 170-180% training cost-performance gain and 300% network bandwidth increase — accelerating open-weight model development by non-hyperscaler labs
- MCP vs Skills is now a production architectural fork for agent extensions — MCP runs containerized JSON-RPC with overhead, Skills runs in the agent's environment with no isolation; most orgs are choosing inconsistently across teams
- Prompt injection is confirmed #1 OWASP LLM vulnerability with no single fix — defense requires dual-LLM architecture (Planner/Executor Split) that doubles inference cost; Google already absorbing this in Gmail at scale
- S&P 500 hit 7,230 and Nasdaq 25,114 record highs during active US-Iran naval war with Strait of Hormuz closed — Spirit Airlines liquidated as first corporate energy casualty, EU 25% auto tariffs threatened
- Berkshire Hathaway sits on $373B cash under new CEO Greg Abel, who has underperformed S&P by 30+ points — largest uncommitted capital pool in corporate history facing pressure to deploy, first shareholder meeting imminent
The AI vendor landscape restructured in a single week — OpenAI left Microsoft's exclusive orbit for AWS, open-weight models closed to within 5-8 points of frontier at a fraction of the cost, and the Pentagon drew an explicit line excluding Anthropic from classified work — all against $725B in committed capex that depreciates on 4-year GPU cycles, not 20-year infrastructure schedules. The organizations that renegotiate cloud AI contracts and build model abstraction layers this quarter will capture leverage that disappears by fall; the organizations that wait will discover their incumbents already restructured pricing without them.