Amazon's $50 billion bet on OpenAI: A frontal assault on Nvidia and a bitter setback for Microsoft

Forget "Peak Amazon": This historic 50-billion-dollar deal makes the company untouchable

"Peak Amazon" refers to the point at which Amazon's growth reached its peak – and after which it no longer continued to increase at the same pace, but slowed down, stagnated, or underwent structural changes. The term is primarily used in economics and e-commerce.

It's a financial and strategic earthquake that has completely reshaped the certainties of the tech industry overnight: With an unprecedented $50 billion capital injection, Amazon is joining OpenAI as a major partner. This deal, the centerpiece of a historic $110 billion funding round, is driving the valuation of ChatGPT's creators to an incredible $840 billion. But there's far more to this news than just big numbers. It's the ultimate proof that Amazon has long since transcended the boundaries of mere e-commerce. Through exclusive cloud contracts, the massive deployment of its own AI chips as a frontal assault on Nvidia's monopoly, and a parallel multi-billion-dollar bet on competitor Anthropic, Amazon is elevating itself to the largest infrastructure giant in the global AI economy. While exclusive partners like Microsoft are suddenly left behind and forced to rethink their strategies, Amazon's extremely aggressive expansion strategy is silencing all its previous critics. The following behind-the-scenes look reveals why this deal will define the cloud market for the next decade – and what risks such a hot investment cycle entails.

Related to this:

• Circular deals involving cloud services? Is Amazon joining Microsoft and Nvidia in investing $50 billion in OpenAI?

Why the e-commerce giant has long since mutated into the world's most powerful infrastructure company, finally proving the skeptics wrong

The news that shook the tech world on February 27, 2026, wasn't just a cash injection, but a tectonic shift: OpenAI had completed by far the largest private funding round in history, raising a total of $110 billion. The fact that Amazon contributed the lion's share with $50 billion, Nvidia $30 billion, and SoftBank another $30 billion, shifted the center of gravity of artificial intelligence in a way that goes far beyond a mere investment deal. OpenAI's pre-money valuation skyrocketed to $730 billion; including the new capital, its valuation reached approximately $840 billion, making it the second most valuable private company in the world after SpaceX. Anyone still believing in a "Peak Amazon" narrative in this context has fundamentally underestimated the implications of this corporation's strategic repositioning.

The record financing and its architecture

OpenAI's funding round surpasses all previous benchmarks for private equity funding, nearly tripling its own record of $40 billion from the previous year. However, the sheer number obscures the strategic intricacies behind it. Amazon's $50 billion commitment consists of an immediately available tranche of $15 billion, followed by another $35 billion tied to conditions that have not yet been publicly disclosed. According to the industry publication The Information, the second tranche is contingent on OpenAI either going public or achieving the milestone of Artificial General Intelligence, with an IPO in the fourth quarter of 2026 considered the far more likely scenario.

For Amazon, this is by far the largest single investment in an external company in its history. The return is not just an estimated seven percent stake in OpenAI, but a multidimensional strategic alliance that could redefine the cloud landscape of the next decade. Sam Altman himself emphasized the desire to create meaningful value for everyone and described SoftBank, Nvidia, and Amazon as long-standing partners who share its ambitions.

Why OpenAI urgently needs capital

The question of why a company that hasn't yet turned a profit is raising hundreds of billions of dollars can only be answered in the context of the escalating costs of the AI ​​arms race. OpenAI projects total revenue of more than $280 billion by 2030, with earnings expected to be split almost equally between its consumer and enterprise divisions. In fiscal year 2025, the company reported annualized revenue of over $20 billion, a massive jump from the roughly $6 billion of the previous year. ChatGPT itself has now surpassed 900 million weekly users and boasts more than 50 million paying subscribers, giving it one of the fastest adoption rates in the history of consumer technology.

But costs are outpacing revenue growth. Inference costs, the ongoing operating costs of the AI ​​models, quadrupled in 2025, pushing the adjusted gross margin down from 40 percent in 2024 to 33 percent. OpenAI has revised its planned compute spending downward from an initial $1.4 trillion to around $600 billion by 2030, an adjustment signaling that its initial expansion strategy had exceeded realistic revenue expectations. Before the current funding round, OpenAI had only $40 billion in cash reserves, and with a projected total consumption of $40 billion by 2028, the company faced financial exhaustion as early as 2027. The $110 billion in funding solves this critical problem and theoretically provides enough capital to reach the targeted break-even point around 2030.

AWS as the exclusive AI cloud hub

The core of the Amazon-OpenAI partnership is not simply the capital injection, but a profound infrastructural integration. Amazon Web Services will become the exclusive third-party cloud reseller for OpenAI Frontier, the AI ​​provider's most advanced enterprise platform. Frontier enables organizations to build, deploy, and manage teams of AI agents that operate across real-world business systems with shared context, integrated governance, and enterprise-grade security. For enterprise customers moving from experimentation to production AI deployment, this is a crucial differentiator.

Furthermore, OpenAI and Amazon will jointly develop a stateful runtime environment, accessible to AWS customers via Amazon Bedrock. This environment gives AI models access to computing power, storage, identity data, and previous context, enabling them to manage ongoing workflows and projects across various software tools and data sources. OpenAI will also leverage two gigawatts of computing power based on Amazon's own Trainium chips.

The financial scope of this cloud partnership is remarkable: OpenAI has committed to spending a total of $100 billion on AWS services over the next eight years. This significantly expands upon an existing agreement signed in November 2025 for $38 billion over seven years. For AWS, this represents a massive, contractually secured revenue guarantee that extends far beyond the AI ​​hype and manifests itself in concrete, long-term cash flows.

Amazon's chip offensive as a silent disruptor

Anyone who still primarily perceives Amazon as an online retailer is overlooking one of the most dramatic transformations in recent economic history. Amazon's in-house chip division, consisting of the Trainium AI accelerators and the Graviton general-purpose processors, has achieved an annual revenue run rate of over $10 billion, with year-over-year growth exceeding 100 percent. CEO Andy Jassy commented on this development, saying that many observers are unaware of just how powerful Amazon has become as a chip company.

To put this figure into perspective: Advanced Micro Devices (AMD), one of the world's most established semiconductor companies, had total annual revenue of approximately $34.6 billion in fiscal year 2025. Amazon's chip revenue thus already represents nearly 29 percent of AMD's total revenue, and if one considers only AMD's data center segment, the share is likely to be significantly higher than the often-cited 50 percent. Trainium2, with 1.4 million chips already shipped, forms the backbone of much of Amazon's Bedrock inference. Trainium3, unveiled at AWS Re:Invent in December 2025, offers four times the computing power, energy efficiency, and memory bandwidth of its predecessor, and initial customer tests indicate a reduction in AI training and inference costs of up to 50 percent. Trainium4, slated for release in 2027, is already under development

The strategic logic behind developing its own chips is immediately clear: By reducing its dependence on Nvidia, Amazon can offer its cloud customers lower costs while simultaneously protecting its own profit margin. OpenAI's explicit commitment as part of the deal to utilize two gigawatts of Trainium-based computing capacity is a tremendous vote of confidence in Amazon's silicon strategy and should further strengthen the Trainium ecosystem.

The 244 billion backlog as a built-in growth guarantee

An often underestimated indicator of a cloud provider's future profitability is its order backlog, also known as Remaining Performance Obligations (RPO). AWS reported a backlog of $244 billion at the end of the fourth quarter of 2025, a 40 percent year-over-year increase and a 22 percent increase compared to the previous quarter. This backlog represents multi-year contractual commitments from enterprise customers and AI-native companies and effectively underpins the expansion plan for 2026 and beyond.

In the context of the hyperscaler landscape, AWS is thus part of a larger dynamic in which the major cloud providers are collectively approaching a backlog of nearly one trillion dollars. Microsoft recently reported RPOs of 625 billion dollars, although around 45 percent of that, or 281 billion dollars, is attributable to a single customer: OpenAI. Oracle recorded 130 billion dollars, and Google Cloud around 90 billion. What distinguishes AWS is the relative diversification of its backlog. While Microsoft exhibits a concerning concentration of OpenAI, AWS's backlog is spread across a broad base of enterprise customers across various industries and regions.

Amazon's financial strength in the overall picture

Amazon's financial results for fiscal year 2025 illustrate why the company is in a position to make such large-scale strategic investments. Total revenue increased by 12.4 percent to $716.9 billion, operating income rose by 16.6 percent to $80.0 billion, and net income grew by 31.8 percent to $78.2 billion. Diluted earnings per share reached $7.17, a 29.7 percent increase year over year.

AWS, as a growth driver, accelerated to revenue growth of 24 percent in the fourth quarter of 2025, the fastest pace in 13 quarters, achieving quarterly revenue of $35.6 billion. This translates to an annualized run rate of approximately $142 billion. The cloud division's operating income for the fourth quarter was $12.5 billion. This means AWS alone generates more operating profit than many DAX-listed companies achieve in total revenue.

However, free cash flow is the Achilles' heel of this growth story. Operating cash flow was [missing value] in 2025, but capital expenditures consumed it, leaving only free cash flow. For 2026, Amazon has announced plans to increase capital expenditures to around $200 billion, the majority of which will go towards AWS data centers and AI infrastructure. This equates to spending of roughly $550 million per day. CEO Andy Jassy emphasized that these investments are a response to rising demand and not a speculative grab for revenue.

The valuation question as a value argument

In an environment where the valuations of the Magnificent Seven are frequently criticized as overheated, Amazon stands out with a remarkable valuation anomaly. Its current price-to-earnings ratio is [missing value], with a market capitalization of [missing value]. For comparison, Nvidia trades at a P/E ratio of 36, AMD at over 76. Even Microsoft, whose cloud growth is slower than that of AWS and whose backlog has a concerning OpenAI concentration, trades at a P/E ratio of around 25. [claim:2]

Crucial to the valuation is the inherent versatility of Amazon. The company is not only a cloud provider, but also the world's leading e-commerce operator, a growing advertising business, a streaming provider, a logistics network, and, more recently, a significant semiconductor manufacturer. The $50 billion investment in OpenAI and the equally profitable stake in Anthropic add another dimension: Amazon now has a strategic stake in the world's two leading AI labs, OpenAI and Anthropic, and simultaneously possesses the infrastructure on which these models are trained and operated.

Our US expertise in business development, sales and marketing - Image: Xpert.Digital

Industry focus areas: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry

More information here:

• Expert Business Hub

A thematic hub offering insights and expertise:

• Knowledge platform covering global and regional economies, innovation and industry-specific trends
• A collection of analyses, insights, and background information from our key areas of focus
• A place for expertise and information on current developments in business and technology
• A hub for companies seeking information on markets, digitalization, and industry innovations

Microsoft between dependence and distance

Microsoft's absence from the current funding round is a telling signal. The software giant, which has been OpenAI's closest partner for almost a decade and invested over $13 billion, is deliberately holding back. The reasons lie in what is increasingly perceived as a problematic dependency. Analyst Brent Thill of Jefferies put OpenAI's share of Microsoft's backlog at 45 percent of its $625 billion in outstanding commitments, a concentration he considers worrying. Microsoft's stock plummeted the day after the quarterly results, losing almost $360 billion in market capitalization.

Microsoft emphasizes that the existing partnership remains unchanged. Azure will continue to be the exclusive cloud provider for the APIs to OpenAI models and for the startup's own products. Exclusive licenses and access to intellectual property will also remain with Microsoft. However, the fact that OpenAI will now pay AWS $100 billion over eight years significantly dilutes this cloud exclusivity. AWS will be the exclusive third-party distributor of the Frontier enterprise platform, while Microsoft will retain only API exclusivity. This creates a dilemma for Microsoft: On the one hand, OpenAI is by far the most important growth driver of Azure's business; on the other hand, every additional OpenAI partnership with competitors gives the startup additional negotiating power and reduces its strategic dependence on Redmond.

Related to this:

• Not OpenAI, not Amazon: This is the real winner of the $38 billion deal: Nvidia

Nvidia's throttled ambitions

Nvidia's role in this funding round also warrants a more nuanced examination. In September 2025, the chipmaker announced in a letter of intent its intention to invest up to $100 billion in OpenAI to jointly build at least 10 gigawatts of computing capacity. However, CEO Jensen Huang backtracked in the following months, clarifying to journalists in Taipei that it had never been a binding commitment.

Behind the scenes, Huang expressed concerns to industry colleagues, according to the Wall Street Journal: He criticized a lack of discipline in OpenAI's business practices and voiced worries about growing competition from Google and Anthropic. Publicly, Huang denied any dissatisfaction and called the claims nonsense, but emphasized that the investment would not reach the originally stated sum. Ultimately, Nvidia's stake amounted to $30 billion, which is still the largest single investment in the chipmaker's history, but only a little less than a third of the originally anticipated amount. For Nvidia, the investment is a double-edged sword: On the one hand, it secures the relationship with one of its largest chip customers; on the other hand, a certain circular logic arises when the chip vendor simultaneously invests in its customer.

Amazon's double AI bet: OpenAI and Anthropic

What sets Amazon apart from all other investors is the fact that the company now holds strategic stakes in the world's two leading AI labs. Amazon has invested a cumulative $8 billion in Anthropic, the developer of the Claude model. In the fourth quarter of 2025, Amazon valued its Anthropic stake at $60.6 billion, a sevenfold increase in its invested capital. The company expects to realize a further book gain of $15 billion in the first quarter of 2026 when convertible bonds are converted into non-voting preferred stock. Anthropic is already valued at approximately $183 billion and is in talks about another funding round that could drive its value up to $350 billion.

In parallel, Amazon now holds an estimated seven percent stake in OpenAI and a significant stake in Anthropic, estimated at around eight percent. Both AI companies utilize Amazon's cloud infrastructure and its proprietary chips. Anthropic has committed to purchasing one million Trainium chips and, together with AWS, operates Project Rainier, one of the world's largest AI computing clusters. This dual strategy gives Amazon a unique position: Regardless of which of the two leading AI labs ultimately prevails, Amazon benefits from both its investments and the cloud and chip revenues generated by both companies. It's a classic platform strategy where the infrastructure provider wins, no matter who wins the application race.

The 200 billion euro investment offensive and the question of return on capital

Amazon's announcement that it plans to increase its capital expenditures to around $200 billion in 2026 initially caused nervousness on Wall Street. Investments already amounted to $131.8 billion in 2025, a 59 percent increase from $83 billion in 2024. The $200 billion mark represents a further increase of over 50 percent and significantly exceeds analysts' expectations, who had originally anticipated around $150 billion.

The majority of these investments are flowing into AWS data centers and AI infrastructure. In the fourth quarter of 2025 alone, AWS added more than one gigawatt of capacity. The key question for investors is whether the AI ​​economy can absorb this capital at acceptable returns. AWS CEO Matt Garman emphasized at the Re:Invent conference that the crucial metric for customers is value for money—how much computing power a dollar buys. If Trainium3 does indeed cut the costs of training and inference in half, this would not only protect Amazon's margins but also further stimulate demand for cloud capacity.

Nevertheless, the massive investment program is putting short-term pressure on free cash flow. The $139.5 billion in operating cash flow generated in 2025 was barely sufficient to cover the investments. With $200 billion in planned expenditures and a similar operating cash flow assumed, negative free cash flow is even a possibility in 2026. Amazon will argue that these investments are being channeled into long-term infrastructure control and will pay off in the form of sustainable cloud growth. The alternative—risking capacity bottlenecks while demand accelerates—would be far more dangerous for its market position.

The geopolitical dimension of the AI ​​infrastructure arms race

OpenAI's $110 billion funding round cannot be viewed in isolation from broader geopolitical dynamics. If OpenAI plans to invest roughly $600 billion in compute capacity by 2030, and if the major hyperscalers are collectively investing hundreds of billions annually in data centers, then this is an infrastructure offensive of historic proportions. The development of this capacity is taking place predominantly in the United States, reinforcing American dominance in the AI ​​value chain.

The average wait time for a power grid connection in North America has now reached four years, and vacancy rates in colocation data centers are at a historic low of 2.3 percent. At least 8 gigawatts of new capacity are under construction, 73 percent of which is already pre-leased. These physical constraints of power and space give companies like Amazon, which invested early in their own infrastructure, a structural competitive advantage that cannot be overcome in a single quarter.

What the skeptics overlook

The "peak Amazon" thesis, the claim that the company had passed its growth zenith, was essentially based on three arguments: the slowdown in e-commerce growth after the pandemic, the supposed margin compression due to high investments, and the narrative that AWS had missed the AI ​​train. All three arguments have proven to be miscalculations.

The e-commerce business is normalizing, but remains the global benchmark for digital commerce and generates stable cash flows that help fund the investment program. Margins may suffer in the short term due to the investment cycle, but operating income continues to grow robustly at 17 percent. And the claim that AWS missed the AI ​​train is refuted by the facts: 24 percent revenue growth, a $244 billion backlog, over $10 billion in chip sales, and now the exclusive cloud partnership with the world's largest AI brand.

What skeptics systematically underestimate is the optional nature of the advertising business. Amazon's advertising division has been growing at double-digit rates for years and benefits from advertisers increasingly shifting budgets from Google and Meta to Amazon's first-party data ecosystem. In a world where AI-powered product recommendations and purchase intent data are revolutionizing advertising effectiveness, Amazon's combination of retail data, cloud infrastructure, and access to AI models is a unique weapon in the advertising market.

The risks of an infrastructure supercycle

An honest analysis, however, must also identify the risks of this unprecedented investment cycle. The fundamental question is: Does the demand for AI justify the cumulative investments of hundreds of billions of dollars, or is this a classic overinvestment cycle where capacity is growing faster than monetizable use cases?

Economist Sebastian Mallaby of the Council on Foreign Relations points out that spending will increase massively in the coming years, with OpenAI likely consuming a total of $40 billion by 2028. The fact that OpenAI has revised its compute spending forecast from $1.4 trillion to $600 billion suggests that even internal recalibration is underway. If AI model development becomes less capital-intensive than anticipated due to algorithm-based efficiency gains, as demonstrated by DeepSeek, infrastructure providers could be sitting on excess capacity.

For Amazon, the risk is lower than for pure AI investors due to the diversification of its cloud customers and the increasing demand for traditional cloud services beyond pure AI workloads. The $244 billion backlog is not a hype indicator but represents contractually secured cash flows over several years. Even in a scenario of moderate AI disappointment, AWS would continue to grow robustly because the fundamental cloud migration of the enterprise world is far from complete.

The new power structure of artificial intelligence

The $110 billion funding round marks a turning point in the structure of the global AI industry. The existing duopoly of Microsoft and OpenAI is being complemented by a triangle in which Amazon acts as an equal infrastructure partner. Microsoft retains API exclusivity and the largest equity stake in OpenAI at 27 percent, but Amazon's role as the exclusive Frontier reseller and Trainium supplier gives the company a counterweight that fundamentally alters the dynamics of negotiations.

For the industry as a whole, this situation means a further consolidation of power among the few companies that possess both the capital and the physical infrastructure to operate AI at scale. Start-ups and medium-sized businesses are increasingly becoming consumers of this infrastructure without having the ability to compete on a level playing field. The three hyperscalers AWS, Azure, and Google Cloud together form an oligopoly that controls the AI ​​value chain from chips and data centers to modeling platforms.

With its $50 billion bet on OpenAI, its investment in Anthropic, its own chip development, and its massive infrastructure expansion, Amazon has positioned itself far beyond traditional retail. The company has become perhaps the world's most complete technology infrastructure provider: from silicon to cloud computing to modeling platforms, from end consumers to business customers to AI labs. Those who call this "Peak Amazon" are confusing a strategic investment cycle with a zenith. In reality, the rise of the infrastructure giant Amazon has only just entered its next phase.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your native language!

Konrad Wolfenstein

I and my team are happy to be available to you as your personal advisor.

You can contact me by filling out the contact form here or simply call me at +49 7348 4088 965. My email address is: [email protected]

I'm looking forward to our joint project.

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

Benefit from Xpert.Digital's extensive, five-fold expertise in a comprehensive service package | R&D, XR, PR & Digital Visibility Optimization - Image: Xpert.Digital

Xpert.Digital possesses in-depth knowledge across various industries. This allows us to develop tailored strategies precisely aligned with the requirements and challenges of your specific market segment. By continuously analyzing market trends and monitoring industry developments, we can act proactively and offer innovative solutions. The combination of experience and expertise generates added value and provides our clients with a decisive competitive advantage.

More information here:

• Benefit from Xpert.Digital's 5 areas of expertise in one package – starting from just €500/month