The Data-Center Boom Is Sparking a Third Wave of Inflation Demand for memory chips is pushing prices higher. Will AI’s promise of increased productivity come in time to temper that inflation?
(…) The data centers used for AI require sophisticated computing equipment, cooling systems to keep that equipment from overheating, electric and fiber-optic cables and backup generators to prevent power disruptions. Based on announced and planned developments, Van Nieuwerburgh estimates that spending on the AI build-out through 2032 could come to about $8 trillion—nearly five times the market value of the entire New York City property market.
With so much demand, prices are rising for many of the things that go into the AI build-out. And because those things are used for more than just AI, those price increases are spilling over into the broader economy.
Memory and storage chips, for example, are used in a broad array of consumer-electronics products that includes everything from videogame consoles to cars. Nintendo, Microsoft and Sony have all raised prices on devices. Higher price tags are coming to Apple products, too, according to Chief Executive Tim Cook, who told The Wall Street Journal that the jump in costs was unlike anything he had seen “in any area in over 40 years.”
If AI is as revolutionary as many economists predict, it could eventually cool inflation. That is the lesson from past technological revolutions, which boosted workers’ productivity, making it easier for businesses to meet demand without raising prices. Kevin Warsh, now the Federal Reserve chairman, has previously made that case. (…)
Even under an accelerated timeline, economists at UBS reckon it will be at least a couple of years before AI would start helping to lower inflation. (…)
Already, this is beginning to show up in the inflation data. Consumer prices for computer software and accessories were up about 15% from a year earlier in May, according to the Labor Department. There could be more price increases in the pipeline: The Labor Department’s measure of wholesale electronic components and accessories was up 27% from a year earlier last month. (…)
AI is a shock to demand that could persist for years. (…)
In some instances, the AI build-out could also add to labor costs. Wages for workers who are in demand from data-center construction have been picking up: Average hourly earnings for electrical and wiring-installation contractors were up 6.5% in April from a year earlier, which compared with 3.6% for all private-sector workers. (…)
Earlier this year, Goldman Sachs economists forecast that data centers will account for nearly half of U.S. growth in power demand through 2030. As a result, they saw consumer electricity prices rising about 6% annually this year and next.
To be sure, economists don’t foresee the AI build-out fueling anything like the inflation surge the U.S. experienced when the economy reopened following the Covid-19 crisis. Items like smartphones and videogames represent just a tiny fraction of what people spend every year. Even electricity accounts for only about 2.5% of consumer spending, according to the Labor Department.
Instead, it could serve to keep inflation broadly elevated. Economists expect the May reading of the Fed’s preferred measure of inflation, due out from the Commerce Department on Thursday, will show prices were 4.1% higher than a year earlier. The central bank aims to get inflation to 2%—a level not seen in over five years, as a series of temporary-seeming factors pushed prices higher.
“The more these things happen, the more likely it is that people think, ‘Hey, this is a pattern, maybe I shouldn’t expect inflation to come back down,’” said Jón Steinsson, an economist at the University of California, Berkeley.
That is only one inflation story this year. Add the broad demand pressures stemming from
- the US-Iran war pushing governments and companies to hoard resources and inventories to secure supply “just-in-case”.
- worldwide increases in military spending as governments have realized the need for self-defense.
Importantly, these are all urgent spending that are largely price insensitive.
This Yardeni chart (my rectangles) illustrates the ratcheting up in prices for a large set of commodities used in manufacturing …
… translated into sharply higher producer prices (which exclude tariffs) …
… quickly finding their ways into consumer prices …
… compounded by rising import prices (be mindful of the scale, now 5-10%) …
… which have also been ratcheting up significantly. Could we be in a true regime change?
Thankfully, wages are not adding to inflation so far, nor are they subtracting from inflation still in the +3.5% range:
Purchasing managers, in a clear cost-push situation, are easily pushing their output prices up 4.5%. The composite includes both goods and services.
Yesterday, the Richmond Fed released its quarterly CFO Survey which includes firms that range from small operations to Fortune 500 companies across all major industries. Respondents include chief financial officers, owner-operators, vice presidents and directors of finance, accountants, controllers, treasurers, and others with financial decision-making roles.
Financial decision-makers’ outlooks worsened this quarter amid heightened concern over rising costs and prices in the second quarter of 2026. Since the last survey, CFOs added 1.1 percentage points to their firm’s unit cost and price growth projections for 2026.
AI Demand Begins to Justify Massive Cost of Data-Center Buildout
(…) Global AI sales, excluding China, reached $25 billion in the first quarter of 2026, exceeding the industry’s estimated $21 billion in depreciation costs tied to investments in data centers and chips for the second consecutive quarter.
While the milestone suggests that AI companies are beginning to cover the cost of their capital spending, the margins are thin. Depreciation charges still consume more than two thirds of revenue, leaving a small buffer to cover other costs such as power, labor and financing. (…)
Much of the AI boom has been measured from the supply side, through disclosures from public semiconductor companies like Nvidia Corp. and hyperscalers like Alphabet. Demand has been harder to quantify because many of the most important AI labs, including OpenAI and Anthropic, remain private.
Generative AI revenue, excluding China, reached $110 billion over the past 12 months and is scaling three times faster than any previous information technology wave including the internet, mobile applications and the cloud, according to the report.
The figures are based on a dataset that Exponential View built tracking AI spending across more than 1,000 companies. They used sources including company filings, executive statements, press reporting and cloud-provider disclosures, and then adjusted the figures to avoid double-counting between layers of the AI supply chain.
The analysis assumes a six-year depreciation life for IT equipment including graphics processing units, or GPUs, the chips used to train and run advanced AI models. Some investors argue this is optimistic given the rapid pace of chip innovation, which can render older hardware less valuable within a few years.
If GPUs lose economic value faster than assumed, companies could face higher depreciation charges, asset writedowns or earlier replacement costs. Michael Burry, the investor known for betting against the US housing market before the 2008 financial crisis, has described understated depreciation as “one of the most common frauds of the modern era.”
However, data in the report suggests older chip models are not collapsing in value. The rental price for an hour of access to Nvidia’s H100 chip remains almost 80% of its launch level. “Even into its fourth year, it is completely in demand,” Azhar said, noting it’s become more expensive over the last year, as demand for AI compute outstripped supply of Nvidia’s new Blackwell chips.
That chimes with comments from Matt Garman, the chief executive officer of Amazon Web Services, who said in February the company had not retired six-year-old Nvidia A100 servers due to continuing demand.
The report also shows more users are moving toward open-weight and Chinese AI models such as DeepSeek. Data from OpenRouter, a platform that gives developers access to multiple AI models, shows the share of tokens requested from Google, OpenAI and Anthropic models fell to 33% in June 2026 from 72% a year earlier.
Azhar said that reflects power users moving toward cheaper and faster models for simpler tasks. (…)
That does not necessarily spell trouble for leading foundation-model companies, he added, but it raises the bar for charging higher prices. They will need to compete with “additional services, with more lock-in, and with all of the things that allow you to charge a premium,” he said.
- Micron Delivers the AI Reassurance Wall Street Was Craving
The US memory-chip maker released a quarterly sales forecast that crushed estimates, signaling its growth run remains strong. The results should help rebuild investor confidence following a tech selloff sparked by worries over AI.
How good was the report? Very.
- Revenue will be roughly $50 billion in the fiscal fourth quarter. Analysts had estimated $43.2 billion.
- Profit will be about $31 a share, compared with a projection of $25.31.
- Third-quarter earnings increased to $25.11 a share. A year ago, they were $1.91.
- Adjusted gross margin more than doubled.
Perhaps most importantly, Micron said it has secured 16 strategic customer agreements, which average three years in length. That suggests it can mitigate the boom-and-bust cycles that have plagued the memory-chip industry.
China Issues New Energy Plan at Transition Inflection Point
China published a five-year plan for building a new energy system, aiming to map out a way forward for a sector that’s starting to run up against the constraints of its rapid pivot toward clean electricity.
Every half-decade, Chinese leaders publish a five-year plan outlining economic and societal goals, and then follow it up with several sectoral schemes with more detailed targets and strategies. This year, the broader plan came out in March, and sectoral plans have begun trickling out in recent weeks, including ones on jobs and urban renewal. (…)
It called for China to peak its coal and oil during the 2026-30 period, double non-fossil fuel energy over the next decade, and focus on developing technologies like hydrogen and nuclear fusion. It also sought to make progress on a major gas pipeline from Russia, and boost capacity of generating technologies like nuclear, offshore wind and pumped hydro storage. (…)