By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panelists generally agree that Amazon's AI strategy is promising, but they have differing views on its sustainability and profitability. While some see high-margin potential in AI infrastructure services and internal consumption driving retail margin expansion, others caution about capex intensity, cyclical demand, and lock-in risks.
Risk: Capex intensity and potential lock-in risks, as highlighted by Claude and ChatGPT.
Opportunity: The AI-retail flywheel effect, as proposed by Grok.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
When we talk about Amazon (AMZN), it is mostly about the product that we saw or ordered and whether they were decent or a waste of money. The retail business made Amazon an e-commerce giant, rewarding investors with gains of more than 15,700% over the last 20 years. Many still see Amazon as the website or app where they can simply order things. Very few may have noticed that Amazon has quietly transformed into a tech and artificial intelligence (AI) powerhouse that is becoming a full-stack AI infrastructure empire.
It may not be entirely correct to say that the Amazon investors knew it doesn’t exist anymore. Retail and e-commerce are still deeply embedded in the company’s core business. However, while retail made Amazon, AI and technology are building its future.
And its Q1 earnings reveal just how huge this transformation has become. Let’s dig in deeper.
Amazon’s Retail Business Is Still Going Good, but It Is No Longer the Whole Story
Amazon reported first-quarter revenue of $181.5 billion, up 17% year-over-year (YoY), while adjusted earnings rose 75% to $2.80 per share. Its retail business still continues to expand rapidly, with unit growth of 15% YoY, which marked the highest level since the later stages of Covid-19 lockdowns. Amazon also added more than 600 notable new brands during the quarter. Its grocery business exceeded $150 billion in gross sales in 2025, making it the second-largest grocer in the U.S.
CEO Andrew Jassy made it clear that Amazon now sees itself at the center of one of the largest technology shifts in decades. According to him, the company has “never seen a technology grow as rapidly as AI.” AI is now embedded across Amazon’s entire business, driving a 28% YoY increase in its cloud computing business, AWS. Management said AWS’s AI business has already reached a revenue run rate of more than $15 billion within just three years of the current AI boom, making it nearly 260 times larger than AWS’s cloud business was in its early years.
Amazon Is Becoming an AI Infrastructure Empire
What separates Amazon from other tech companies is that it is not just building AI applications. It is, in fact, trying to control every layer of the AI infrastructure stack. Its AI ecosystem now includes model-building tools through SageMaker, AI inference services, frontier AI models on Bedrock, enterprise AI agents, custom AI chips, data center infrastructure, cloud computing services, and agentic AI platforms.
Amazon Bedrock has become one of the company’s most important AI products, with nearly 80% of Fortune 100 companies using it. Its biggest weapon might be chips. While most investors associate AI chips with Nvidia (NVDA), Amazon revealed that its "custom silicon business is now one of the top three data center chip businesses in the world." Its chip business grew nearly 40% sequentially during Q1, with an annual revenue run rate now exceeding $20 billion and growing at triple-digit percentages YoY. Amazon also stated that if its chip business operated independently and sold chips externally like traditional semiconductor companies, its annual revenue run rate would already be approximately $50 billion.
Its Trainium AI chips are becoming important to its AI strategy, with over $225 billion in revenue commitments tied to Trainium. The company revealed that Trainium3, which started shipping in 2026, has already seen nearly all of its capacity booked by customers. Meanwhile, Trainium4, which is still 18 months away from availability, has already been heavily reserved. Amazon has already secured multiyear, multi-gigawatt training commitments from both OpenAI and Anthropic, while companies like Uber (UBER) are increasingly adopting Trainium as well.
Amazon, like Intel (INTC), feels that AI is more than simply a GPU-driven opportunity. It is also dramatically increasing CPU demand. The company revealed that Meta alone has committed to using tens of millions of its Graviton CPUs for AI workloads. What’s more, its autonomous vehicle business, Zoox, has now driven nearly 2 million miles and transported more than 350,000 riders. Meanwhile, its satellite internet business, Amazon Leo, is almost ready for commercial launch. Delta Air Lines (DAL), JetBlue (JBLU), AT&T (T), Vodafone (VOD), NASA, and other large clients have already made commitments to the company.
The Amazon Investors Once Knew Is Evolving Into Something Much Bigger
Many investors continue to associate Amazon with retail. But these developments show that Amazon has expanded far beyond traditional retail. It is becoming a global AI infrastructure provider, a semiconductor company, a cloud giant, an advertising platform, a healthcare technology business, a logistics network, a robotics operator, a satellite communications player, and an enterprise AI ecosystem all at the same time. And these could shape what Amazon will become in the next decade.
With such a diversified business, I believe AMZN is an excellent buy-and-hold stock for the long haul.
Overall, on Wall Street, AMZN stock has earned a consensus “Strong Buy.” Of the 57 analysts covering the stock, 48 have a “Strong Buy,” six have a "Moderate Buy" rating, and three analysts rate the stock as a “Hold.” AMZN stock has climbed 18% year-to-date (YTD). But its mean target price of $314.69 implies 14% potential upside from current levels. Plus, the high target price of $370 suggests that shares could climb as much as 34% over the next year.
On the date of publication, Sushree Mohanty did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Barchart.com
Four leading AI models discuss this article
"Amazon's transition from a retail-first company to an AI infrastructure utility is creating a defensive moat that will allow them to capture the majority of the value in the enterprise AI stack."
Amazon is successfully executing a vertical integration strategy that mirrors the 'full-stack' dominance of the early cloud era, but applied to AI. By controlling the silicon (Trainium/Graviton), the orchestration layer (Bedrock), and the application environment (AWS), they are effectively creating a walled garden that increases customer switching costs. The $20B run rate in custom silicon is the real story; it signals that Amazon is successfully commoditizing the hardware layer to protect its cloud margins from Nvidia's pricing power. While retail provides the cash flow, the valuation re-rating will be driven by the sustainability of these high-margin AI infrastructure services as they scale beyond early adopters.
The massive capital expenditure required to maintain this 'full-stack' lead could lead to margin compression if the AI demand cycle hits a plateau or if open-source models render proprietary infrastructure less critical. Furthermore, Amazon's reliance on custom silicon assumes they can maintain pace with Nvidia's R&D cycle, which is a high-stakes gamble that could leave them with obsolete inventory if they miscalculate the next architectural shift.
"AMZN's vertical AI stack and $20B+ chip run-rate create durable cost edges, targeting AWS margin expansion to 40%+ by 2027."
Amazon's Q1 crushed: $181.5B revenue (+17% YoY), EPS $2.80 (+75%), with AWS AI run-rate at $15B and custom chips at $20B ARR (40% sequential growth). The full-stack AI bet—SageMaker, Bedrock (80% Fortune 100 adoption), Trainium chips ($225B commitments)—positions AMZN to undercut NVDA/others on costs, boosting AWS margins toward 40% as Trainium3/4 scale in 2026+. Retail/grocery strength ($150B+ sales) provides cashflow ballast. Unlike MSFT's app focus, AMZN owns infra layers, enabling 20-25% EPS CAGR if AI capex yields. Undervalued at ~35x forward P/E vs. growth.
AWS overall growth slowed to 17% (from 30%+ peaks), with AI still <20% of revenue; massive data center buildout risks capex overruns if Trainium adoption falters amid NVDA/MSFT dominance. Retail's thin margins remain exposed to consumer slowdowns.
"Amazon's AI infrastructure story is real but the article mistakes bookings and internal consumption for external revenue, while ignoring that record capex spending is masking margin pressure in the near term."
The article conflates revenue run-rate with actual profitability. AWS's $15B AI revenue run-rate sounds massive until you remember AWS overall generates ~$100B annually at ~30% operating margins — the AI subset's margin is undisclosed and likely lower given infrastructure capex intensity. Amazon's chip business at '$50B run-rate if independent' is pure fiction: that's internal consumption, not external revenue. The Trainium commitments ($225B) are multi-year bookings, not cash. Most critically: the article ignores that Amazon is now in a capex arms race with Meta, Microsoft, and Google. Q1's 75% EPS growth masks that capex as a % of revenue is accelerating. The retail business's 15% unit growth is solid but mature; the 'AI empire' thesis requires years of capex before meaningful incremental profit emerges.
If Amazon's custom chips and Bedrock truly capture 80% of Fortune 100 adoption and Trainium capacity is fully booked years in advance, the company may have already won the infrastructure layer — making current valuation a bargain despite near-term capex drag.
"AMZN's AI infrastructure push could unlock durable upside if AI demand and profitability hold up, but the risk remains that execution, competition, and capex reality could derail the thesis."
While the article markets AMZN as an AI infrastructure empire, there are three notable gaps. First, AI demand remains cyclical; AWS growth could decelerate if enterprise budgets tighten or price competition intensifies, risking margin pressure. Second, the big-ticket claims around a 15 billion dollar AI run rate and Trainium bookings may be optimistic or rely on multi-year commitments that could slip. Third, the plan is capital-intensive across chips, data centers, and satellites, with execution risk and geopolitical headwinds. Retail remains the near-term earnings engine, so the stock's upside depends on sustained profitability and free cash flow from the AI push.
Strongest counter: AI demand may not be durable; AWS margins could compress if chip costs rise or pricing wars intensify, and Trainium bookings might prove overstated or cancelation-prone.
"Amazon's AI infrastructure play is a commodity utility trap unless Bedrock achieves significant software-layer differentiation."
Claude is right to flag the 'run-rate' obfuscation, but misses the real risk: Amazon isn't just fighting a capex war, they are fighting a software-defined lock-in battle. If Bedrock doesn't achieve high-margin stickiness, they are just a utility provider with depreciating hardware assets. The $225B in commitments is a vanity metric if those customers churn to Azure or GCP when their proprietary models outperform Amazon's. Amazon’s retail margins are the only real safety net here.
"Amazon's AI stack creates a retail flywheel boosting margins via internal applications, overlooked by panel."
Everyone obsesses over AWS capex drag, but ignores the AI-retail flywheel: Trainium/Bedrock powers next-gen warehouse robotics and demand forecasting, potentially expanding retail op margins from 3% to 6%+ ($15B+ incremental EBIT at scale). Claude calls retail 'mature,' but that's the underrated AI leverage point others miss—internal consumption of infra drives 2-3x ROI vs external sales.
"Internal AI-retail synergies are real but unproven at scale; claiming 2-3x ROI without disclosed fulfillment center Trainium penetration rates is premature."
Grok's retail-AI flywheel is intriguing but needs pressure-testing: warehouse robotics ROI assumes Trainium adoption at scale, yet Amazon's own fulfillment centers are still piloting these systems. A 3% → 6% margin expansion ($15B EBIT) requires flawless execution across logistics, not just chip availability. More critically: if this internal ROI is as strong as Grok claims, why hasn't Amazon already saturated internal Trainium demand, making external sales the marginal revenue driver? The flywheel thesis only works if internal consumption doesn't cannibalize external pricing power.
"Bedrock lock-in risk and pricing pressure threaten external AI revenue, potentially offsetting any capex-driven margin benefits."
Claude raises the bookings vs cash concern, but the bigger miss is the self-reinforcing lock-in risk of Bedrock. Even with 80% Fortune 100 adoption, external revenue hinges on pricing power; if enterprises push back on platform fees or switch to open models, revenue collapses while capex continues. Also, the AI hardware race remains Nvidia-dominant and faces regulatory scrutiny; Bedrock-only moat could erode fast if Nvidia expands its ecosystem or if open-source models undercut AWS pricing.
The panelists generally agree that Amazon's AI strategy is promising, but they have differing views on its sustainability and profitability. While some see high-margin potential in AI infrastructure services and internal consumption driving retail margin expansion, others caution about capex intensity, cyclical demand, and lock-in risks.
The AI-retail flywheel effect, as proposed by Grok.
Capex intensity and potential lock-in risks, as highlighted by Claude and ChatGPT.