Podcast Episode
Meta increased its 2026 capital expenditure forecast to between $125 billion and $145 billion, up from a prior range of $115 billion to $135 billion. Alphabet now projects full-year spending of $180 billion to $190 billion and signalled that 2027 spending would rise significantly. Microsoft told investors it expects $190 billion in calendar-year 2026 capital expenditures, while Amazon reiterated plans to spend close to $200 billion. A substantial share of that spending flows directly to Nvidia, whose GPUs remain the backbone of AI training and inference infrastructure worldwide.
AWS announced plans to deploy more than one million Nvidia GPUs across its cloud regions, spanning both Blackwell and the next-generation Rubin architectures. Google Cloud said it would be amongst the first providers to offer Nvidia Vera Rubin NVL72 rack-scale systems in the second half of 2026.
Big Tech's $725 Billion AI Spending Surge Sends Nvidia to New Heights
April 30, 2026
0:00
2:47
The four largest cloud providers have unveiled combined 2026 capital expenditure plans approaching $725 billion, much of it flowing to Nvidia's Blackwell Ultra and forthcoming Vera Rubin GPU architectures. Alphabet, Amazon, Meta, and Microsoft each raised or reaffirmed aggressive AI infrastructure spending targets, while Nvidia CEO Jensen Huang projects orders could approach $1 trillion by 2027.
Hyperscalers Open the Floodgates
The four largest cloud providers have collectively committed to spending approximately $725 billion on AI infrastructure in 2026, an unprecedented wave of capital expenditure that places Nvidia at the centre of the artificial intelligence economy. Alphabet, Amazon, Meta, and Microsoft all disclosed their quarterly results after the market close on 29 April, with each company either raising or reaffirming aggressive infrastructure spending targets.Meta increased its 2026 capital expenditure forecast to between $125 billion and $145 billion, up from a prior range of $115 billion to $135 billion. Alphabet now projects full-year spending of $180 billion to $190 billion and signalled that 2027 spending would rise significantly. Microsoft told investors it expects $190 billion in calendar-year 2026 capital expenditures, while Amazon reiterated plans to spend close to $200 billion. A substantial share of that spending flows directly to Nvidia, whose GPUs remain the backbone of AI training and inference infrastructure worldwide.
Blackwell Ultra Ramps Up
Nvidia's Blackwell Ultra platform, first unveiled at GTC 2025 and further detailed at GTC 2026, has been shipping since the second half of 2025 and continues to ramp up production. The GB300 NVL72 rack-scale system, featuring 72 Blackwell Ultra GPUs and 36 Grace CPUs, delivers up to 50 times higher throughput per megawatt and 35 times lower cost per token for agentic AI workloads compared with the previous Hopper generation. For training, Blackwell Ultra offers roughly 1.9 times the performance of standard Blackwell GPUs at scale.AWS announced plans to deploy more than one million Nvidia GPUs across its cloud regions, spanning both Blackwell and the next-generation Rubin architectures. Google Cloud said it would be amongst the first providers to offer Nvidia Vera Rubin NVL72 rack-scale systems in the second half of 2026.
Vera Rubin and Beyond
The Vera Rubin platform, Nvidia's next full architectural leap beyond Blackwell, is expected to begin shipping in the second half of this year and promises 10 times the performance per watt of its predecessor. Jensen Huang also previewed Kyber, a future architecture featuring 144 GPUs in vertical compute trays that will underpin Vera Rubin Ultra systems targeted for 2027. At GTC 2026, Huang projected that orders for Blackwell and Vera Rubin systems could approach $1 trillion by 2027, doubling the company's earlier $500 billion demand estimate. Nvidia's stock reached an all-time high earlier this week.Published April 30, 2026 at 9:23am
