Big Tech Is Spending $700 Billion. These Companies Get Paid.
Johannes Gutenberg printed his first Bible in 1455 and effectively went bankrupt.
His financier, a goldsmith named Johann Fust, took over Gutenberg’s press, his inventory, and the entire business. Within a generation, Fust’s heirs (and a network of Dutch and Venetian printer-bankers who followed them) had turned movable type into one of the most profitable industries in Europe.
While Gutenberg got the credit, the vast majority of wealth flowed to those who capitalized on the idea. Every transformative technology unfolds the same way: Someone builds the marvel; someone else owns the supply.
And last week, four of the largest companies on the planet confirmed which side of that split is about to get very, very rich…
Four Big Tech Earnings Reports, One Clear Message
Every quarter, Wall Street asks the same question: is the AI spending boom real, or is it a story companies are telling to justify valuations? Last week, Microsoft (MSFT), Alphabet (GOOGL), Amazon (AMZN), and Meta (META) answered that question definitively.
Together, they committed more than $700 billion in 2026 capital spending to build out AI infrastructure, explicitly guiding for that number to climb again in 2027.
The picks-and-shovels trade is the visible part of this cycle, but what forms around it (what gets built on top of all that infrastructure once it goes online) is where the next phase of the AI fortune gets made. And it leads directly to a project Elon Musk has been waiting 27 years to launch. We’ve got the full story for you here.
But the important takeaway from Big Tech’s earnings is what they confirmed about the trajectory of the entire AI buildout.
Microsoft Earnings: AI Demand Is Exploding
Microsoft reported earnings that make the AI bull case look conservative.
Its AI business surpassed $37 billion in annualized revenue, growing 123% year-over-year. And its Cloud operations exceeded $54 billion in quarterly revenue, up 29%. Azure — Microsoft’s cloud computing platform — grew 40%, its fastest rate in years, even as it lapped a strong prior year.
The application layer is just as impressive. Weekly engagement with Microsoft 365 Copilot — the AI assistant embedded inside Microsoft’s Office suite — has reached the same level as Outlook, meaning people are using it as habitually as email. It now has over 20 million paid subscriptions, with seat additions up 250% year-over-year. Accenture (ACN) alone accounts for 740,000 seats. Johnson & Johnson (JNJ), Bayer, Mercedes, and Roche each committed to 90,000 seats or more.
Microsoft is spending accordingly. Capital expenditures for Q4 alone will exceed $40 billion. For calendar year 2026, Microsoft expects to invest roughly $190 billion in infrastructure. Its remaining performance obligation — essentially, revenue already contracted but not yet recognized — hit $627 billion. And the company’s message on future spending was unambiguous: demand continues to exceed available capacity, and it is building to close that gap.
Alphabet Earnings: AI Is Driving Real ROI
Alphabet — Google’s parent company — reported one of the cleanest AI ROI quarters in history.
For the very first time, Google Cloud hit $20 billion in revenue, growing 63% year-over-year. Enterprise AI solutions became Cloud’s primary growth driver. And revenue from products built on Google’s AI models grew nearly 800% year-over-year.
The demand curve here is getting steeper. Google Cloud’s backlog nearly doubled in a single quarter, reaching $462 billion — half a trillion dollars in contracted but not-yet-recognized revenue, accumulated in essentially 90 days. Just over half of that will convert to revenue within the next 24 months.
The most persistent AI bear thesis was that AI-powered search would cannibalize Google’s ad business. This quarter demolished it. Search queries are at all-time highs. AI Overviews and AI Mode are driving more searches, including more commercial queries that attract premium advertiser rates. Search revenue grew 19%, hitting $60 billion in a single quarter.
On future spending, Alphabet raised its 2026 CapEx guidance to $180- to $190 billion and explicitly guided for 2027 CapEx to “significantly increase” compared to 2026. That is corporate-speak for: we are nowhere near the peak.
Amazon Earnings: The Hidden AI Chip Story
Amazon delivered exactly what analysts expected from Amazon Web Services (AWS) — and then buried the most important number in the chip business.
AWS grew 28% year-over-year — its fastest rate in 15 quarters — on a $150 billion annualized revenue base.
Growing 28% when you are already a $150-billion business should not be possible. It is happening anyway.
But the most underappreciated story in Amazon’s quarter is not AWS. It is chips. Amazon’s custom silicon business — primarily its Trainium AI chip and Graviton CPU — has an annualized revenue run rate of over $20 billion, growing triple digits. If Amazon accounted for internally consumed chips the way standalone chip companies do, the revenue run rate would exceed $50 billion. It built that in approximately five years.
Why does this matter for investors? Because Amazon’s Trainium chip offers roughly 30% to 40% better price performance than comparable Nvidia GPUs. Amazon expects Trainium to save tens of billions of dollars per year in CapEx and provide several hundred basis points of operating margin advantage versus third-party chips. That is a structural cost advantage that widens with every new chip generation. Trainium2 is largely sold out. Trainium3, which just started shipping, is nearly fully subscribed. And much of Trainium4 — still 18 months away from broad availability — has already been reserved. The company has $225 billion in total Trainium revenue commitments.
First-quarter CapEx measured $43.2 billion. Moreover, on the company’s earnings call, CEO Andy Jassy called AI “a once-in-a-lifetime opportunity where every application that we know of is going to be reinvented” and pledged to “invest a significant amount of capital over the coming years.”
Meta Earnings: AI Is Fueling Engagement and Ads
The AI story at Meta is different from these other three hyperscalers.
Microsoft, Alphabet, and Amazon primarily monetize AI by selling infrastructure and software to other businesses. Meta’s AI ROI flows through engagement improvement — better recommendations mean people spend more time on the platform, which generates more ad impressions and, thus, more revenue. It is an indirect mechanism; but for 3.5 billion daily users, it is extraordinarily powerful.
The company reported $56.3 billion in quarterly revenue, up 33% year-over-year, with a 41% operating margin. Family of Apps ad revenue grew 33%, driven by both a 19% increase in ad impressions and a 12% increase in the price per ad. Total video time on Facebook increased more than 8% globally in Q1 — the largest quarterly gain in four years. Instagram saw a 10% lift in real-time user engagement from ranking improvements alone.
Meta also launched Muse Spark, the first model from its newly formed Meta Superintelligence Labs — and notably, the first closed-source model in Meta’s history, marking a clean break from the open Llama strategy. That shift signals what the lab is actually for: Meta is not building AI tools for developers to remix. It is building the infrastructure for what Zuckerberg calls “personal superintelligence for everyone” — AI integrated directly into the daily lives of 3.5 billion users across every Meta platform and device. Weekly business AI conversations grew 10x in a single quarter to over 10 million.
The capital commitment number is staggering: Meta’s contractual obligations increased by $107 billion in a single quarter through multiyear cloud deals and supply chain agreements. The company raised its 2026 CapEx guidance to $125- to $145 billion, up from $115- to $135 billion, citing higher memory prices as the primary driver. And management acknowledged that they have ‘continued to underestimate’ their compute needs — a remarkable admission from a company that has been aggressively ramping capacity for two straight years.
AI Infrastructure Stocks Set to Benefit
Collectively, these four hyperscalers have committed $700 billion-plus in 2026 capital expenditures to build the infrastructure driving the AI boom — and even more in 2027.
Spending at this scale creates powerful structural tailwinds for every company on the receiving end of that spending.
This is not a sector-specific story. It’s a macroeconomic force.
When Microsoft spends $190 billion on infrastructure, that money flows throughout the AI hardware stack: GPUs, custom chips and networking, semiconductor manufacturing equipment, power and thermal management…
Those companies then hire more people, pay suppliers, and generate their own profits, creating a multiplier effect throughout the broader economy. It is why corporate earnings growth keeps exceeding expectations. And it is why the stock market, despite tariff fears and geopolitical noise, keeps finding reasons to rise.
The companies on the receiving end of that $700 billion-plus?
Semiconductors and Chip Equipment
- Nvidia (NVDA) remains the primary beneficiary of AI compute spending, supplying GPUs to all four hyperscalers.
- Broadcom (AVGO) is building custom AI chips for both Google (TPUs) and Meta (via its AI-specific silicon partnership), and its networking solutions are critical to every major data center.
- Marvell Technology (MRVL) is building custom chips for Amazon and Microsoft.
- AMD (AMD) is gaining share in both AI training and CPU markets, with Meta rolling out significant AMD deployment alongside new Nvidia systems.
- Micron Technology (MU) is the primary beneficiary of the memory price surge that every hyperscaler cited as a CapEx headwind — higher memory prices are good for memory manufacturers.
- Applied Materials (AMAT), KLA Corporation (KLAC), Teradyne (TER), and Lam Research (LRCX) supply the equipment used to manufacture every chip going into these data centers.
- Monolithic Power Systems (MPWR) provides power management semiconductors critical to the efficiency of AI compute.
- Sandisk (SNDK) and Seagate (STX) benefit from the storage buildout accompanying every new data center.
Power and Cooling Providers
Bloom Energy (BE), along with Eaton (ETN) and Vertiv (VRT), sits at the center of what may be the most acute infrastructure bottleneck in the entire AI buildout: power. U.S. data center energy demand is projected to nearly double between 2025 and 2028 — from 80 to 150 gigawatts — the equivalent of adding another Spain to the American grid in three years. Microsoft added a full gigawatt of capacity this quarter and expects to double its total footprint within two years. Every one of those gigawatts requires the transformers, switchgear, and thermal management systems that Eaton and Vertiv supply — and the onsite power generation that Bloom provides.
Data Center Networking
The interconnects market — the networking hardware that links data center components together — is a critical and often overlooked beneficiary.
Credo Technology (CRDO) provides copper interconnects — the dominant near-term solution as data centers scale.
Lumentum (LITE) and Coherent (COHR) build the longer-term solution: optical interconnects, which use light instead of electricity to transmit data, offering superior bandwidth, lower latency, and reduced power consumption.
No matter the investment window, all are positioned to benefit from the data center buildout.
The Bottom Line: Follow the AI Spending
The AI trade is not a momentum trade built on narrative. It is a fundamental trade anchored in the largest capital investment cycle in the history of technology, validated by real revenue, real margins, and real customer commitments.
And the companies on the receiving end of that spending remain the most compelling investment opportunity in the market.
So long as the hyperscalers keep spending — and they just emphatically confirmed that they will — the economy keeps expanding, the market keeps pushing higher, and the AI infrastructure names keep working.
That’s the visible part of the cycle.
But what matters just as much is what forms around it.
Because once the buildout is underway, the leverage shifts to the systems that sit on top of it — especially the ones that control how capital moves.
That’s where our attention is now… And it leads directly to what Elon Musk is building inside X.
