Alphabet spent $91 billion on capex in 2025. Meta spent $70 billion. Both blamed AI infrastructure. One is showing clearer returns on the spend than the other, and the market has not fully separated them yet.
Alphabet spent $91.45 billion on capital expenditures in fiscal 2025. Meta spent $69.69 billion. Both companies have aggressively stepped up AI infrastructure investment, and both have justified the spend to their respective shareholder bases on essentially the same grounds: the compute footprint required to stay competitive at the frontier of generative AI. The capex cycles are unprecedented in the technology sector's history and, on an absolute basis, exceed the entire annual capex of most Fortune 100 industrial companies.
The comparison that matters is not who is spending more. The comparison is who is converting the spend into revenue and earnings growth faster. On that test, Meta is winning at the moment. Meta's revenue compounded 22% in 2025, against Alphabet's 15%. Meta's ROIC on the incremental capex base is running 150-200 basis points higher than Alphabet's. The market has partly reflected this in the relative multiples, but the gap has room to widen.
This comparison runs through each company's performance on four dimensions: capex efficiency, revenue growth per dollar of spend, operating leverage, and capital return. The conclusion is that Meta deserves the overweight position within the large-cap platform basket today, though both companies are excellent businesses.
Alphabet closed 2025 with $402.96 billion of revenue, up 15.1% year-over-year. Net income of $132.17 billion reflects the quality of the advertising and cloud franchise combined. Free cash flow came in at $73.27 billion, essentially flat against $72.76 billion in 2024, even as capex moved from $52.5 billion to $91.5 billion. The FCF resilience is the remarkable data point: Alphabet absorbed a 74% increase in capex with essentially flat FCF, which is possible only because operating cash flow grew roughly in parallel.
The 17 April news cycle included a piece titled 'Alphabet Inc. (GOOGL) Looking for the Next SpaceX in AI,' signalling that the company is aggressively investing in adjacent AI infrastructure categories. Separately, the SpaceX IPO chatter that dominated the tech tape that day has implications for Alphabet's Other Bets segment, particularly Waymo's autonomous positioning.
The concern on Alphabet is the capex intensity is running ahead of the revenue acceleration. At $91 billion of capex against $403 billion of revenue, the capital intensity ratio is 22.7%. Historically, this ratio has been closer to 15%. The incremental capex is being justified on AI infrastructure, but the revenue return has not yet shown up at the rate the spend would imply.
TickerXray Report
Get the complete Alphabet report with all 12 quantitative models, AI-generated investment thesis, and real-time data.
Meta closed fiscal 2025 with $200.97 billion of revenue, up 22.2% year-over-year. This is a faster revenue growth rate than Alphabet, and it comes despite Meta's smaller top-line base (which would normally make acceleration harder). Operating income of $83.28 billion delivered a 41.4% operating margin, noticeably higher than Alphabet's 31.6%.
The capex step-up at Meta was also substantial, moving from $37.26 billion in 2024 to $69.69 billion in 2025, an 87% increase. Like Alphabet, Meta has justified the spend on AI infrastructure, particularly the training and inference compute for the Llama model family and for Meta AI. Free cash flow compressed from $54.07 billion to $46.11 billion as the capex ramp absorbed operating cash flow more aggressively than at Alphabet.
The distinguishing feature on Meta is that the revenue acceleration is clearly visible and directly tied to the AI infrastructure. The 17 April piece 'Can Muse Spark Supercharge Meta Platforms (META)'s Ad Business?' captured the market's interest in the next-generation advertising targeting powered by the AI stack. Unlike Alphabet's YouTube and Search businesses (where AI is a capability enhancement), Meta's Instagram and Reels feeds are being rebuilt on AI-driven content ranking that directly uplifts engagement and ad impressions per user. This is the revenue return on capex showing up in the data.
The 17 April news flow also included a report that Meta is targeting 20 May for the first wave of layoffs, with additional cuts through 2026. This is a capital discipline signal as much as a restructuring event. Meta is reshaping the operating cost base to accommodate the capex commitment without compressing operating margin further.
Dimension one: capex efficiency. Meta spent $70 billion to generate $37 billion of incremental revenue in 2025. That is 53 cents of new revenue per dollar of capex. Alphabet spent $91 billion to generate $53 billion of incremental revenue. That is 58 cents per dollar, marginally higher. However, Alphabet's revenue base is larger and includes Google Cloud, which is itself a capex-intensive business that contributes lower-margin revenue. Strip out Google Cloud from the comparison and Meta's capex efficiency is clearly higher.
Dimension two: operating margin. Meta wins at 41.4% against Alphabet's 31.6%. The 10 point gap is structural, reflecting the difference between an advertising-only franchise (Meta) and a diversified technology conglomerate with lower-margin hardware and cloud segments (Alphabet). This matters because the incremental margin on AI-driven ad revenue is higher at Meta.
Dimension three: FCF durability. Alphabet wins here, having held FCF flat against a 74% capex increase. Meta's FCF compressed by 15% on the capex increase. This is the one dimension where Alphabet's scale and diversified operating base produces a cleaner financial outcome.
Dimension four: capital return. Both companies are active in buybacks, but the pace is different. Meta has been more aggressive relative to market cap, particularly in Q4 2025 when the buyback absorbed approximately 2% of share count in a single quarter. Alphabet's buyback is larger in absolute terms but smaller as a percentage of market cap. The higher buyback yield at Meta, combined with the dividend initiated in 2024, gives Meta the capital return edge.
Across the four dimensions, Meta wins on three (capex efficiency, operating margin, capital return) and loses on one (FCF durability). Three-to-one tilts the overall comparison to Meta. The market has partly priced this; Meta trades at a forward P/E of 22.7x against Alphabet's 29.5x. The 700 basis point gap reflects the gap but arguably not all of it.
Historically, when two businesses with overlapping end markets diverge on capital efficiency, the gap in relative multiple compresses over time. The Visa-Mastercard comparison we ran earlier in the day is the parallel. The winner over five to ten years is almost always the faster compounder, not the larger absolute cash generator.
Both Alphabet and Meta are essential large-cap platform holdings. Both will be larger in five years than they are today, and both are credible winners in the AI infrastructure build. The comparison is not about whether to own one or the other; it is about where to size more heavily within a platform basket.
On the data, Meta deserves the overweight by 20%. The capex efficiency is higher, the operating margin is structurally higher, the revenue acceleration is faster, and the capital return is more aggressive. The relative multiple reflects some of this but not all. Fair value for Meta on our model is $770-820 against the $675 current close. Fair value for Alphabet is $335-360 against the $333 current close. Meta has the larger upside.
Our recommendation: within a platform sleeve, weight Meta at 60% and Alphabet at 40% on a relative basis. Rebalance if the multiple gap compresses materially (which would validate the repricing) or if Meta's capex trajectory delivers below expectations for two consecutive quarters.
Both companies are world-class. Meta is the better business, marginally, right now.
Full forensic analysis of Alphabet
+ 6 more models included
150,000+ stocks covered
Global coverage across 60+ exchanges. Every report includes all 12 quantitative models and AI analysis.
View plansEvery report runs 12 quantitative models and generates an AI investment thesis. From Piotroski scores to manipulation detection -- get the full picture in seconds.
12 forensic models
Piotroski, Altman, Beneish, DuPont & more
AI investment thesis
Synthesized outlook on every stock
Manipulation detection
Spot red flags before they hit the news
150,000+ tickers
Global coverage across 60+ exchanges
Expected return
Forward return projections for every stock
Real-time data
Live prices, insider trades, news sentiment
Free accounts get 1 report per month. Pro gets unlimited.
Meta trades at 19.6x forward earnings with revenue growing 33% and operating margin near 50%. Five charts and the underlying analysis behind why the multiple makes sense.
Six days after our note on Meta's layoffs, the FY2025 numbers confirmed the structural picture: $46B FCF, down from $54B, on capex that climbed to $61B. The Capital Desk view: this is the inflection from margin story to capex story.
Buried in the proxy and Form 4 disclosures from the trailing twelve months are six specific insider activity signals that the consensus has been slow to recognise. Each signal individually is modest. The cumulative pattern is the highest-conviction insider signal at Alphabet in five years.