By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel's discussion centered around the sustainability of high valuations for AMZN and GOOGL, driven by their significant investments in AI and data centers. While some panelists (Grok) argued that these companies' energy independence strategies will sustain returns, others (Gemini, Claude, ChatGPT) expressed concerns about the capital intensity, regulatory risks, and potential margin compression due to competition and energy costs.
Risk: Energy costs and grid constraints could erode margins and lower returns on invested capital for AMZN and GOOGL, potentially compressing their high multiples.
Opportunity: Successful execution of energy independence strategies could provide a durable edge, sustaining returns on invested capital above 15% even as capex peaks.
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 →
Amazon.com, Inc. (NASDAQ:AMZN) was one of the stocks on Jim Cramer’s radar as he highlighted AI winners to buy for 2026. Cramer expects huge returns on the company’s data center investments, as he remarked:
Data centers are about to give the believers extraordinary returns. I mean, unbelievable returns. The money’s worth it. We just can’t tell it yet because right now there are only two publicly traded winners: Amazon and Alphabet. After talking to Andy, I now believe that both Amazon and Alphabet have ensured years and years of colossal profits because they’re spending tens of billions now to make hundreds of billions later.
Photo by Sunrise King on Unsplash
Amazon.com, Inc. (NASDAQ:AMZN) sells consumer goods and digital content through online and physical stores, provides advertising and subscription services, operates Amazon Web Services for cloud computing, develops electronic devices, produces media content, and offers programs supporting third-party sellers and content creators. Cramer mentioned the company during the May 4 episode and said:
I’m talking about Amazon and its CEO, Andy Jassy. Higher interest rates can fell many a company, but if you want to guess who’d be the last man standing, you could do a lot worse than betting on Amazon with a stock that rallied $3.79 today. If there were a Kalshi bet about which company could thrive with a crimped consumer, it will somewhat oddly be Amazon, because their goal is always to keep prices as low as possible, making the ultimate trade-down play.
There’s a reason we own this one for the Charitable Trust. I’m in awe of how Amazon’s become all encompassing in so many aspects of our lives today… When I hear things like that, I try to figure out how much Amazon means to America, means to you. No one has ever been that big a factor to our growth since Standard Oil got broken up for monopolizing the oil market over a century ago…
The tech giant reported a really fabulous quarter last week, driven by their booming Amazon Web Services business for cloud infrastructure and AI. While this is still one of the largest retailers in the world, AWS alone has an annual revenue run rate of $150 billion. And they keep adding more to the story on a pretty regular basis. From setting up their own low-earth orbit satellites for internet service to offering their supply chain services to anyone who wants it, it’s at the heart of the computer-driven economy.
While we acknowledge the potential of AMZN as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
Four leading AI models discuss this article
"Amazon's valuation is increasingly sensitive to the efficiency of its AI CapEx, which faces significant execution risk if enterprise AI adoption fails to accelerate revenue growth."
Cramer’s thesis relies on the assumption that massive CapEx in data centers acts as a moat, but he ignores the diminishing returns of AI infrastructure spending. While AWS is a cash cow, its $150B run rate faces intense margin pressure from Microsoft’s Azure and specialized GPU-cloud providers. Amazon’s retail segment is currently subsidizing this AI build-out; if consumer spending hits a wall, the 'trade-down' benefit won't offset the drag of massive depreciation costs from these data centers. Investors are pricing in perfect execution, ignoring that 'hundreds of billions' in future revenue requires a breakthrough in AI-driven enterprise productivity that has yet to materialize at scale.
If Amazon successfully leverages its proprietary Trainium/Inferentia chips to decouple AWS margins from Nvidia’s pricing, their cost-advantage could create a structural barrier that competitors cannot replicate.
"AWS's operating margins hitting 37% and $150B run-rate validate Cramer's thesis, but execution on $100B+ capex is make-or-break for re-rating above 45x P/E."
Cramer's hype on AMZN and GOOGL hinges on AWS and Google Cloud's data center capex—tens of billions annually—poised for 'hundreds of billions' later via AI demand. AWS's $150B run-rate (up ~17% YoY last quarter) drives 40%+ of AMZN's operating income despite retail drag, with margins expanding to 37% on efficiency gains. AMZN trades at 40x forward P/E (vs. 15% EPS growth consensus), pricing in perfection amid low-earth orbit bets like Kuiper. GOOGL's similar path, but MSFT's Azure/OpenAI lead narrows the 'only two winners' claim. Resilient in high rates, but consumer trade-down masks e-comm weakness.
AI capex could prove a bubble if demand disappoints—AMZN's $75B 2024 spend (2x revenue growth) risks write-downs like Meta's 2022 pivot, eroding free cash flow. Regulatory scrutiny on Big Tech dominance intensifies under potential Trump admin.
"AWS and Google Cloud will likely remain highly profitable, but the article conflates profitability with stock returns — and current multiples already embed significant AI upside, leaving little margin for error on capex ROI or monetization timing."
Cramer's thesis rests on a specific bet: that AWS and Google Cloud's capex spending today (tens of billions annually) converts to 'hundreds of billions' in profits later. The math isn't absurd — AWS already runs at ~$150B revenue with reported operating margins near 30%. But the article conflates two separate claims: (1) these companies will be profitable, which is likely true, and (2) returns will be 'extraordinary' and 'unbelievable.' That's a valuation claim, not a cash-generation claim. AMZN trades at ~50x forward earnings; GOOGL at ~23x. The capex thesis doesn't automatically justify current multiples unless AI adoption accelerates faster than consensus expects or margins expand beyond historical norms. The article also omits: rising competition (Microsoft, Oracle, AWS losing share to smaller players in certain workloads), regulatory risk, and the possibility that capex ROI disappoints if AI monetization stalls.
If data center returns disappoint — either because AI adoption is slower than priced-in or because capex ROI falls short of historical cloud margins — both stocks could re-rate sharply lower despite 'colossal profits' in absolute terms. Cramer is also selling conviction without quantifying what 'years and years' of profits means at current valuations.
"The profits implied by AWS and Google Cloud depend on an uninterrupted AI compute cycle and sustained pricing power, which is doubtful given capex intensity, competition, and regulatory risk."
Despite the hype around AMZN and GOOGL as AI winners, the article glosses over critical risks. The claim that tens of billions in capex yield hundreds later relies on an uninterrupted, margin-rich AI compute cycle that may not materialize quickly. Hyperscale cloud margins are at risk of compression from ongoing capacity buildouts and price competition (MSFT, ORCL, others). Regulatory and data-privacy risk rises in the US/EU; energy costs and cooling challenges bite returns; macro headwinds could delay enterprise cloud spend and AI adoption. The onshoring narrative is uncertain, and valuations may already reflect best-case outcomes.
But if AI adoption accelerates unabated and pricing power remains durable, AWS and Google Cloud could surprise on margins even with capex, making the upside more than priced in.
"The transition to energy-intensive AI infrastructure will permanently compress ROIC and increase capital intensity beyond current market expectations."
Claude is right to highlight the valuation-versus-cash-generation gap, but everyone is ignoring the 'energy bottleneck.' AWS and Google aren't just building chips; they are becoming de facto utility companies. If grid constraints or environmental regulations force them to build proprietary power infrastructure, the capital intensity will far exceed current depreciation estimates. This isn't just a software margin game anymore; it’s a capital-intensive infrastructure play that will permanently lower ROIC (Return on Invested Capital).
"Hyperscalers' nuclear and SMR power deals create an unmatchable energy moat, protecting ROIC amid grid constraints."
Gemini, energy bottlenecks aren't a new crisis--AMZN secured 960MW nuclear capacity via its June 2024 Talen deal, while GOOGL partners with Kairos Power for SMRs by 2030. This forward power contracting flips the script: it builds a supply moat competitors can't match without similar scale, sustaining ROIC above 15% even as capex peaks. Others fixate on costs but miss this energy independence as a durable edge.
"Energy independence is necessary but insufficient to justify current valuations—it's a cost floor, not a margin expander."
Grok's energy moat argument is clever but overstates durability. A 960MW nuclear deal and future SMR partnerships don't solve the core problem: energy costs remain a variable input that scales with compute demand. Competitors can replicate power contracting at similar scale—it's capital, not IP. The real question Grok dodges: does energy independence improve ROIC enough to justify 40x forward P/E on AMZN when consensus EPS growth is 15%? Power deals are table stakes, not a structural edge.
"Energy-cost risks and regulatory headwinds can erode the supposed durable moat from power agreements, threatening margin resilience and high multiples unless AI monetization accelerates."
Gemini's energy moat claim assumes power contracts alone sustain ROIC; but scale brings grid, policy, and price volatility risks that can erode margins faster than capex efficiency gains. A 960 MW nuclear deal is not a permanent shield—regulators, carbon pricing, and energy-price spikes could raise costs or force expensive hedges. If energy costs become a material drag, high multiples on AMZN/GOOGL may compress unless AI monetization accelerates meaningfully.
The panel's discussion centered around the sustainability of high valuations for AMZN and GOOGL, driven by their significant investments in AI and data centers. While some panelists (Grok) argued that these companies' energy independence strategies will sustain returns, others (Gemini, Claude, ChatGPT) expressed concerns about the capital intensity, regulatory risks, and potential margin compression due to competition and energy costs.
Successful execution of energy independence strategies could provide a durable edge, sustaining returns on invested capital above 15% even as capex peaks.
Energy costs and grid constraints could erode margins and lower returns on invested capital for AMZN and GOOGL, potentially compressing their high multiples.