Applied Materials reports record AI-driven growth

AI demand moves from narrative to tool orders

Applied Materials’ latest investor message is unusually direct: AI data centers are reshaping the semiconductor equipment cycle. At the Goldman Sachs Communacopia + Technology Conference on September 9, CEO Gary Dickerson said AI, especially the data-center segment, is now an “enormous driver” for the company’s business, and he framed the shift in wafer starts as a structural change rather than a single-quarter inventory swing [1].

The clearest comparison came from leading-edge foundry demand. Dickerson said smartphone wafer starts were about four times larger than data-center wafer starts a few years ago; today the two are roughly equal, and in a couple of years he expects data-center wafer starts to be about twice smartphone wafer starts [1]. That is the kind of reversal that explains why equipment companies, not only chip designers, have become central to the AI trade.

Applied’s role is not to sell the accelerator, the memory stack or the cloud service. It sells and services the manufacturing systems needed to make those chips at scale. That makes the company a second-order AI beneficiary, but not a marginal one. If hyperscalers and AI labs keep ordering custom accelerators, high-bandwidth memory and advanced packages, chipmakers need more deposition, materials engineering, inspection, process control and packaging capability.

A record quarter, and a stronger setup

The financial backdrop gives the AI story credibility. Applied Materials reported fiscal third-quarter revenue of $9.12 billion, up 25% year over year, while non-GAAP earnings per share rose 41% to a record $3.50 [3]. The company also generated record quarterly operating cash flow of $3.04 billion, with GAAP gross margin of 50.3% and non-GAAP gross margin of 50.4% [3].

Management’s fiscal fourth-quarter guide adds another layer. Applied continues to point to revenue of about $10.25 billion, plus or minus $500 million, and non-GAAP EPS of $4.02, plus or minus $0.20 [3]. At the midpoint, that implies a sequential revenue step-up from an already record fiscal Q3, showing that the AI equipment cycle is not confined to a single strong print.

The market is watching whether this is a peak or a plateau. The September 9 TradingKey analysis argued that the stock’s technical setup was improving after a rebound from a demand zone, but it also highlighted risks including China export restrictions, execution pressure and the possibility that tighter financing could eventually slow AI infrastructure projects [3]. Those caveats matter because equipment stocks often discount orders well before factories begin producing chips.

Customer forecasts are still rising

At Citi’s 2026 Global TMT Conference on September 8, Applied CFO Brice Hill said the company’s rolling eight-quarter forecasts from major customers, especially DRAM and leading-edge logic manufacturers, had increased throughout the year [2]. MarketBeat reported that Hill tied those rising forecasts to AI systems demand and to cloud-service capital-expenditure expectations exceeding $700 billion for U.S. companies and approaching $1 trillion globally [2].

The message is important because Applied’s demand signal comes from chipmakers’ production plans, not from consumer excitement around AI chatbots. If customers are raising eight-quarter equipment forecasts, they are making concrete assumptions about future capacity needs. That is why Hill’s comments support the view that AI infrastructure investment is migrating into the manufacturing supply chain.

Applied is also tracking more than 100 fab projects globally, with more than 10 new projects added in each of the past two quarters [2]. Clean-room availability remains a medium-term constraint, according to the same reporting, which means the limiting factor is not simply whether chipmakers want tools, but whether fabs and facilities can be readied fast enough [2].

Advanced packaging and DRAM become the pressure points

The fastest-growing parts of wafer-fab equipment are directly tied to AI systems. Dickerson said leading-edge foundry-logic, DRAM including high-bandwidth memory, and advanced packaging account for about 80% of wafer-fab equipment growth this year, with a similar mix expected in 2027 [1]. That concentration explains why Applied’s AI exposure is deeper than a generic semiconductor rebound.

Advanced packaging is especially important because AI accelerators increasingly combine logic, memory and chiplets into dense systems. Applied’s advanced-packaging business generated about $1.4 billion in revenue last year and is expected to grow more than 70% this year [2]. The company is also investing in panel-level packaging, though that remains a developing opportunity rather than a high-volume production business today [2].

DRAM may be the larger swing factor. Hill said DRAM is entering a more significant greenfield investment cycle as AI-related memory needs rise [2]. He estimated that 100,000 wafer starts of greenfield DRAM capacity could require roughly $10 billion of process equipment, while upgrading an existing fab may require only about a quarter of that investment [2]. For Applied, the distinction between upgrades and new fabs is therefore economically decisive.

Broadcom’s forecast adds a demand-side check

The broader AI chip story also supports Applied’s case. Broadcom, a major custom-silicon and networking supplier, now expects AI semiconductor revenue of about $58 billion in fiscal 2026, rising to $115 billion in fiscal 2027 and $230 billion in fiscal 2028 [4]. That 2028 figure is a forecast, not achieved revenue, but it is a striking signal of how much compute demand large AI customers are communicating to suppliers.

Broadcom CEO Hock Tan separately described the path as supply-constrained, with data-center power, site readiness and construction timelines becoming real-world limits on growth [5]. That dovetails with Applied’s own clean-room and fab-readiness comments: the bottleneck is not merely chip design, but the physical infrastructure needed to turn AI spending into shipped systems [2], [5].

Some stock analysis has extrapolated Broadcom’s forecast into a possible share price above $900 by 2028, assuming the AI revenue guide is achieved and margins remain strong [4]. That is an outlook, not a result, and it should be treated as a market scenario rather than a certainty. For Applied, the more relevant takeaway is not Broadcom’s hypothetical valuation, but the scale of manufacturing demand implied by custom accelerators, networking chips, HBM and advanced packages.

The question now: durability

Applied Materials’ record AI-driven growth story is therefore not just about one strong quarter. It is about whether AI infrastructure spending can keep pulling the semiconductor supply chain into a multi-year capacity expansion. The company’s latest conference comments point to rising customer forecasts, broader fab tracking, faster advanced-packaging growth and a DRAM cycle that may require new greenfield capacity [1], [2], [3].

The risks remain visible. Export controls can redirect or delay tool demand, advanced packaging has to scale reliably, and the economics of AI data centers still depend on power, financing and end-user monetization. But as of the latest 72-hour window, Applied’s message is that the AI buildout is showing up where it matters for an equipment maker: in customer forecasts, fab plans, process complexity and cash flow.

Moore’s law may no longer arrive neatly as a smaller transistor. For Applied Materials, it is increasingly arriving as a purchase order for more complex tools, more packaging capacity and more ways to manufacture the AI era at industrial scale.

Sources from the last 72 hours

1. [1]Applied Materials, Inc. (AMAT) Presents at Goldman Sachs Communacopia + Technology Conference 2026 TranscriptSep 9, 2026, 9:25 PM UTC
2. [2]Applied Materials Says AI Boom Is Driving Semiconductor Equipment Demand HigherSep 8, 2026, 12:03 PM UTC
3. [3]Applied Materials Stock Breaks Wedge as Rising Fab Forecasts Put $493 in FocusSep 9, 2026, 3:00 PM UTC
4. [4]Broadcom Forecasts $230 Billion in AI Semiconductor Revenue in 2028. The Stock Could Reach $900 Per Share as a Result.Sep 9, 2026, 10:21 PM UTC
5. [5]Broadcom at Goldman Sachs conference: ai growth meets supply limitsSep 8, 2026, 8:12 PM UTC

AI-generated article based on recent web research, then preserved as a dated editorial snapshot.