Micron: Memory Crunch Runs to 2028

Consumers hoping the memory crunch will ease once AI data-center spending cools may want to sit down. Micron, one of the three major memory chip makers, has told investors that supply and demand for memory and storage will stay badly out of balance well past the point most forecasts had penciled in for relief — and that robots, not servers, may be the next big drain on supply.

The warning arrived alongside the company's fiscal Q4 2026 results, which brought booming profits and a set of guidance comments that paint a much longer runway for the shortage than buyers of PCs, phones, and SSDs would like.

Micron's guidance gets blunter

In the guidance accompanying its results, Micron framed the next two calendar years in unusually stark terms. According to Wccftech reports, the company observed: "We expect memory and storage supply-demand conditions to be much tighter in calendar 2027 and 2028 than they were in 2026."

That line alone pushes the expected timeline for relief out by a full year beyond the widespread 2027 prediction. But Micron went further, addressing the question of whether new manufacturing capacity might close the gap.

"Even with additional industry DRAM cleanroom space plans, with robust demand trends including new upside requests from customers, we do not have line of sight to when supply and demand will return to balance."

— Micron, in guidance accompanying its fiscal Q4 2026 financial results

The phrasing matters. Micron is not saying supply will be tight; it is saying the company cannot currently see the point at which it stops being tight — even after accounting for planned cleanroom expansion across the industry.

The AI bubble question, answered sideways

A common assumption among consumers is that if the AI boom deflates — if data-center operators stop buying accelerators and memory at the current pace — RAM prices should fall back to earth. Micron's guidance suggests that assumption may be too simple.

The company made the point, highlighted on X by Shay Boloor, chief market strategist at Futurum Equities, that even if the current AI bubble deflates and server RAM demand in data centers eases as a result, a high level of demand could still come from what Micron calls 'physical AI' — meaning robots.

That is a different demand curve than the one the industry has been planning around. Data-center memory demand is concentrated, forecastable, and tied to a handful of very large buyers. Robot demand, if it materializes, would be distributed across a manufacturing base that does not exist at scale yet.

Robots and the 200GB problem

The numbers Micron attached to its robot forecast are the part most likely to get quoted back at the company in the years ahead. The company said "humanoid robots could need 200GB+ of DRAM and multiple terabytes of storage per unit with several physical AI customers already sampling next gen products."

Micron also described this as potentially a "significant driver of memory and storage demand by the end of this decade."

To put the per-unit figure in context: a single humanoid robot, by Micron's estimate, would carry more DRAM than most consumer desktops and laptops sold today, plus storage measured in terabytes. If humanoid production scales the way robotics firms have promised, that is a new category of memory buyer layered on top of the data-center demand that already has supply stretched.

Micron also noted that several physical AI customers are already sampling next-generation products — meaning this is not purely a forecast about a distant market, but a claim that design activity is underway now.

Where the 2030 idea came from

Micron's comments feed into an existing narrative that the RAM crisis lasts until 2030. One of the earliest public mentions of that timeline came from Nvidia's CEO, according to the reporting.

The robot angle reinforces that framing rather than replacing it. Under the argument Micron is making, even a meaningful pullback in AI data-center spending would not necessarily restore balance, because a second wave of demand — physical AI — could absorb the slack.

That said, the robot timeline is the softest part of the forecast. Robotics hardware at humanoid scale remains early, and per-unit memory estimates depend on designs that have not shipped in volume.

The incentive to talk up the future

There is a reason to read Micron's robot forecast with some caution. In these kinds of forward-looking statements, it is in the firm's interest to build up its future prospects and drive more investment — a point raised directly in the analysis of the guidance.

Micron is not a neutral observer of the memory market. It sells the product whose scarcity it is describing. Guidance that extends the shortage into 2028 and dangles a robot-driven demand wave at the end of the decade supports a narrative of durable pricing power.

None of that makes the supply picture wrong. The cleanroom-capacity caveat — that even added DRAM manufacturing space does not close the gap — is a specific, checkable claim about industry capacity that goes beyond promotional language.

What the supply numbers actually say

Stripped of the robot speculation, the core of Micron's guidance contains a small set of concrete claims worth tracking:

• Memory and storage supply-demand conditions are expected to be much tighter in calendar 2027 and 2028 than in 2026.
• Micron says it has no line of sight to when supply and demand return to balance, even with additional industry DRAM cleanroom space plans.
• Humanoid robots could need 200GB+ of DRAM and multiple terabytes of storage per unit.
• Micron frames physical AI as a possible significant driver of memory and storage demand by the end of this decade, with several customers already sampling next-gen products.
• The results in question cover Micron's fiscal Q4 2026, with booming profits.

The company has not offered a date by which it expects balance to return — and by its own account, it cannot currently identify one.

Why consumers keep waiting

For anyone watching the price of a RAM kit, an SSD, or a laptop, the practical takeaway from Micron's guidance is that the wait for normal pricing has no visible end date attached to it. The shortage that has defined component pricing through 2026 is being projected forward, not backward.

The 2027 timeline had already been widely discussed. Micron's contribution is to place 2028 on the same level, and to argue that capacity additions alone will not be enough to fix it.

That leaves buyers — individuals building a PC, businesses refreshing fleets, and OEMs setting next year's bill of materials — planning against a supply environment with no announced inflection point.

What to watch from here

The verifiable signals to track are capacity, not commentary. Whether the additional DRAM cleanroom space Micron references actually comes online on schedule will determine whether its own caveat holds. Whether humanoid robot programs move from sampling to volume production will determine whether the physical AI demand forecast is a real second wave or a slide in an investor deck.

Micron's own framing leaves both questions open. The company says it expects tighter conditions in 2027 and 2028, and that it cannot see the return to balance — a statement that is notable precisely because it is a negative claim about the company's own visibility.

For the moment, the memory market's defining feature is not a spike or a crash but duration: a shortage that Micron is now telling investors extends through the end of the decade's first two years, with a robot-shaped question mark sitting behind it.