Micron beat revenue expectations thanks to the AI boom, but Wall Street still punished the company and the reason reveals what investors now fear most

Published On: March 24, 2026 at 3:45 PM
Follow Us
A close-up of a Micron high-bandwidth memory (HBM) chip used in AI data centers, reflecting the company's 2026 revenue surge.

Micron just delivered another reminder that the AI boom is not only about flashy models and faster GPUs. It is also about the unglamorous parts of the stack, especially memory, and the massive industrial buildout needed to keep data centers stocked.

The big takeaway is simple and a little uncomfortable. As Micron and its rivals race to expand capacity, the environmental footprint of AI starts to look less like an abstract climate debate and more like a very practical question about electricity, water, and wastewater in real communities.

Bigger bets

Micron reported fiscal second-quarter 2026 revenue of about $23.86 billion and told investors it expects fiscal third-quarter revenue of about $33.5 billion, plus or minus $750 million. It also posted an outlook gross margin of roughly 81 percent, which shows how tight supply and strong AI-driven demand can reshape a whole market cycle.

But the stock reaction showed what Wall Street is really watching. Micron said it is boosting its fiscal 2026 capital spending plan by $5 billion and expects fiscal 2026 capex to be above $25 billion, with another meaningful step up projected in fiscal 2027.

The Taiwan expansion

Part of that spending story is physical, not theoretical. Micron has been expanding its global footprint, including its Tongluo P5 site in Taiwan that it acquired from Powerchip Semiconductor Manufacturing Corp, with plans that include retrofitting existing cleanroom space and starting construction of a second cleanroom by the end of fiscal 2026.

Micron has also highlighted major buildouts in the U.S., including Idaho and New York, alongside work in places like Japan, Singapore, and India.

In practical terms, that means more cleanrooms, more tools, more construction, and a lot more demand for local infrastructure that most people never connect to “AI.”

A row of orange industrial robotic arms on an automated assembly line using laser precision to manufacture electronic components.
Advanced robotic arms, like those pictured in this high-tech facility, are increasingly being integrated with AI to optimize manufacturing speed, reduce waste, and manage complex production tasks.

Water is the quiet constraint

Here is the part that tends to get buried under earnings headlines. Chip manufacturing is famously water-intensive because wafers need repeated cleaning with ultrapure water, and facilities also need controlled humidity and cooling.

Sources tracking the industry note that a single chip manufacturing facility can use millions of gallons of water per day, with some estimates reaching as high as 10 million gallons daily for a large plant.

That demand has an ecological side that goes beyond the water meter. Semiconductor processes can generate complex wastewater streams that require heavy treatment, and researchers continue to flag the environmental risks tied to hazardous and persistent waste from chip production. If you have ever lived through watering restrictions during a dry spell, you already know why this becomes a local political issue fast.

Power demand is not slowing down

Water is one pressure point, but electricity is the other. The International Energy Agency has put hard numbers around the trend, estimating data centers used about 1.5 percent of global electricity demand in 2024, around 415 TWh, and projecting that consumption could more than double to around 945 TWh by 2030, with AI a major driver.

Now connect that to Micron’s spending. The company has told investors that cleanroom facility-related capex is a major factor behind its higher spending outlook, which is a reminder that scaling chips is not just about buying machines, it is about running energy-hungry buildings that operate around the clock.

Your electric bill does not care whether the electrons are going to a server rack or a cleanroom air handler.

Micron’s sustainability promises

Micron is not ignoring the environmental side, at least on paper. The company has published goals that include an aspirational net zero target for operational and purchased-energy emissions by 2050, a 42 percent absolute reduction in Scope 1 emissions by 2030 from a 2020 baseline, and a 75 percent water conservation target in fabs by 2030 through reuse, recycling, and restoration.

The nuance is that goals do not automatically scale with output. A capex boom can create an opportunity to build more efficient facilities with higher recycling rates, but it can also amplify local impacts if grid upgrades and water planning lag behind construction.

That is why investors, regulators, and neighbors increasingly want proof in the form of permits, verified conservation rates, and transparent reporting, not just a glossy PDF.

Defense and policy pressures

This is also where “Military and Defense” quietly enters the picture. Governments have been treating microelectronics as a strategic asset, and the CHIPS and Science Act gave the U.S. Department of Commerce $52.7 billion over five years to boost semiconductor manufacturing and research programs.

A close-up of a Micron high-bandwidth memory (HBM) chip used in AI data centers, reflecting the company's 2026 revenue surge.
Despite beating revenue expectations with $23.86 billion in Q2, Micron faces investor scrutiny over its plan to spend more than $25 billion on new facilities.

At the same time, U.S. strategy documents have emphasized domestic capabilities to strengthen security and competitiveness, reflecting how deeply chips are tied into national resilience. The catch is that resilience is not just about geography, it is also about reliable water and power, especially as climate volatility and heat waves make both harder to guarantee.

What to watch next

If Micron’s forecast holds, this buildout is likely to keep accelerating. The company’s own investor materials point to a fiscal 2027 capex step-up and construction-related capex rising by more than $10 billion year over year as it expands globally, which is a signal that the physical footprint of AI is still growing.

So what should readers keep in mind? Watch where these plants and data centers are built, how they secure renewable power, and whether they can actually hit high water reuse and conservation targets without shifting the burden downstream to rivers, aquifers, and municipal systems. AI runs on electrons and water, not magic.

The press release was published on Micron Investor Relations.

Leave a Comment