The AI spending arms race just kicked into overdrive. Google, Meta, and Microsoft delivered a collective message to Wall Street this earnings season: we're doubling down on artificial intelligence infrastructure, and the bill is getting bigger. Combined, these tech giants are now projecting over $200 billion in capital expenditures through 2026, with Google alone raising its 2025 guidance to as much as $93 billion.
The numbers tell the story of an industry that's not just betting on AI anymore - it's going all-in. Alphabet didn't just beat earnings expectations when it reported its first-ever quarter topping $100 billion in revenue. The company dramatically revised its capital expenditure forecast upward, now projecting spending between $91 billion and $93 billion for 2025, a massive jump from its earlier guidance of $75 billion to $85 billion according to company filings.
But that's just the opening act. Finance chief Anat Ashkenazi told investors to expect 'a significant increase' in capex for 2026, signaling Google's AI infrastructure buildout is far from complete. The revelation sent a clear message: the search giant sees no ceiling on AI demand.
Meta matched that intensity, hiking the low end of its capex guidance to $70 billion from $66 billion. CEO Mark Zuckerberg defended the massive outlay during the earnings call, telling analysts that 'being able to make a significantly larger investment here is very likely to be a profitable thing.' The comment reflects Meta's confidence in monetizing its AI capabilities across its social media empire.
Microsoft completed the spending triumvirate with its own surprise. CFO Amy Hood revealed that first-quarter capex reached $34.9 billion, well above the $30 billion figure the company had estimated just three months ago in July. More significantly, Hood indicated that capex growth in fiscal 2026 will exceed even 2025's pace, suggesting Microsoft's AI infrastructure demands are accelerating.
The synchronized spending surge comes as these companies race to build the computational backbone for next-generation AI services. Data centers, specialized chips, and networking infrastructure don't come cheap, but the alternative - falling behind in the AI race - appears far costlier to these executives.
