Micron Reports This Week — the 940% Earnings Test That Could Settle the AI Argument

This blog has been circling one date for two weeks. The weekend guide called it the scheduled catalyst; Friday’s rebound story treated it as the test the shrug was buying time for. Now it is here: Micron reports fiscal fourth-quarter earnings at the end of September, and per Investor’s Business Daily the Street is modeling roughly 940% year-over-year earnings growth. Numbers like that are not forecasts anymore — they are bars. Expectations that extreme function as a referendum on the memory shortage that has powered this stock past $1,000 and, by extension, on the entire AI supply chain built on the assumption that memory stays scarce and expensive.

What makes this print unusual is that both camps have publicly agreed it matters. The shortage thesis and the glut thesis do not disagree that Micron’s guidance is the datapoint — they disagree about which way it points.

The shortage case: three posts of evidence, one more week to prove it

This blog’s June coverage laid out the thesis: Micron posted $13.6 billion in quarterly revenue and warned memory supply would stay tight through 2026. The follow-through has been blunt. Micron crossed $1,000 a share in August, AMD crossed a trillion dollars in market value last week, and the squeeze has reached companies nowhere near the chip trade — Apple’s new CEO is reportedly planning layoffs as memory-chip costs rise, the same cost pressure that had Apple calling price hikes unavoidable back in June. When the world’s most supply-advantaged consumer company starts cutting costs over memory prices, the shortage is not a narrative. It is a purchase order.

And the demand side of the argument keeps writing bigger checks: OpenAI is reportedly negotiating a $30 billion funding round at a $1.4 trillion valuation — a number that only makes sense if compute demand, and the memory underneath it, keeps compounding. A strong Micron print with confident guidance does not just validate one stock. It validates the assumption the whole capital cycle rests on.

The glut case: Burry’s shorts and the cycle nobody remembers

Against all of it stands Michael Burry, who has called the Nasdaq-100 historically overvalued and historically top-heavy, and has increased shorts on Micron, the iShares Semiconductor ETF, Nebius Group, and Palantir. His argument is not that AI is fake — it is that memory is an industry with a century-long habit: shortage drives prices, prices fund capacity, capacity kills prices. He points to expanding Chinese DRAM and NAND production, a risk analysts at the Center for Strategic and International Studies have flagged as future surplus and potential dumping, and to executives already disputing the shortage itself — Acer’s CEO Jason Chen, asked about memory shortages, answered bluntly that there is no continuous shortage possible with China’s production capacity consistently increasing. Burry’s own framing: over the next two years the shortage blows off, and memory has a down cycle again.

Note what the glut case does not require. It does not require AI demand to collapse. It only requires supply to arrive on schedule — which is what supply has done in this industry, every cycle, for decades. That is what makes Micron’s guidance the perfect test: the shortage case needs the company to describe tightness extending into 2027; the glut case needs only the faintest hint that pricing has peaked. One sentence from one executive call lands on one side or the other.

Before the number: the pre-earnings checklist

There is no reliable way to predict the print — 940% growth expectations cut both ways, because a beat that merely meets the bar can still sell off on guidance. What you can control is what you own and what you will do about it. The first step is seeing the exposure in one place: InvestSheet syncs positions from Robinhood, Fidelity, Schwab, and 35+ other brokerages into a single Google Sheet, so chip exposure shows up once with live formulas no matter which account holds it:

With the numbers visible, run three checks. First, the single-name check: has MU or any chip runner drifted past your position cap after a month that took Micron past $1,000 and AMD past a trillion? Second, the theme check: direct chip shares plus semiconductor weights inside your funds — do they exceed the limit you set when the trade was smaller? Third, the plan check: write down now what you will do on a beat versus a miss — a trim rule, a rebalance rule, a do-nothing rule — so Thursday’s reaction is executed, not improvised. Earnings weeks are won before the open, not during it. The sheet does not tell you which way Micron goes. It tells you which way you are exposed — and that is the part you can actually fix tonight.

Frequently asked questions