CNBC: AI boom lifts toilet, steel prices

CNBC said in a May 23 video that the AI buildout is pushing up demand not only for GPUs and servers, but also for transformers, copper, steel, cooling systems and other industrial inputs tied to new data centers. The point of the report was that the spending wave behind generative AI is now reaching markets far upstream from semiconductors, including construction materials and basic building products. That argument lines up with recent analyst estimates showing hyperscalers raising capital spending plans again and with industry research on power, metals and campus construction needs. ### How does an AI boom end up affecting toilets and building materials? Hyperscale data centers are physical construction projects before they become software platforms. CNBC’s framing was that each new AI campus requires large amounts of electrical equipment, piping, cooling, concrete, steel and copper, which can tighten supply for other buyers using the same industrial base. That is how a surge in server demand can spill into products used in housing, commercial construction and renovation. (cnbc.com) McKinsey said in an August 2025 report that global capital expenditures on data-center infrastructure, excluding IT hardware, are expected to exceed $1.7 trillion through 2030, largely because of AI expansion, edge computing and high-performance computing. McKinsey also said campuses are moving from tens of megawatts toward hundreds of megawatts and, in some cases, gigawatt scale. (youtube.com) ### How big is the spending wave behind this? CNBC reported on April 30 that Wall Street analysts now estimate total AI capital expenditures could rise to between $800 billion and $900 billion in 2026 and exceed $1 trillion in 2027. Jefferies analysts said in a note cited by CNBC that “demand outpaces supply and pricing increases,” while Bank of America’s tally showed higher spending plans from Alphabet, Amazon, Meta and Microsoft. (mckinsey.com) Amazon projected about $200 billion in buildout for 2026, according to CNBC, while Bank of America’s tally cited by the network put Alphabet at $185 billion, Meta at $135 billion and Microsoft at $190 billion. Those figures help explain why suppliers of power gear, metals and construction services are now part of the AI story. ### Why are copper and power equipment getting so much attention? (cnbc.com) S&P Global said in its January 8 report “Copper in the Age of AI” that AI and data centers are becoming a new source of copper demand because expanded electricity supply is essential to digital infrastructure. The report framed copper as foundational to electrification, digitalization and technologies including data centers, electric vehicles and defense. (cnbc.com) McKinsey said data-center electricity demand could reach 1,400 terawatt-hours globally by 2030, or 4% of total global power demand. In the United States, McKinsey said data-center electricity use could rise from 147 terawatt-hours in 2023 to 606 terawatt-hours by 2030, equal to 11.7% of total U.S. power demand. (spglobal.com) ### Are consumers already seeing spillover effects? Consumer Reports said on March 20 that AI data centers are already contributing to higher household electricity bills in some regions, while also competing for water, land and other local resources. The publication cited 3,069 operating U.S. data centers and 1,489 additional sites planned or under construction as of March 2026, with a growing share of them classified as hyperscale. (mckinsey.com) CBS News reported on April 30 that a memory-chip shortage tied to AI demand is pushing up computer prices, citing Oxford Economics. Bernard Yaros, lead economist at Oxford, told CBS that prices for computers, software and accessories had risen more than 3% per month recently, reversing a decades-long decline in computer costs. ### What should readers watch next? (consumerreports.org) The next evidence will come from company capital-spending updates, utility interconnection queues and commodity-market research. Hyperscaler earnings calls from Alphabet, Amazon, Meta and Microsoft have become one of the clearest places to track how much new AI infrastructure is still being ordered, while S&P Global and similar industry reports are mapping how that demand flows into copper, power and construction supply chains. (cbsnews.com) (cnbc.com)