AI buildout strains capacity, reshapes chip demand

Alphabet CEO Sundar Pichai told investors that surging artificial intelligence demand has turned infrastructure capacity into the company’s biggest near‑term challenge. On Alphabet’s fourth‑quarter earnings call, he cited constraints in power availability, land for data centers and supply chains as key bottlenecks in scaling infrastructure to meet what he called “extraordinary” demand.

Pichai said Alphabet has spent years developing custom chips such as tensor processing units, but the current AI wave is testing how fast those investments can be expanded. He described management’s top priority as ramping capacity while maintaining efficiency, and stressed the need to “get our investments right for the long term” without losing operational discipline.

Alphabet reported fourth‑quarter revenue of $113.83 billion, up 18% year over year and ahead of analyst estimates. The company said all major business segments grew at double‑digit rates, highlighting how AI is already feeding into its broader revenue base even as infrastructure constraints mount.

Start‑ups such as Anthropic are contributing to that demand. Industry blog Semianalysis recently described Anthropic’s Claude Code tool as an inflection point in so‑called agentic AI, and forecast it could drive more than 20% of daily GitHub code commits by year‑end, up from about 4% now. The blog compared Claude Code’s potential impact to the early launch of ChatGPT.

Claude Code allows developers to generate software through natural‑language prompts, a shift some in the sector refer to as “vibecoding.” Analysts cited in the report expect this to accelerate the development and deployment of AI agents that can execute multi‑step tasks independently, expanding AI usage beyond model training and simple inference.

Arm Holdings is among the hardware vendors positioned for this shift. Arm’s CPU designs are widely licensed, and its chief financial officer, Jason Child, argued that agentic AI will be CPU‑intensive because CPUs handle orchestration and communication between agents. He said Arm expects CPU demand to rise as these workloads proliferate and noted that the company’s data center royalty revenue more than doubled year over year, with management now projecting data center to surpass mobile as its largest category in coming years.

AI training and inference also require large amounts of high‑performance memory and advanced manufacturing tools. Samsung, Micron Technology and ASML have all seen sharp share‑price gains over the past year as investors focus on their roles in the AI infrastructure stack.

Samsung’s market value has reached about $772.8 billion after a 217% gain over 12 months, helped by a near tripling of operating profit in its fourth quarter of 2025 as demand for DRAM used in AI systems increased. Micron, valued at roughly $469.5 billion, reported 57% year‑over‑year revenue growth and a 180% increase in net income in its fiscal first quarter of 2026, driven by similar trends in memory pricing and volumes.

ASML, which supplies the extreme ultraviolet lithography equipment needed to produce the most advanced chips, recorded 20.6% revenue growth and a 32.4% rise in net income in 2025. Orders for its lithography systems rose 48% year over year to 28 million units of capacity, underscoring how foundries are expanding to meet AI‑related chip demand.

Other hardware and manufacturing specialists are also aligning more closely with AI and data center customers. Flex reported strong third‑quarter results and highlighted new partnerships with Nvidia and LG, describing expanded investments in advanced technologies that support AI and cloud infrastructure. The company said these relationships move it closer to key parts of the hardware stack that large data centers and industrial automation clients depend on.

Equipment and systems vendors in adjacent markets are seeing knock‑on effects as well. Caterpillar’s power and energy segment reported record sales tied to large AI data center projects and a record backlog. Investors have linked Caterpillar’s recent share‑price performance to expectations that its power systems will continue to benefit from data center build‑outs.

Together, these developments show how AI demand is reshaping capital spending across semiconductors, manufacturing services and heavy equipment. While companies are reporting strong near‑term growth, several executives, including Alphabet’s Pichai, have emphasized that constraints in power, land and supply chains will remain central challenges as the sector scales.