TL;DR
· AMAT showcased DRAM and advanced packaging roadmaps at the June 25 masterclass, claiming its DRAM equipment market share has risen to No. 1 globally.
· Next-generation DRAM, HBM, and hybrid bonding require more deposition, etch, and metrology steps; AMAT covers 15 of 19 TSV process steps.
· Further share gains still depend on customer CapEx, 3D DRAM yield, and panel-level packaging ramp-up timeline.

Applied Materials (AMAT) highlighted its memory and packaging equipment roadmap for the AI era at the June 25 DRAM and Advanced Packaging Masterclass. Deutsche Bank reiterated a Buy rating, citing AI servers driving DRAM, HBM, and advanced packaging toward more complex manufacturing processes.

This is not just about new equipment launches. To investors, the key point is that the bottleneck in AI servers extends beyond GPUs themselves. The high-bandwidth memory, DRAM stacks, chip interconnects, and packaging substrates that work around GPUs are also becoming harder to manufacture. The more complex the process, the more steps equipment vendors can participate in, and the more CapEx may shift from "buying more wafer capacity" to "buying more sophisticated materials engineering equipment."

The most direct figure AMAT provided is that its DRAM market share has climbed from below 15% in 2013 to No. 1 globally today. The company aims to demonstrate that the AI memory upgrade is not a one-time demand but a systematic shift spanning DRAM transistors, interconnects, bonding, packaging, and metrology inspection.

AI Servers Don't Just Buy GPUs – Memory Manufacturing Is Also Getting Harder

Over the past two years, market attention on AI hardware has mostly focused on GPUs and advanced process nodes. But as large-scale cluster deployments go mainstream, memory bandwidth, power consumption, and packaging density have become key constraints on overall system performance.

HBM is a prime example. It boosts bandwidth by stacking multiple DRAM layers, but the higher the stack, the thinner the wafers, increasing difficulty in warpage, void filling, alignment, and defect detection during manufacturing. For equipment vendors, this means more deposition, etch, clean, polish, bonding, and metrology steps enter the serviceable addressable market.

What AMAT emphasized this time is not a single point tool but its coverage span. In the TSV process for HBM packaging, 15 of the 19 materials engineering steps can be covered by its products. TSV is the critical process of etching vertical through-silicon vias and filling them with metal to enable multi-layer interconnects, and it's a core link as HBM stacks move to higher layer counts.

Among the new tools, Avila 2 CVD addresses bowing issues in thinner HBM DRAM wafers; Nokota VMax 2 enables void-free filling of smaller TSVs; and OPTA Quad CMP brings chemical mechanical polishing in advanced packaging under real-time control. These products all point to the same thing: The AI memory upgrade does not drive a single equipment segment but simultaneously complicates multiple process steps.

Five Process Challenges in DRAM

AMAT summarizes next-generation DRAM changes into five key directions: EUV patterning, advanced transistors and interconnects, CMOS bonded array, 4F² vertical transistor DRAM, and 3D DRAM.

Behind these terms lie the real-world problems DRAM faces as it continues scaling down and improving performance. Traditional planar structures struggle to simultaneously meet density, power, and cost requirements. Memory makers need to introduce more complex patterning, finer interconnects, higher aspect ratio structures, and stacking/bonding approaches closer to those used in logic chips.

EUV patterning increases demands on high-precision etch; FinFET, copper interconnects, and epitaxy processes make transistor and interconnect steps more reliant on materials engineering; CMOS bonded array separates the array and peripheral logic, manufacturing them separately before bonding; and 4F² vertical transistors and 3D DRAM push further challenges toward high-aspect-ratio silicon channels, conductor etch, and e-beam metrology.

This is also why AMAT emphasizes its shift in DRAM market share. In 2013, the company's share in DRAM equipment was below 15%; today it claims to be No. 1 globally. If DRAM transitions from 2D scaling to more 3D structures and bonded architectures, its past coverage in deposition, etch, epitaxy, and metrology may continue to amplify.

However, this should not be simply interpreted as "DRAM technology upgrades equal an inevitable continued rise in AMAT's share." The pace at which memory makers adopt new structures, yield ramp speed, CapEx constraints, and the response from competitors in etch, deposition, and metrology will all determine actual order placement.

Packaging Moves from Silicon Interposer to Large Panels, Hybrid Bonding Takes Center Stage

Beyond DRAM itself, advanced packaging is the other main theme for AMAT.

Higher-density, lower-power interconnects are needed between AI accelerators and HBM, putting pressure on traditional silicon interposer solutions in terms of area and cost. AMAT's proposed direction is larger panel-level substrates, moving from 310×310mm and 510×515mm to 600×600mm.

The larger the panel, the greater potential benefit from larger single-processing area and lower packaging cost, but manufacturing difficulty also rises significantly. Deposition, etch, electroplating, planarization, and defect control on large-area substrates become harder to maintain uniformity. AMAT has invested in digital lithography, panel PVD/CVD/etch, and has complemented its large-area copper plating capabilities through the NEXX acquisition.

More attention is drawn to the Kinex hybrid bonding system. It integrates plasma surface activation, cleaning, bonding, and metrology, and AMAT calls it the industry's first integrated die-to-wafer hybrid bonding system. The value of hybrid bonding lies in making interconnects between chips denser, shorter, and more power-efficient, suited for future packaging roadmaps where high-bandwidth memory and logic chips are tightly integrated.

On process control, AMAT also launched VeritySEM 7AP and SEMVision G7AP for hybrid bonding pad, TSV, and microbump critical dimension measurement, as well as defect review and classification. Advanced packaging is evolving from "packaging chips" into a "high-precision process similar to front-end manufacturing," raising the importance of metrology and defect inspection.

Whether to Buy Still Depends on Customer CapEx and Yield

Deutsche Bank's positive view is based on a premise: that AI's demand for performance per watt will keep pushing up memory and packaging capital intensity, and AMAT's portfolio across deposition, etch, CMP, bonding, and metrology is comprehensive enough.

This explains why the market is willing to refocus on semiconductor equipment vendors like AMAT. If AI investment stays limited to GPU procurement, the beneficiary chain is relatively concentrated. But if HBM, DRAM, and advanced packaging all require new processes and new equipment, the equipment vendors' addressable scope expands significantly.

However, risks are equally clear. First, memory customers' CapEx is cyclical; strong AI demand does not necessarily mean DRAM makers will expand production indefinitely. Second, 3D DRAM, 4F² vertical transistors, and hybrid bonding all require yield verification; lab roadmaps are not the same as mass production. Third, while panel-level packaging has cost appeal, maintaining uniformity and defect control on large substrates remains an industry challenge.

AMAT has repositioned itself from "a DRAM equipment share chaser" to "a core equipment platform for AI memory upgrades." What will truly support this narrative going forward is not the number of product names, but how many new processes customers bring into production lines, and whether those processes genuinely generate more sustainable orders.

AMAT historical recommendation and target price change chart, showing stock price upside and rating adjustments from 2023 to 2026.

Click to learn about open positions at BlockBeats

Welcome to join the BlockBeats official community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Twitter Official Account: https://twitter.com/BlockBeatsAsia