I’ve been tracking the semiconductor equipment trade for a while now, and there’s a pattern I keep noticing.
Every time I think the artificial intelligence (AI) infrastructure story has been fully priced in, another analyst note lands, reframing the whole thesis. Citi’s latest call on Applied Materials (AMAT) is exactly that kind of note.
Shares of 58-year-old Applied Materials surged more than 7% intraday on June 17, trading near $610, after Citi raised its price target to $710 from $550, according to a note shared with TheStreet.
The catalyst wasn’t a new product or an earnings beat. It was a memory chip most people have never thought about twice: NAND flash.
CEO Gary Dickerson had already flagged the momentum building beneath the surface.
“The rapid global build-out of AI computing infrastructure, combined with Applied’s strong leadership positions in leading-edge logic, DRAM, and advanced packaging, provides an exceptionally strong foundation for sustained, multi-year revenue and profit growth,” Gary said in a company statement following Q2 earnings.
Citi is arguing that the AI story is no longer just about GPUs and high-bandwidth memory. It’s spilling into NAND in a way the market hadn’t fully connected yet.
Also Read: Applied Materials Inc. (AMAT) Stock Price Latest News and Updates
The agentic AI memory squeeze driving Citi’s NAND thesis
Here’s the mechanism I find most compelling in Citi’s note. It’s worth slowing down to explain because it’s genuinely a new wrinkle in the AI infrastructure story.
Agentic AI systems, or models that perform multi-step reasoning and tool use, require a KV cache, which stores intermediate computational state during inference.
More Applied Materials:
- Morgan Stanley resets Applied Materials stock price forecast
- Citi massively revamps Applied Materials stock price target
- Applied Materials makes unexpected AI move as demand explodes
As these workflows get more complex, the KV cache footprint balloons, according to the Citi note. The problem is that high-bandwidth memory (HBM) and DRAM, the chips typically used for this, are both expensive and currently supply-constrained.
Citi’s analysts framed the dynamic plainly in the note.
Rather than indicating weaker demand, this underscores a widening gap between required and available memory.
Companies are responding by offloading that intermediate state to NAND flash — cheaper, higher-capacity storage that can absorb the overflow.
I think this part of the thesis deserves the most attention. It’s not that NAND demand is rising because phones and laptops need more storage. It’s that the entire architecture of AI inference is being redesigned around NAND as a release valve for a DRAM shortage. That’s a structural shift, not a cyclical bump.
Citi also pointed to ongoing innovation in high-performance NAND formats such as HBF and XL-Flash as reinforcing the trend, according to the note.
How big Citi thinks the wafer equipment market could get
Citi’s revised numbers on the broader wafer fabrication equipment (WFE) market are aggressive, and they extend further out than most forecasts I’ve seen.
Citi’s bull case WFE estimates, according to the note:
- 2026: Approximately $145 billion
- 2027: Approximately $200 billion
- 2028: Approximately $250 billion, implying another 25% growth on top of 2027
Citi attributed the 2028 confidence to continued capacity constraints and expansion at TSMC and memory makers, along with recent progress at Intel and Samsung foundries, according to the note.
My read on this: Forecasting three years out in a cyclical industry like semiconductor equipment is unusual.
Related: Citi resets AMD stock price target on key move
Citi is essentially betting that the AI buildout has enough duration to outlast the typical boom-bust pattern this sector has lived through for decades.
For Applied Materials specifically, Citi modeled total revenue growth of 30% year over year (YoY) in calendar year 2027 and 22% YoY in 2028, including 35% and 25% growth from its Silicon segment and 14% and 13% from Applied Global Services (AGS), according to the note.
Applied Materials’ record quarter set the stage for this upgrade
Citi’s bullish revision didn’t come out of nowhere. Applied Materials reported record Q2 fiscal 2026 results in May, and the numbers gave Wall Street plenty to work with.
Q2 fiscal 2026 highlights:
- Revenue: $7.91 billion, a record, up 11% YoY
- Non-GAAP EPS: $2.86, a record, up 20% YoY
- GAAP EPS: $3.51, a record, up 33% YoY
- Non-GAAP gross margin: 50.0%
Source: Applied Materials second-quarter 2026 Results
Dickerson guided semiconductor equipment business growth of more than 30% for calendar year 2026, up from prior commentary, according to the company statement.
The AI growth Applied had been preparing for is now “in full force,” CFO Brice Hill said, noting the company has increased its build plan, inventory positions, and logistics capacity to keep pace.
Related: Applied Digital adds $1B in market cap on major announcement
I also think the Micron partnership is worth watching closely. Applied and Micron Technology (MU) are jointly developing next-generation DRAM, HBM, and NAND solutions, combining Applied’s EPIC Center research capabilities in Silicon Valley with Micron’s innovation center in Boise, Idaho, according to the company statement.
That collaboration positions Applied directly inside the memory architecture redesign Citi is describing.
What AMAT’s valuation tells you about how much is already priced in
AMAT now trades at a price-to-earnings (P/E) ratio of 57 times, according to GuruFocus data. The stock is up 137.89% year to date and 252.87% over the past year, according to Yahoo Finance data as of June 17, 2026. The S&P 500 returned 9.75% and 25.58% over those same periods.
Citi also raised price targets on Lam Research (LRCX) and KLA Corp (KLAC) in the same note, framing this as a sector-wide call on NAND equipment demand, rather than an isolated Applied Materials story.
My own take, having watched this sector cycle through booms and corrections before, is that a P/E above 57 leaves very little room for disappointment. Citi’s thesis is well constructed and grounded in a real architectural shift in how AI systems handle memory.
But betting on three years of sustained 25%-plus WFE growth is still a bet on duration, and duration is the variable Wall Street has gotten wrong most often in this industry’s history.
