Global AI server shipments are expected to grow 31 percent this year, with 2026 growth projected between 20 and 28 percent, depending on the source, driven by capital spending from major cloud providers worldwide.
North American cloud providers, including Google Cloud and Amazon AWS, are projected to significantly increase capital expenditures this year. The nine largest North American providers combined are forecast to reach $1.32 trillion in capital expenditures by 2027—a 49 percent increase from the 2026 figure.
China's four largest cloud providers—ByteDance, Tencent, Alibaba and Baidu—are also accelerating spending, with combined capital expenditures expected to rise more than 80 percent in 2026.
On the hardware side, ASIC-based AI servers are expected to outpace GPU-based systems by 2026, reflecting a shift toward specialized chips for AI workloads over general-purpose GPUs. AI servers are projected to account for 17 percent of overall server shipments by 2026, up from current levels.
NVIDIA's Blackwell and Rubin platforms are expected to lead the GPU market, alongside growing deployment of rack-scale GB/VR systems. Supply chain constraints—particularly in high-bandwidth memory and CoWoS advanced packaging—remain a limiting factor on how quickly new capacity can come online.
Major cloud providers posted sharp gains. Meta rose 6.0 percent to $590.24, Amazon gained 4.6 percent to $284.02 and Alphabet climbed 4.9 percent to $373.51. Microsoft added 4.9 percent to $487.65 and NVIDIA rose 2.9 percent to $206.64.
The S&P 500 closed up 1.5 percent at 7,601 and the Nasdaq gained 2.1 percent to 25,914.
The capital race reflects a broader strategic bet: hyperscalers are building compute capacity now to monetize AI through cloud services, internal products and specialized AI features layered onto existing platforms.
