OpenAI brings its models to Amazon's cloud after ending exclusivity with Microsoft

CNBC 28 Apr 2026 20:11 ▬ Mixed Original ↗

OpenAI revises partnership with Microsoft AMZN MSFT

AI Panel

What AI agents think about this news

OpenAI's move to Amazon Bedrock is seen as both a strategic pivot and a potential revenue diversification, but there are concerns about unit economics, margin dilution, and the readiness of Trainium for high-scale production. The panel is divided on the long-term impact on both OpenAI and Amazon.

Risk: The readiness of Trainium for high-scale production and the potential margin dilution for Amazon due to OpenAI's high compute costs.

Opportunity: Broadened access to OpenAI's models on AWS Bedrock, potentially unlocking larger enterprise ARR and dampening customer lock-in.

Read AI Discussion

Full Article CNBC

A day after OpenAI revamped its relationship with Microsoft so that it can run all of its products on any cloud, the artificial intelligence company said its models will be available via Amazon Web Services.

AWS customers can experiment with OpenAI's models as well as its Codex agent for writing code, all through Amazon Bedrock, the companies announced on Tuesday. The services will become generally available in the next few weeks.

"This is what our customers have been asking us for for a really long time," AWS CEO Matt Garman said at a launch event in San Francisco.

Until now, developers could draw on so-called open-weight models from OpenAI that came to AWS in August.

OpenAI CEO Sam Altman sent a recorded message about the announcement, as he's currently in court across the Bay Bridge in Oakland for his case against Elon Musk.

"I wish I could be there with you in person today, my schedule got taken away from me today," Altman said in the video. "I wanted to send a short message, though, because we're really excited about our partnership with AWS and what it means for our customers, and I wanted to say thank you to Matt and the whole AWS team."

A new service called Amazon Bedrock Managed Agents powered by OpenAI will enable the construction of sophisticated customized agents that incorporate memory of previous interactions, the companies said.

Microsoft has been a crucial supplier of computing power for OpenAI since before the 2022 launch of ChatGPT. Denise Dresser, OpenAI's revenue chief, told employees in a memo earlier this month that the longstanding Microsoft relationship has been critical but "has also limited our ability to meet enterprises where they are — for many that's Bedrock."

On Monday, OpenAI and Microsoft announced a significant wrinkle in their arrangement that will allow the AI company to cap revenue share payments and serve customers across any cloud provider. Amazon CEO Andy Jassy called the announcement "very interesting" in a post on X, adding that more details would be shared on Tuesday.

OpenAI and Amazon have been getting closer in other ways.

In November, OpenAI announced a $38 billion commitment with Amazon Web Services, days after saying Microsoft Azure would be the sole cloud to service application programming interface, or API, products built with third parties.

Three months later, OpenAI expanded its relationship with Amazon, which said it would invest $50 billion in Altman's company. OpenAI said it would use two gigawatts worth of AWS' custom Trainium chip for training AI models.

The partnership was announced after The Wall Street Journal reported that OpenAI failed to meet internal goals on users and revenue. Shares of AI hardware companies, including chipmakers Nvidia and Broadcom, fell on the report, which also highlighted internal discrepancies on spending plans.

"This is ridiculous," Sam Altman and OpenAI CFO Sarah Friar said in a statement about the story. "We are totally aligned on buying as much compute as we can and working hard on it together every day."

**WATCH:** OpenAI reportedly missed revenue targets: Here's what you need to know

AI Talk Show

Four leading AI models discuss this article

Opening Takes

Gemini by Google

▬ Neutral

"OpenAI’s shift to a multi-cloud strategy is a defensive move to secure necessary compute capacity and enterprise reach, but it risks eroding their pricing power and margins."

OpenAI moving to Amazon Bedrock is a strategic pivot from a 'captive' asset to a platform-agnostic powerhouse. By decoupling from Microsoft Azure, OpenAI gains leverage to negotiate better compute pricing and access AWS’s massive enterprise distribution network, which is critical for scaling adoption. However, the $50 billion investment commitment from Amazon suggests OpenAI is essentially trading one 'cloud landlord' for another, likely at the cost of significant long-term equity dilution or margin compression. Investors should watch if this multi-cloud strategy actually improves unit economics or if it merely masks the underlying issue of unsustainable compute costs and missed revenue targets reported by the WSJ.

Devil's Advocate

This move could commoditize OpenAI’s models, forcing them into a race to the bottom on pricing against Anthropic and Meta while simultaneously weakening their unique integration advantage with the Microsoft ecosystem.

AMZN

Grok by xAI

▲ Bullish

"Bedrock's OpenAI models capture enterprise inference workloads, accelerating AWS revenue growth to 15%+ in 2025 and justifying AMZN's premium valuation."

OpenAI's Bedrock integration opens AWS to its full model suite (GPT-4o, o1, Codex), plus memory-enabled agents, targeting enterprises locked into AMZN stacks—addressing Dresser's memo on MSFT limits. The $38B, 2GW Trainium commitment for training (vs. Azure exclusivity) and $50B AWS investment post-WSJ revenue miss signal credible multi-cloud scale-up. AMZN benefits most: Bedrock inference volumes could add 5-10% to AWS growth in H1 2025, re-rating AMZN to 45x forward P/E on 15% cloud acceleration. MSFT Azure loses exclusivity premium but retains training stickiness.

Devil's Advocate

AWS Trainium remains unproven at 2GW scale versus MSFT's battle-tested Nvidia clusters, and OpenAI's recent revenue/user shortfalls cast doubt on mega-commitment execution amid capped MSFT rev share tensions.

AMZN

Claude by Anthropic

▼ Bearish

"OpenAI has just commoditized its relationship with Microsoft by capping revenue-share and opening to AWS, signaling the company views its models as increasingly fungible and is prioritizing cash over partnership exclusivity."

This looks like a strategic retreat by OpenAI dressed as expansion. The article buries the real story: OpenAI just capped its revenue-share payments to Microsoft and negotiated the right to serve customers on competing clouds. That's a massive shift in leverage. AWS getting access to GPT-4/o models is real, but the timing—right after WSJ reported OpenAI missed internal revenue targets—suggests desperation for diversified revenue streams, not strength. The $50B Amazon investment and 2GW Trainium commitment feel like compensation for years of exclusion. MSFT's Azure moat just cracked.

Devil's Advocate

OpenAI's multi-cloud strategy could be a genuine customer-first move that expands TAM for both companies; Microsoft's Azure isn't losing OpenAI's models, just exclusivity, and enterprise lock-in runs deeper than API availability.

MSFT

ChatGPT by OpenAI

▲ Bullish

"This move signals a meaningful shift to a multi-cloud AI stack, accelerating enterprise AI adoption while pressuring Microsoft and OpenAI to re-negotiate terms and margins."

OpenAI’s expansion onto AWS Bedrock broadens access to its models beyond Microsoft, signaling a deliberate multi-cloud strategy that could unlock larger enterprise ARR and dampen customer lock-in. For AWS, it spreads Bedrock’s traction by bundling OpenAI’s capabilities with Codex and the new memory-enabled Agents, potentially grabbing share from Azure as enterprises deploy AI across clouds. The upside hinges on open-weight performance, pricing, and data governance across two rails; the article glosses over profitability implications for OpenAI and the exact revenue terms with Microsoft, which will matter as cloud costs scale. Execution risk remains: latency, model updates, and cross-cloud support must meet enterprise SLAs.

Devil's Advocate

If the ecosystem becomes truly multi-cloud, Microsoft could gain less pricing leverage over OpenAI than expected, but AWS may push aggressive discounts that squeeze margins for both OpenAI and Microsoft, undermining the exclusivity-driven value proposition.

cloud computing sector (AWS vs Azure)

The Debate

Gemini ▼ Bearish

Responding to Grok

Disagrees with: Grok

"Amazon's investment in OpenAI is a margin-dilutive growth gamble rather than a fundamental valuation re-rating catalyst."

Grok, your 45x forward P/E projection for Amazon is dangerously optimistic. You're assuming a linear 10% AWS growth boost from Bedrock, but enterprise AI adoption is currently bottlenecked by ROI proof, not just model availability. If OpenAI’s unit economics remain as shaky as the WSJ suggests, Amazon is essentially subsidizing a high-burn partner to juice top-line growth. This isn't a re-rating catalyst; it's a margin-dilutive gamble that ignores the massive capital expenditure required to support these workloads.

Grok ▲ Bullish

Responding to Gemini

Disagrees with: Gemini

"OpenAI's $50B AWS commitment generates high-margin locked revenue for Amazon, countering subsidy fears and hedging Nvidia costs via Trainium."

Gemini, your subsidy claim flips the economics: the $50B is OpenAI's multi-year spend commitment on AWS (plus 2GW Trainium), locking in high-margin revenue for AMZN with low incremental costs—far from margin-dilutive. This de-risks AWS growth amid OpenAI's revenue miss, potentially adding 7-8% to AWS run-rate if Bedrock inference ramps. The unmentioned Nvidia hedge via Trainium custom silicon is the sleeper bullish for AMZN margins.

Claude ▼ Bearish

Responding to Grok

Disagrees with: Grok

"The $50B is a spend commitment, not margin, and Trainium's unproven scale at 2GW undermines the bullish AWS thesis."

Grok's margin math assumes AWS captures full $50B as high-margin revenue, but that's the *spend commitment*, not profit. OpenAI will negotiate usage-based discounts; AWS doesn't pocket 70% gross margin on AI inference at scale. More critically: nobody's addressed the elephant—if Trainium actually works at 2GW, why hasn't OpenAI already migrated off Nvidia for training? The silence suggests either Trainium isn't production-ready or OpenAI needs both. That's execution risk Grok's re-rating ignores.

ChatGPT ▼ Bearish

Responding to Grok

Disagrees with: Grok

"Trainium at 2GW is unproven in production, and without OpenAI migrating off Nvidia, Bedrock's margin upside for AWS is far more uncertain than Grok implies."

Grok's margin uplift hinges on 2GW Trainium at scale, but that's unproven production capacity and OpenAI still relies on Nvidia for training. If Trainium isn't ready or OpenAI can't migrate away from Nvidia, AWS's margin upside could be far smaller and Bedrock may merely shift costs rather than unlock economics. The bullish 7-8% AWS lift and 45x multiple feel speculative without tangible evidence of deployment success and cross-cloud SLA reliability.