Meta and Alphabet: Betting $335 Billion on AI in a Single Year
In April 2026, both Meta and Alphabet reported the best quarters of their recent history and got punished for it.
Meta grew revenue 33% year over year — the fastest in nearly five years — and the stock fell 8.6%. Alphabet grew revenue 22%, doubled EPS, and Google Cloud grew 63%. Stock fell too.
The reason wasn't performance. It was the number both companies quietly slipped into their conference calls: combined 2026 AI infrastructure spending of $335 billion.
That's more than France's annual defense budget. More than the entire market cap of Netflix. More than the GDP of Finland. Two companies. One year.
The market didn't love it. Neither did I, at first. But then I did the math on what this means for the next four years — and I've been DCAing into both ever since. This is why.
Meet the two spenders
Meta (META) owns Facebook, Instagram, WhatsApp, and Threads. Total daily active users across the family: north of 3.4 billion. Zuck's bet since 2022 has been that AI dramatically improves ad targeting, video ranking, and content recommendation — the three levers on which Meta's entire economic engine turns. That thesis is why he's committed to spending as much money on AI in a single year as many countries spend on their armies.
Alphabet (GOOG) owns Google Search, YouTube, Android, Google Cloud, Waymo, and roughly a hundred other things. Gemini is their AI. Google Cloud is their infrastructure. Both are AI's largest known payer and Nvidia's largest known customer, respectively. In 2026 Alphabet is spending $190 billion on capex — up from $110 billion two years earlier.
Two of the four largest AI-capex spenders on Earth, in the same portfolio.
Meta — spending more than it earns, on purpose
Meta's Q1 2026 print: revenue $56.3 billion, up 33%; operating income $22.9 billion, up 30%; profits $26.8 billion, up 61%. These are numbers that would have been considered career-defining a decade ago. The stock fell 8.6% anyway.
The reason: Zuck raised 2026 capex guidance from a prior range of $115-135 billion to a new range of $125-145 billion. $145 billion of capex means spending roughly two-thirds of Meta's revenue on physical infrastructure. Servers. Data centers. Land. Power. Networking.
The market's reaction was rational. Meta's free cash flow margin drops meaningfully in a year like this. Return on capital is compressed until the AI investments produce revenue. And there's genuine uncertainty about whether the return ever comes.
What the market missed: Llama 4 is now powering Meta's ad ranking system. That's the single most valuable use case Meta has ever built — targeting is what the entire ad business is priced on. If AI improves ad targeting even 10% relative to the pre-AI baseline, Meta's revenue per user compounds at a materially higher rate for years.
Then in July, Meta announced it will start selling excess AI compute capacity to third parties — turning a portion of that $145B capex into a revenue-generating business, not just a cost center. Stock ripped.
Meta trades around $660 in mid-July. Not cheap. Not obviously overvalued either. Priced for "the capex was worth it," which is exactly what the debate is about.
Alphabet — the sleeping giant everyone forgot about
Alphabet's Q1 2026 print is arguably the most impressive from a big-tech company this decade.
Total revenue: $109.9 billion, up 22% — the fastest in over two years. EPS $5.11 versus $2.62 consensus (nearly double). Google Cloud revenue $20.03 billion, up 63%. Cloud operating margin 32.9% — a shocking number for a business that lost money three years ago. Cloud backlog: $462 billion — nearly doubled in one quarter.
Gemini is quietly becoming the enterprise default in a way most people underestimate. Gemini Enterprise MAUs grew 40% quarter over quarter; total paid subscriptions across Alphabet products hit 350 million; Gemini is processing 16 billion tokens per minute via direct API. That's more inference traffic than OpenAI publicly reports.
Then Sundar dropped the capex number: $190 billion in 2026, with a "significant increase" to come in 2027. Free cash flow margin dropped from 21% to 9.2%. The market sold Alphabet down to $345 after-hours despite the beats.
By mid-July it recovered somewhat. Trading around $370.
The bull case: Google Cloud is now real money — $20B/quarter growing 63%, with 33% margins. Gemini has real usage. Waymo has real ride volume. YouTube ad revenue keeps hitting records. Every one of those is compounding into the capex payoff.
The bear case: the $190B capex works out to roughly $50/quarter of revenue growth needed just to break even on the infrastructure. If AI doesn't generate that, Alphabet becomes a slower-growth story with lower margins for a couple of years while the market waits.
⭐ What $10, $25, or $50 a week becomes by December 2030
Ground rules. Start next Friday. Split whatever you commit 50/50 between META and GOOG. Buy fractional shares every Friday. Stop December 26, 2030. 232 Fridays.
The three scenarios.
- Bear — AI capex disappoints, ad revenue softens, cloud growth stalls. META drifts to $550. GOOG to $320.
$10 a week — $4,640 total in over four and a half years.
| Scenario | End portfolio value | Return |
|---|---|---|
| Bear | ~$4,260 | −8% |
| Base | ~$5,660 | +22% |
| Bull | ~$6,510 | +40% |
In layman terms.
What could break this
AI capex doesn't produce the return. Both companies are betting massive amounts of shareholder capital on infrastructure that only works if AI drives real revenue growth over 2-4 years. If AI adoption stalls, both are stuck with expensive buildings and depreciating servers.
Regulation cracks the ad model. Meta is in a decade-long fight with EU regulators. Alphabet faces antitrust judgment risk on Search. Either could force material changes to the ad businesses that fund AI.
China becomes dominant in AI. If DeepSeek, Qwen, or a Chinese model becomes the global default, Western AI capex isn't just wasted — it's inverted. Low probability, but real.
I stop DCA and reassess if any of those actually happen.
The reflection
Big tech in 2026 is doing something the market has never rewarded well in real time: spending enormous amounts of cash on capex to protect its position over the next decade.
Every time a big-tech company has done this before — Amazon building AWS, Google building GCP, Facebook building the mobile app — the market punished them for the spending and rewarded them enormously three to five years later. The lesson isn't that this pattern always repeats. The lesson is that when the market is punishing companies for investing behind their strongest position, that's usually a good time to be buying.
META and GOOG are the two most obvious versions of this thesis today. Both are being punished for spending. Both are spending because they think AI compounds their existing moats. Both have decades of evidence of being right about this pattern.
Skip one delivery. Buy the two trillionaires. See you in 2030.
---
This is not financial advice. I own both META and GOOG. I'm sharing my personal research and strategy. Past performance and price projections are not guarantees. Always do your own due diligence before investing.