ARM and Micron — Priced to Perfection?

Two stocks have been dominating my research this week. Both are in the semiconductor space. Both have had extraordinary runs. Both are currently consolidating after big moves. And both are generating heated debate about whether the fundamentals justify another leg higher or whether the easy money has already been made.

ARM Holdings and Micron Technology.

Micron is a new addition to the research. The numbers coming out of this company right now are genuinely jaw-dropping — and the valuation is sparking a fascinating debate between bulls and bears over whether this is the opportunity of the cycle or a classic value trap.

Before We Get Into It — A Quick Note on How I Use Fundamentals

For swing trading, the answer is technicals first, fundamentals as the filter.

Fundamentals tell you what a business is worth. Technicals tell you what the crowd is doing right now. In the short term the crowd wins every time. A stock can stay overvalued on a PE basis for months or years while momentum traders make money. The fundamental analysts calling for a return to sensible levels are often right eventually — but catastrophically wrong on timing.

The rule is straightforward. Use fundamentals to decide which stocks deserve to be on your watchlist — positive earnings, growing revenue, a genuine business with a real moat. Use technicals to tell you when to get in and when to get out. And never use fundamentals to justify holding through a broken chart. If the trend breaks, you leave. The thesis can be completely intact and the stock can still fall 30%.

With that said — let’s look at ARM.

ARM Holdings — The Consolidation Everyone Is Misreading

That’s what’s been happening with ARM since earnings. The majority of people watching this stock are misreading it.

What The Chart Is Actually Saying

The daily tells a clear story. From the January lows at 967 ARM ran to the earnings peak at 268 — a 179% move from the bottom. The consolidation pulled the price back from the highs and it has been compressing for around four weeks. But look at what’s happened during that consolidation — the 9 EMA and 20 EMA have caught up to price and are now sitting just below it. The 50 EMA is rising strongly from below. All three EMAs are converging toward price from below in the right order.

That is a coiled spring. Not a breakdown.

RSI has reset from the overbought high all the way back to 54 — almost exactly the midpoint. Textbook healthy reset. Not oversold, not overbought. Pure neutral, energy reloading. Volume during the consolidation has been declining — smaller bars, less conviction from the sellers. The selling pressure has exhausted itself.

The higher low structure is intact. EMAs are converging below the price. RSI reset to neutral. Volume is declining on the pullback. This is consolidation before continuation — not distribution before breakdown.

The bears have two narratives — here’s why they’re both noise

Here’s the context that matters. Antitrust investigations only happen to companies that have become so dominant that they are reshaping entire industries. Google got investigated. Apple got investigated. Nvidia is under investigation. Microsoft spent years in antitrust proceedings. Every single one of those investigations happened while those companies were at or near their peak growth phase — and every single one of those stocks continued higher through the investigation because the underlying business kept compounding.

ARM’s antitrust situation specifically relates to its licensing model. As ARM moves up the value chain — designing more of the chip rather than just licensing the instruction set — competitors and customers are complaining that ARM is using its dominant position unfairly. That complaint is the market telling you ARM has genuine pricing power and a moat that competitors cannot easily replicate. You do not get antitrust investigations for being a minor player. This is a sign of dominance, not weakness. The short sellers using the investigation as a bear catalyst are handing patient buyers a discounted entry.

The second narrative is the TSMC supply queue. ARM has secured production commitments at TSMC for its new AGI CPU chip designs, but customer demand — which doubled to over $20 billion in six weeks — significantly exceeds the supply it can currently fulfil. The bears say this is a near-term revenue headwind and they’re right about that in a narrow sense.

But here’s the thing — demand exceeding supply by that margin is not a problem. It is a validation. When customer demand for your new product doubles to $20 billion in six weeks, the story is working. The supply constraint at TSMC is a near-term operational friction that resolves as the relationship between ARM and TSMC deepens and production allocation improves. The royalty business — which is the majority of ARM’s revenue — is completely unaffected by fab queue position. It grows automatically every time an AI chip using ARM architecture ships, regardless of who makes it.

Do the fundamentals support another leg higher?

EPS is going from $1.63 in FY2025 to $2.17 in FY2027 to $3.06 in FY2028. The jump from FY2027 to FY2028 of 41% is exceptional — that’s the royalty flywheel engaging at scale. As more AI chips using ARM architecture ship, royalty revenue grows automatically without proportional cost increases.

The gross margin is 97.5-98%. That is one of the highest gross margins of any public company on earth. ARM prints money on every chip that ships. There is no manufacturing cost, no raw material risk, no inventory. Pure intellectual property economics.

The valuation sits at 103x forward earnings, above the historical mean of 76x. But — and this is the crucial point that changes how you read it — ARM has always traded at elevated multiples. It listed into premium valuations and has stayed there. Overvalued relative to what? Relative to its own history it’s above the mean but well below the all-time high of 121x. The multiple has actually been compressing during the consolidation as earnings estimates rise faster than the stock price moves. That is exactly what you want to see.

There’s one more piece of the fundamental picture that most people aren’t modelling correctly. ARM recently renegotiated its licensing agreements and pushed royalty rates significantly higher — particularly for AI-optimised chips. The v9 architecture, which is mandatory for leading-edge AI inference chips, carries royalties 2-3x higher than the older v8 architecture. As the installed base shifts from v8 to v9 — which is happening right now — ARM’s royalty revenue per chip shipped increases automatically even if total chip volumes stay flat. Combined with volume growth from AI acceleration, the revenue trajectory could beat current estimates significantly.

Every AI chip, every smartphone, every data centre — they all use ARM architecture. Every single one pays royalties. The demand curve for what ARM is selling is measured in years, not quarters.

Entry at current levels or on any pullback toward the 9 EMA on the daily around $230. Stop below the consolidation lows at $200. Target one is $245 — the upper range of the consolidation. There’s a lot more that goes into this. I’ll average in, average out and leave something on. This fits my strategy and management of trades. Here on Substack, I am just raising ideas; hopefully, you have your own strategy which helps you get in and out of trades. Don’t follow me blindly. A lot more goes into this than entering here and getting out there.

Many say ARM is priced to perfection. Maybe. But sentiment can push stocks further than fundamentals alone justify — especially when the underlying story is getting better every quarter. The trade is not “ARM is cheap therefore buy.” The trade is “ARM is growing at 30% per year with 98% gross margins and a royalty rate that’s about to step up significantly. The chart is coiled. The antitrust noise is a gift.”

Micron Technology — The Most Interesting Valuation Debate in the Market

The numbers coming out of this company are extraordinary. And I mean that in the literal sense — they are the kind of numbers that make you check if you’ve read them correctly.

The Earnings Picture

EPS was $8.29 in FY2025. The estimate for FY2026 is $58.29. For FY2027 it’s $102.74.

EBITDA margins — in FY2027 — are estimated at 86.4%. On a hardware company. That is software-level profitability coming from a business that makes memory chips.

Free cash flow goes from $1.67 billion in FY2025 to $36.4 billion in FY2026 to $75.8 billion in FY2027.

These numbers look almost too good to be true. But they reflect the HBM supercycle — when memory pricing goes up and volumes explode simultaneously, Micron’s highly operating-leveraged business model generates extraordinary margins because the fixed costs of the fabs are already paid. Every additional dollar of revenue above the break-even point falls almost entirely to the bottom line.

Why it’s trading at 7.51x forward earnings

The reason is the market’s memory — no pun intended.

Memory companies are cyclical. The whole sector is. In 2023, Micron’s EPS went from positive $8.35 to negative $4.45 in a single year. The market has seen this before — extraordinary profitability followed by a brutal downcycle that wipes out everything and more. Every investor in Micron knows the cycle will eventually turn. The 7.51x multiple is the market pricing in the next collapse before it’s arrived.

That is the fundamental tension in the Micron trade. The bears say the discount is appropriate because the HBM cycle will turn. The bulls say HBM is structurally different this time and the market is applying the wrong framework.

The Bear Case

On the surface this is logical. In isolation it would be concerning.

But it misses the structural demand picture entirely. NVIDIA’s H100 had 80GB of high-bandwidth memory. Current GB200 parts have 288GB. The next generation will have a terabyte. The generation after that, two terabytes. That is 25 times more memory per AI accelerator at the same time as 25 times more accelerators are being deployed. No Samsung labour dispute resolution changes that trajectory.

The other bear argument is valuation ceiling — even at modest multiples the stock is already pricing in FY2027 numbers. At 7.51x on the average of the next three years’ EPS estimates, the math doesn’t give you a huge amount of upside without multiple expansion. And multiple expansion in a cyclical business is a big ask.

The Bull Case

Micron has been gaining share in HBM specifically because of technology improvements in their HBM3E product. Their partnership with NVIDIA on next generation memory is deepening. When the most important company in the AI infrastructure story is your primary customer and the relationship is getting closer, not further apart, that is a powerful signal.

The quality metrics are accelerating in every direction. Return on Equity (ROE) has gone from 17.2% in FY2025 to nearly 40% on the latest trailing numbers. Gross margin from 39.8% to 58.4%. Return on capital has increased from 14% to nearly 34%. Every number is moving the right way at the same time.

At 10x forward earnings on FY2027 estimates, Micron has significant upside from current levels. The question is whether the cycle cooperates long enough for those numbers to be achieved.

The Practical Setup

You don’t need to resolve the bull versus bear debate to trade this. The chart gave a clean 20 EMA pullback with RSI approaching oversold. The stock bounced decisively from that level. The momentum trade is straightforward — buy the bounce from the 20 EMA, ride the recovery toward the previous high, and exit before the next earnings in late June. You are not making a two-year investment decision. You are trading a technical bounce in a confirmed uptrend backed by exceptional fundamentals.

The Samsung noise is a bear narrative, not a structural change. The 7.51x forward PE means even moderate earnings beats push the stock higher because there is almost no multiple expansion required — the earnings are doing all the work.

Entry around current levels or any dip toward 700. Stop at 600 below the consolidation lows. Target: the previous high at 818. Hard exit before late June earnings. That is my trade. As above, manage your own trade.

What ARM and Micron share is a position at the absolute heart of the AI infrastructure buildout. ARM provides the architecture that every AI chip runs on. Micron provides the memory that every AI accelerator needs in ever-increasing quantities. Neither of these companies is a speculative bet on AI becoming real — AI is real, it’s here, and these companies are already being paid for it in their earnings numbers.

The debate is not whether the demand is genuine. It clearly is. The debate is whether the current prices already reflect everything that’s coming — whether the crowd has already priced in years of growth that haven’t appeared yet.

My view, for what it’s worth, is that both stocks have more road ahead. The royalty rate shift at ARM is a genuine earnings catalyst that the market hasn’t fully modelled. The HBM structural demand at Micron is real and durable in a way that conventional DRAM cycles weren’t. And both charts are set up for continuation rather than breakdown.

But I hold that view loosely. The stop levels exist for a reason. If the higher low structure breaks on either name, I’m out — regardless of how compelling the fundamental case looks on paper.

That’s the discipline that makes the difference between trading and hoping.

Not financial advice. I’m a bloke with a chart and an internet connection, not a financial advisor. Everything I write on here is my own research, my own opinions, and my own trades — for better or worse. The wins are mine. The losses are mine. And if you follow anything I do and it goes wrong, that’s on you, not me. Do your own research, size your positions properly, and never risk money you can’t afford to lose. We’re all just trying to figure this out.