Alphabet, Microsoft, Meta target $700B

AI infrastructure is turning into the biggest spending race in tech. Not product launches. Not hiring. Concrete, chips, power, networking gear, and giant data centers. The news from April 29 was that Alphabet, Microsoft, Meta, and Amazon all made clear this buildout is still accelerating, not cooling, and the combined 2026 number now looks close to $700 billion once you stack their plans together. (cnbc.com) ### What actually changed? Alphabet was the clearest. It raised full-year 2026 capex guidance to $180 billion to $190 billion, up from $175 billion to $185 billion, and said 2027 spending should “significantly increase” again. That is not a company trying to taper off. It is a company telling you the bottleneck is still supply. Google Cloud crossed $20 billio(cnbc.com)higher without compute constraints. (cnbc.com) ### Why is Microsoft part of this? Microsoft did not give a neat full-year capex target in the same way, but the direction was obvious. It said capex will rise above $40 billion in the June quarter alone as it keeps adding AI capacity. Microsoft Cloud revenue hit $54.5 billion, up 29%, and its AI business passed a $37 billion annual revenue run rate, up 123% (cnbc.com)ll outrunning available supply. (microsoft.com) ### What about Meta? Meta is the outlier because it is not mainly building rented cloud capacity for outside customers. It is building for ads, recommendations, video, and its own AI ambitions. But the spending is just as real. Meta posted $19.84 billion in capital expenditures in Q1 alone and said it raised full-year capex guidance because(microsoft.com)s proving this is not just a cloud-company story — AI capex is now core platform strategy. (investor.atmeta.com) ### And Amazon? Amazon’s Q1 release did not spell out a fresh annual capex guide in the press release, but it showed the same pattern. AWS revenue rose 28% to $37.6 billion, operating cash flow hit $148.5 billion over the trailing twelve months, and outside earnings coverage tied Amazon to roughl(investor.atmeta.com)ild mode. (ir.aboutamazon.com) ### So where does the $700 billion come from? Add the visible pieces and the estimate starts to make sense. Alphabet is at up to $190 billion. Amazon is widely framed around roughly $200 billion. Meta’s run rate and raised guidance imply a very large year, and Microsoft’s quarterly pace points to (ir.aboutamazon.com)less important than the direction — all four are still pressing the gas. (cnbc.com) ### Why are investors not panicking more? Because the revenue is showing up fast enough to keep the story intact. Google Cloud just had a breakout quarter. Microsoft’s AI revenue is scaling unusually quickly. AWS reaccelerated. Meta’s ad machine is still throwing off cash. The market is basically tolerating huge capex because these companies can still point to demand, backlog, or monetization — not just vague future promise. (abc.xyz) ### What is the real risk? Efficiency. That is the whole argument now. GPUs, custom chips, memory, storage, networking, land, and power all cost more when everyone is buying at once. If AI revenue keeps climbing, the spend looks smart. If usage grows but pricing compresses, or if returns arrive slower than expected, these budge(abc.xyz)also means the industry is still paying scarcity prices. (cnbc.com) ### Bottom line? The hyperscalers are no longer debating whether AI needs giant infrastructure budgets. They have already decided. The new question is whether they can turn record capex into durable profit before the cost of the race catches up.