5-Min Brief: Big Tech Just Reported Earnings. The AI Spending Numbers Are Hard to Comprehend.

What you need to know — in 30 seconds

• Google, Amazon, Meta, and Microsoft all reported first-quarter earnings this week and every single one raised its AI infrastructure spending forecast for 2026
• Combined, the four companies are now expected to spend more than $700 billion on AI this year — data centers, chips, and computing power
• Wall Street analysts now project that number could cross $1 trillion in 2027
• The companies cutting jobs while making these investments is not a coincidence — it's the same trade-off playing out in real time

Once a quarter, the biggest technology companies in the world have to tell investors exactly how much money they're making and spending. This week was that moment for 2026's first quarter — and the AI spending numbers that came out of it are genuinely hard to get your head around.

Let's try anyway.

The numbers, company by company

Google (Alphabet): Raised its 2026 AI infrastructure spending forecast to between $180 and $190 billion — up $5 billion from previous guidance. Google Cloud revenue grew 63% year over year to $20 billion, more than doubling its growth rate. The company said its Cloud backlog — work already committed but not yet billed — has nearly doubled quarter over quarter to $460 billion. CEO Sundar Pichai said AI is "lighting up every part of the business" and that demand for AI computing is "unprecedented." Alphabet was the only one of the four whose stock rose after earnings, climbing more than 6%.

Amazon: Projected 2026 AI infrastructure spending of around $200 billion — the largest absolute number in the group. Amazon Web Services grew 28% year over year, its fastest pace in 15 quarters. AWS is now adding capacity faster than at any point in its history, and management confirmed demand continues to exceed what they can supply.

Meta: Raised its 2026 spending forecast to between $125 and $145 billion — up from a previous range of $115 to $135 billion. Meta cited higher memory chip costs and additional data center capacity as the drivers. Investors didn't love it — Meta's stock dropped more than 6% after hours despite revenue growing 33% to $56 billion. This is the same company that announced 8,000 layoffs last week to "offset" its AI investments. The numbers now make clear exactly what it's offsetting.

Microsoft: Set its full 2026 AI infrastructure spending at $190 billion — a figure that stunned analysts who had expected around $152 billion. CFO Amy Hood attributed about $25 billion of the increase to rising memory chip and component costs. Azure grew 40% year over year, and Microsoft's AI business has now crossed a $37 billion annual revenue run rate — up 123% year over year. Microsoft warned it still expects to remain capacity constrained through the end of 2026.

Add those together and you get more than $700 billion in planned AI infrastructure spending from just four companies in a single year — up 77% from last year's combined total of $410 billion.

What is all this money actually buying?

Data centers — enormous buildings full of specialized computer chips — are the foundation of everything AI runs on. Every time you ask ChatGPT a question, Claude helps you draft an email, or Gemini summarizes a document, that interaction runs on computing hardware inside one of these buildings.

The chips that power AI — primarily made by Nvidia, though other manufacturers are catching up — are extraordinarily expensive and in extraordinarily short supply. Every major tech company is competing to lock in chip supply years in advance and build the physical infrastructure to house them.

Memory chips in particular emerged as an unexpected pressure point in this week's earnings — both Meta and Microsoft specifically called out rising memory costs as a driver of their higher spending forecasts. AI models require enormous amounts of fast memory to operate, and demand is currently outpacing supply.

What the earnings reports confirmed this week is that this isn't slowing down. Google's CFO said the company expects 2027 spending to "significantly increase" compared to 2026. Microsoft said the same. The investment cycle is not close to its peak.

The $1 trillion question

Wall Street analysts from Evercore and Bank of America both now project that AI infrastructure spending across the tech industry will exceed $1 trillion in 2027 — a number that would have seemed science fiction five years ago.

To put that in perspective: $1 trillion is more than the GDP of the Netherlands, more than Apple's annual revenue, and more than the US government spent on defense in 2024.

The obvious question — one that investors are increasingly asking loudly — is: when does this pay off? Spending $700 billion in a year requires generating more than $700 billion in return to make economic sense. Right now, AI revenue is real and growing fast. Microsoft's AI business tripled. Google Cloud's growth rate doubled. But the revenue is not yet close to matching the capital being deployed.

The companies making these bets are wagering that AI becomes the infrastructure layer for the entire global economy — that every business, every software application, every workflow eventually runs on top of AI in some form. If that happens, the economics work out. If it doesn't, or if it takes longer than expected, the math gets complicated. The stock drops after Meta's and Amazon's reports suggest some investors are starting to ask that question more urgently.

What this week's earnings mean alongside last week's layoffs

Last week we covered Meta cutting 8,000 jobs to help fund its AI spending. This week we see the full picture of what that spending looks like across the industry.

These aren't separate stories. They're the same story: companies are making an enormous, simultaneous bet on AI as the future of their businesses. They're reducing costs in areas where AI can absorb the work — headcount, operational overhead — and redirecting those resources into the infrastructure that makes AI possible.

Whether that bet pays off for shareholders, for workers displaced in the process, and for society broadly is the defining business question of the next decade. What this week's earnings confirmed is that the bet is now fully placed. There's no version of events where these companies pull back.

HumanReadable-AI covers AI news in plain English every weekday. Subscribe below — free, no jargon, always under five minutes.