The Business
Applied Materials (NASDAQ: AMAT) is the largest semiconductor-equipment company in the world — the "materials engineering" layer underneath virtually every chip. Its tools deposit, etch, modify, and inspect the thin films that make a transistor work. The one-line framing: it sells the shovels for the entire chip build, regardless of which chipmaker wins. TSMC, Samsung, Intel, and the memory makers all buy from Applied.
Two pieces of the mix matter most for the AI story right now:
- Advanced packaging & HBM. Applied is the leader in the tools for high-bandwidth memory and 3D chiplet stacking — the exact bottlenecks in building an AI accelerator. Management now expects packaging revenue to grow more than 70% in calendar 2026.
- Leading-edge foundry-logic and DRAM. The equipment intensity of each new node keeps rising, and Applied guided to particularly strong second-half growth in DRAM and leading-edge logic — the capacity being built out for AI compute and memory.
This is a picks-and-shovels business: Applied doesn't need to pick the winning chip, it needs fabs to keep spending on capacity. That is both the bull case (broad exposure to the whole buildout) and the bear case (spending is cyclical — more on that below).
Where the Price Is
As of the September 17 session (quote pulled live from Yahoo Finance):
- Price: ~$417 · essentially flat on the day
- Day range: $413.56–$428.54
- 52-week range: $183.83–$739.67 — the stock is roughly 44% below its high, and still well off the low
- 50-day MA: ~$504 — price is ~17% below it. The intermediate trend is down.
- 200-day MA: ~$415 — price is sitting right on it, within half a percent. This is the line the whole chart is deciding at.
- RSI(14): ~34 — near oversold. This is a beaten-down, washed-out tape, not an extended one.
The technical read is the mirror image of a name like Marvell. AMAT isn't a high-flyer digesting a run — it's a leader that already round-tripped a huge move (it saw $739 at the euphoria peak) and is now testing its 200-day with momentum against it. The risk here isn't "don't pay the top multiple," it's "is this support, or the pause before it keeps going?"
What Q3 Actually Said
The quarter was, by the numbers, the best in the company's history:
- Revenue: $9.12B, a record — up 25% YoY and up 15% sequentially, the highest quarter-on-quarter growth Applied has ever reported
- EPS (non-GAAP): $3.50, a record, up 41% YoY (GAAP EPS $3.17, up 43%)
- Gross margin: 50.4% non-GAAP — the 13th consecutive quarter of year-over-year gross-margin expansion
- Cash: record operating cash flow of $3.04B; $860M returned to shareholders ($440M buyback + $420M dividend), with $12.8B still left on the buyback authorization and a stated policy of returning 80–100% of free cash flow
- Guidance (Q4 FY26): revenue $10.25B ±$500M — up ~51% YoY — and non-GAAP EPS $4.02 ±$0.20 — up ~85% YoY, with gross margin holding ~50.4%
- Outlook: management raised its semiconductor-systems expectations for calendar 2026, said it expects to grow faster than the market, and guided to "another strong growth year" in 2027
Read plainly: revenue, profit, and margins all at records; cash returns aggressive; and a next-quarter guide that accelerates rather than fades. On the fundamentals there is nothing to complain about here.
Why It Fell on Records
So why does a company printing all-time highs trade 44% off its own? The market isn't arguing with the quarter — it's pricing the cycle.
Semicap is cyclical, and the market pays the lowest multiple at the peak. Wafer-fab-equipment spending moves in waves. Equipment stocks characteristically top before the fundamentals do — the tape de-rates while the numbers are still climbing, because it's trying to price the digestion year on the other side. Record results plus a downtrending stock is the classic signature of a market betting the peak is near. That's the single biggest thing weighing on AMAT: not this quarter, but the fear of what calendar 2027–2028 WFE looks like after a historic build.
China and export controls. Applied carries meaningful China revenue and sits directly in the path of U.S. export policy. Every tightening of the rules is a potential air pocket in a slice of the business, and that overhang never fully leaves a semicap name.
AI-capex peak worry. The same macro fear pressuring the whole AI-infrastructure complex — that hyperscaler spending is near its peak rate of change — lands squarely on the company selling the capacity equipment. If capex growth flattens, equipment orders are the first derivative that rolls.
Momentum unwound. The crowd that chased the stock to $739 is gone, and a name that has lost its 50-day by 17% has no momentum bid underneath it until that changes.
Valuation
Here's the tension the price creates. Annualizing the Q4 guide (~$4.02 non-GAAP EPS) puts Applied around a mid-teens forward earnings run-rate, which at ~$417 is roughly 26x — and cheaper still against 2027 estimates if the growth Management is guiding to shows up. On trailing numbers it's mid-30s. For a company compounding EPS at 40%+ with 50% gross margins, a fortress balance sheet, and a buyback shrinking the float, ~26x forward is not a rich multiple — it's roughly a market multiple on a well-above-market grower.
But that's exactly how cyclical peaks look. A low multiple on peak earnings is a trap if the earnings are about to roll; a bargain if they're not. The entire debate on AMAT collapses to that one question — is this peak earnings, or a way station?
The Thesis
Bull case: Applied is the indispensable, diversified leader of the AI chip build — it doesn't have to pick the winning accelerator, it just needs fabs to keep adding capacity, and it's levered to the two tightest knots in the system (HBM and advanced packaging, guided +70%). Records across revenue, profit, and margin; 13 straight quarters of margin expansion; record free cash flow with 80–100% of it returned and $12.8B of buyback left; and a next-quarter guide of +51% revenue and +85% EPS. At ~26x a forward run-rate on that profile, with the stock washed out on the 200-day, you're being handed a leader at a market multiple because the tape is scared of a cycle that management says has another strong year in it.
Bear case: the market prices semicap at low multiples near the top for a reason. If calendar 2027 is a WFE digestion year, today's records are the peak, estimates get cut, and "cheap on peak earnings" gets cheaper — the multiple and the E move down together. On the chart it's already in a downtrend, 17% below its 50-day, with no support proven until it holds the 200-day. And China/export-control risk is a live tail that can dent a real slice of revenue with one policy headline. This is a knife that hasn't been caught yet.
What would change my mind: "digestion" or "pause" language on the next call, an order-growth deceleration in DRAM/leading-edge, or a China export-control escalation would confirm the bear cycle read and take precedence over any valuation argument. On the other side, a decisive reclaim of the 50-day (~$504) on volume would say the de-rate is finished and the market has decided 2027 is a growth year after all.
Trade Plan
Setup: a best-in-class business the market is pricing for a cycle top. That's a more interesting kind of setup than a priced-for-perfection name — here the pessimism is doing the work for you — but it comes with a downtrending chart, so this is stalk-and-scale, not chase.
Levels that matter:
- ~$415 (200-day) — the line the stock is deciding at right now. Holding and basing here is the constructive tell; a decisive close below it opens real downside and says the cycle-top bears are winning.
- ~$504 (50-day) — the trend-repair level. Until AMAT reclaims it, the intermediate trend is down and rallies are suspect.
- $183.83 — the 52-week low and the reference point if a full cyclical de-rate plays out. Naming it isn't a forecast; it's the floor of the range.
- $739.67 — the prior euphoria high. Only relevant much later, and only if the cycle extends.
How I'd treat it: not a name to swing full-size at a single price. The constructive approach is a starter while it defends the 200-day, with the RSI basing out of oversold, and adds on either a hold-and-turn off ~$415 or a reclaim of the 50-day on real volume. Losing the 200-day decisively is the signal to stand aside and let it find a lower base, not to average into it.
Invalidation: a decisive weekly close below the 200-day (~$415) on heavy volume shifts this from "washed-out leader building a base" to "cyclical de-rate still in force," and the next real references are materially lower. Size for a cyclical name that already fell 40%+ once — this can keep trending.
Horizon: this is a cycle call as much as a company call. Own the thesis only to the extent you believe the AI-capacity build has another leg — management is guiding as if it does. Trade the 200-day; hold the business only if the orders keep confirming.
Bottom Line
Applied Materials delivered the best quarter in its history and guided to an even bigger one, and the stock is down 44% from its high anyway — because the market isn't voting on the quarter, it's voting on the cycle. Both things are true at once: the business is firing on every cylinder, and equipment stocks de-rate ahead of the peak they're afraid is coming. The resolution is on the other side of this WFE cycle. If the AI-capacity buildout has another year in it — which is exactly what a +51% revenue guide is telling you — then ~26x forward on a washed-out leader is a gift and the 200-day is a base. If 2027 turns into a digestion year, the records are the top and patience gets rewarded with lower prices. Applied is the highest-quality way to own the shovels of the AI build; the debate isn't the company, it's where we are in the cycle. I'd stalk the 200-day, let it prove a base, and scale rather than chase — and I'd treat a lost 200-day, or any "digestion" language from management, as the tell that the cycle bears were right.
Price and technical data pulled live from Yahoo Finance at the September 17, 2026 session. Fundamentals from Applied Materials' fiscal Q3 2026 release (reported August 13, 2026) and earnings-call coverage. Valuation figures are approximate and based on the company's own Q4 guidance annualized. Levels are point-in-time and will move.