Skip to content 
Meta just posted revenues of $51.2 billion. The company beat analyst expectations. Its business is growing at 26% year-on-year. You’d think Wall Street would be celebrating, right?
Wrong.
Meta’s stock crashed over 8% in after-hours trading, wiping out a staggering $160 billion in market value. That’s roughly the entire GDP of Hungary vanishing in a single evening. And this could be Meta’s second-biggest one-day wipeout on record.
What went wrong? Well, nothing really. Except that Zuckerberg decided to tell investors he plans to spend even more money on AI. A lot more money.
The Spending Spree
Here’s where things get interesting. Meta raised its 2025 capital expenditure guidance to between $70-72 billion. But that wasn’t the scary part. The company also casually mentioned that spending increases in 2026 would be “notably larger” than the roughly $30 billion it expects to add this year.
Let that sink in. We’re talking about spending that could push Meta’s annual capex well past $100 billion. That’s more than the GDP of countries like Morocco or Ecuador.
And Meta isn’t alone in this spending frenzy. Microsoft reported quarterly capex of $34.9 billion — a whopping 74% increase year-on-year, and $5 billion more than its own forecast. For the full year, Microsoft’s data center spending hit $120 billion, up from $88 billion the previous year.
Alphabet (Google’s parent company) wasn’t far behind, with third-quarter capex reaching $24 billion, nearly double last year’s $13.1 billion. The company disclosed an eye-watering $155 billion order backlog for AI infrastructure.
Combined, these companies and Amazon are forecast to spend $350 billion this year and $410 billion in 2026 on AI infrastructure. Some analysts think it could touch $500 billion by 2027.
What Are They Building?
So where is all this money going? Data centers. Lots and lots of data centers.
Microsoft’s CEO Satya Nadella announced plans to double the company’s data center footprint over the next two years. He’s building what he calls “planet-scale” cloud infrastructure. One site in Wisconsin — claimed to be the world’s most powerful data center at 2 gigawatts — is set to come online next year.
These aren’t your typical server farms. These are massive facilities packed with specialized AI chips, requiring enormous amounts of electricity and cooling systems. They’re the backbone needed to train and run increasingly powerful AI models that these companies are racing to develop.
The AI Arms Race
Why the urgency? Because nobody wants to be left behind in what might be the most important technology race of our generation.
Meta’s Zuckerberg is chasing what he calls “personal superintelligence” — AI that’s smarter than humans. He’s reportedly offering top engineers $100 million deals to join a new “superintelligence” lab inside the company. His argument? It’s better to “aggressively frontload building capacity” so Meta is prepared for the “most optimistic cases” where superintelligence arrives sooner.
Google’s Sundar Pichai is touting the company’s “full stack” AI strategy — everything from making their own AI chips (called TPUs) to training their Gemini language models. He claims Gemini is already generating “billions in quarterly revenues” from business customers, while the consumer app has hit 650 million monthly users, chasing ChatGPT’s 800 million.
Microsoft, as OpenAI’s largest backer with a 27% equity stake worth $135 billion, is racing to supply computing power not just for its own AI ambitions but also for its star partner.
The Winners and Losers
But here’s where Wednesday’s earnings got really interesting. While Meta got hammered, Alphabet’s stock surged over 6% in after-hours trading.
Why? Alphabet delivered the goods. The company reported quarterly revenue exceeding $100 billion for the first time ever. Net income jumped 33% to $35 billion. Its core search and advertising business grew 15% (analysts expected 11%), while cloud revenues climbed 34% to $15.2 billion.
Microsoft fell somewhere in between, with shares dipping 3.6%. The company beat revenue and profit expectations, but growth in its Azure cloud computing business — at 39% — disappointed some bullish investors who expected even more, especially after the company’s expanded deal with OpenAI.
The Investor Dilemma
This brings us to Wall Street’s big question: When does the AI spending pay off?
Investors are essentially being asked to trust that these companies know what they’re doing. That the hundreds of billions being poured into AI infrastructure will eventually generate returns that justify the investment.
Meta’s situation is particularly tricky because the company admitted earlier this year that its generative AI push wasn’t expected to generate meaningful revenue in 2025 or even 2026. Yet here’s Zuckerberg, asking for patience and bigger budgets.
His pitch? In the best case, Meta gets superintelligence faster. In the worst case, “we would just slow building new infrastructure for some period while we grow into what we build.” Essentially: trust me, this will work out.
The Ripple Effects
Interestingly, the AI boom isn’t just benefiting tech companies. Caterpillar — yes, the heavy machinery company — saw its stock jump 11.6% after reporting that demand for power generation equipment for data centers was driving record order backlogs. Its $39.8 billion backlog was propelled by orders for reciprocating engines specifically for data center applications.
When construction equipment makers are getting a boost from AI spending, you know we’re talking about serious infrastructure investment.
The Bottom Line
What we’re witnessing is possibly the most expensive technological bet in corporate history. These companies are spending GDP-level sums on infrastructure for technology that’s still evolving, with business models that are still being figured out.
Alphabet seems to have cracked the code — showing investors both massive spending and meaningful revenue from AI. Meta is asking for faith in a longer-term vision. Microsoft is caught in the middle, spending heavily but with results that haven’t quite matched the most optimistic expectations.
For now, Wall Street’s message is clear: spending is fine, but show us the money. And if you can’t show us the money yet, at least show us you’re not going to keep spending forever without a clear path to returns.
The AI race is on. The question is: who’s going to win, and at what cost?
