The Big Tech companies of Silicon Valley, including Alphabet (Google), Microsoft, Meta, and Amazon released their Q1 earnings on Wednesday. And as expected, they all revealed details of their extraordinary spending on AI infrastructure in 2026. These companies are preparing to spend at a scale that goes far beyond anything the technology industry has seen before. Google’s parent company alone has committed to spending as much as $190 billion on AI this year alone.
To put that number in context, India’s Finance Minister Nirmala Sitharaman allocated an all-time high of Rs 7.85 lakh crore (roughly $87 billion) to defence in the Union Budget for 2026–27. Google’s AI spending commitment is more than double of that figure.
In fact, it is larger than the entire GDP of countries such as Hungary or Ecuador. And, according to Google’s own Chief Financial Officer, in 2027 it will spend more.
Where will Google spend this much of the money?
In the recent quarterly earnings call, CFO Anat Ashkenazi said the company is seeing “unprecedented internal and external demand for AI compute resources”. Roughly 60 per cent of the budget, she revealed, will be spent on servers, while 40 per cent will be allocated to data centres and networking equipment.
To be more specific, a significant portion of this spending is being funnelled into core infrastructure, particularly servers, chips, and data centres.
CEO Sundar Pichai noted that Gemini, Google’s flagship AI model, is now processing more than 16 billion tokens per minute via direct API. Paid monthly active users of Gemini Enterprise grew 40 per cent last quarter, with major enterprise deals signed with companies including Bosch, Mars, and Merck.
And with AI adoption growing at this pace, the underlying infrastructure demands grow alongside it. These models are constantly retrained, fine-tuned, and deployed across billions of users simultaneously. Every search query, every AI-generated response, every automated workflow runs on infrastructure that must be fast, reliable, and always available. Google will be putting the money into maintaining and scaling the models.
Meanwhile, Google has been rapidly expanding its global network of AI-optimised data centres, many of which are powered by its custom-built Tensor Processing Units (TPUs). These chips are not merely incremental upgrades, they are specifically designed to handle the scale and complexity of modern AI workloads, and are increasingly positioned as alternatives to Nvidia’s GPUs. Google is now planning to deepen its investment in building these chips further.
At the same time, the company’s cloud backlog has nearly doubled to $460 billion. In other words, Google’s massive budget is not speculative spending. The demand is already locked in, and Google is planning to meet it.
Energy is another increasingly unavoidable factor. AI data centres are extraordinarily power-hungry, and Google’s expansion plans come with parallel investments in renewable energy, advanced cooling technologies, and long-term carbon-free energy commitments.
Then there is the competitive pressure
Google is not alone with a massive budget to build AI. Every major player is spending aggressively to stay ahead.
The four hyperscalers — Alphabet Inc, Amazon, Meta, and Microsoft — are together projected to spend as much as $725 billion in capital expenditure in 2026, a 77 per cent jump from last year’s already record $410 billion.
Microsoft alone is expected to invest around $190 billion this year, with roughly two-thirds of that directed towards GPUs, CPUs, and AI tools. Meta, meanwhile, has raised its capital expenditure forecast to as much as $145 billion as it doubles down on its own AI ambitions. Hence, Google is spending heavily not only to meet demand, but also to ensure it does not fall behind.
– Ends
Source link
[ad_3]
