Micron vs AMD vs SanDisk vs ASML vs Seagate: A 6-Round AI Infrastructure Showdown

Why These Five and Not the Obvious Names

The AI semiconductor conversation has a NVIDIA-and-TSMC problem. Those two suck up most of the oxygen. Everyone knows them, everyone analyzes them, and most active investors already hold a position. The alpha — if it's still there — is a layer below.

To build a GPU you need HBM. To make HBM you need EUV lithography. To run an AI data center you need massive cold storage. That's where Micron, SanDisk, ASML, AMD, and Seagate live.

I ran the five through six head-to-head financial rounds: net margin, forward revenue growth, cash ROIC, levered FCF margin, profit-adjusted PE, and debt-to-equity. Three points for first, two for second, one for third. The scoreboard at the end matters less than the shape of how each company won and lost.

Round-by-Round at a Glance

Round	1st	2nd	3rd
Net Margin	Micron 41.5%	SanDisk 34.2%	ASML 29.7%
Revenue Growth Forecast	Micron 194.1%	SanDisk 162.9%	AMD 41.6%
Cash ROIC	Seagate 50.3%	ASML 40.1%	SanDisk 34.4%
Levered FCF Margin	SanDisk 33.8%	ASML 26.6%	AMD 22.9%
Profit-adjusted PE	Micron 0.27	SanDisk 0.61	ASML 1.39
Debt-to-Equity	SanDisk 5.9%	AMD 6%	ASML 13%

The Five Characters

SanDisk (SNDK) — The balanced champion. Hit the podium in four of six rounds, twice as the leader. A 5.9% debt-to-equity paired with a 33.8% FCF margin is rare in any capital-intensive industry. Its only 'weakness' was placing third in cash ROIC (34.4%) — still in the medals.

Micron (MU) — Cheapest profit, highest growth. Three first-place finishes (net margin, growth, profit-adjusted PE). It lost the overall fight because capex pressure compressed cash ROIC and FCF margin. I read that as a phase, not a flaw.

ASML — Wide moat, steady hand. Growth came last at 19.7%, but 40.1% ROIC, 26.6% FCF margin, and 13% debt-to-equity tell the stability story. The EUV monopoly doesn't fully show up in any of these six rounds.

AMD — Premium price, average efficiency. Made the podium once (revenue growth, 41.6%), but stumbled with a 13.4% net margin and a 3.89 profit-adjusted PE. You're paying roughly 14x more per unit of profit than you would for Micron.

Seagate (STX) — Cash ROIC king, balance-sheet hostage. The 50.3% ROIC headline is impressive until you see it was generated on top of 381.6% debt-to-equity.

How I'd Actually Construct This

I wouldn't treat these five as one bucket.

SanDisk + Micron is the core pair. SanDisk gives you balance, Micron gives you the asymmetric setup. SanDisk plays downside defense, Micron plays upside expansion.

ASML sits in its own category. Even if the memory cycle cools, EUV demand doesn't. Portfolio insurance with growth optionality.

AMD is a wait-for-better-entry. Not bad — just expensive relative to what you can buy in this group right now. When AMD's valuation closes the gap toward Micron's, the calculus changes.

Seagate is the one I'd skip in this cycle. AI storage demand is real, but a 381.6% debt-to-equity does not coexist comfortably with any rate shock or demand slowdown.

The Limits of This Framework

I want to be honest about what six rounds of ratios miss.

ASML's EUV monopoly is structural and doesn't compress into a quant metric. AMD's data center GPU share trajectory doesn't either. Neither does the NAND/DRAM mix difference between Micron and SanDisk.

I use this kind of comparison as a starting point for where to dig, not as a verdict on what to buy. It ranks priorities — it doesn't make conclusions.

FAQ

Q: Why did SanDisk beat Micron when Micron won more first-place rounds?

A: Scoring rewards consistency on the podium, not raw wins. Micron took three firsts but missed the medals entirely on cash ROIC and FCF margin. SanDisk placed in nearly every round, which compounds across six categories.

Q: Is the profit-adjusted PE a standard metric?

A: It's a custom ratio that forces forward PE to answer for actual net margin. Two companies with the same headline PE can look very different once profitability is factored in. Useful for cross-company comparison in a single sector.

Q: Should I just buy the index instead?

A: A sector ETF gives you average exposure, but the spread between best and worst in this group is enormous (SanDisk 13 points vs Seagate 3). The framework's point is to separate the leadership names from the laggards within the AI infrastructure theme.