Skip to content

Applied Materials ​

ASML 负责“画图纸”(光刻):ASML 拥有极紫外光(EUV)和深紫外光(DUV)光刻机的绝对垄断地位。它的作用是利用光线,将芯片设计的电路图案“投影”到涂有光刻胶的硅片上。

AMAT 负责“盖房子”(沉积、刻蚀、改性):一旦 ASML 投射出了图案,就需要 AMAT 的设备来进行物理和化学操作。AMAT 提供最广泛的产品线,包括把材料铺上去(CVD/PVD 沉积)、把多余材料挖掉(刻蚀)、把表面磨平(CMP)以及注入离子改变导电性。

ASML 的 EUV 曝光线宽越来越细(进入 3nm、2nm 节点),光刻胶边缘粗糙度(LER)和随机缺陷等问题极其严重。ASML 必须依赖 AMAT 提供极其精密的底材沉积、硬掩模版(Hardmask)材料和选择性刻蚀设备,才能确保 EUV 印出来的图案在后续加工中不会坍塌或变形。双方在逻辑芯片的 GAA(环绕栅极晶体管)和 DRAM 制程中是深度绑定的。

双重曝光 vs Centura Sculpta(图案成型系统) ​

随着制程缩小,单次 EUV 曝光已经不够用了,晶圆厂原本需要购买更多的 EUV 光刻机来进行EUV 双重曝光(Double Patterning),这会让晶圆厂成本剧增。

AMAT 推出了名为 Centura Sculpta 的设备。它可以在 ASML 完成一次 EUV 曝光后,利用特殊的定向刻蚀技术,把原有的图案按照特定方向“拉长”或“塑形”,从而达到双重曝光的效果。

台积电一直对 ASML High-NA EUV 的经济性持谨慎态度,认为其成本过于高昂。台积电明确表示,在 2nm(N2)甚至接下来的 A16(1.6nm)节点,都不会把 High-NA EUV 作为主力大规模使用。

为了在不买 High-NA 的情况下实现 2nm 的线宽,台积电大量采用了现有的标准 EUV 光刻机,并结合 AMAT 的 Centura Sculpta 图案成型系统。

Sculpta 帮助台积电在某些关键层(如通孔和金属走线)上只需进行一次光刻,然后通过物理/化学“定向拉伸”成型,省去了耗时且昂贵的“光刻-刻蚀-光刻-刻蚀(LELE)”循环。这为台积电节省了巨额的资本开支。

与台积电类似,三星在其先进制程(SF2 和 SF1.4 节点)的初期产能爬坡中,也感受到了巨大的成本压力。三星同样引入了 AMAT 的 Sculpta 技术,来优化标准 EUV 的使用效率。

由于存储芯片对成本极为敏感,三星更是明确表示 High-NA EUV 在近期内不适合 DRAM 的大规模制造,因此 AMAT 的材料工程和刻蚀优化方案在三星内部占据了重要的战略地位。

Intel 是全球第一个下单并接收 ASML High-NA EUV 光刻机的厂商,试图以此夺回制程领先地位。但即便如此激进,Intel 在其 20A 和 18A 节点中,也同时大量引进了 AMAT 的 Centura Sculpta,因为单纯依靠光刻微缩的成本实在太高,Intel 必须用 AMAT 的设备来对冲 ASML 带来的成本飙升。

在这场博弈中,AMAT 的 Centura Sculpta 目前取得了阶段性的商业胜利。 台积电和三星都选择用 AMAT 的这套方案来最大化现有标准 EUV 光刻机的投资回报。而 ASML 的 High-NA EUV 虽然代表了人类光学的物理极限,但受限于极高的设备和运行成本,目前只能在晶圆厂的研发线和少数最尖端的层级中缓慢渗透,其真正的大规模量产爆发被台积电和三星向后推迟了。

High-NA EUV vs 先进封装 (Chiplet) ​

ASML大力推销造价高达 3.5 亿至 4 亿美元的 High-NA EUV(高数值孔径 EUV)光刻机。ASML 认为,继续把单块芯片(Monolithic)的晶体管做小才是王道。

AMAT 的叙事:单靠光刻微缩太贵且良率难以控制,未来的重心在先进封装和小芯片(Chiplet)架构。AMAT 正大力推广通过晶圆级封装、混合键合(Hybrid Bonding)、硅通孔(TSV)来将多块成熟芯片拼装在一起。

在先进封装流程中,光刻的需求相对降低(多使用成熟制程光刻机),而对 AMAT 的沉积、电镀、CMP、电子束量测需求呈指数级上升。AMAT 试图用“材料工程+封装”来替代“极致光刻”的价值。

Applied Materials Investor Relations ​

Applied Materials Quarterly Earnings Calls ​

Q3 2026 ​

The company generated record cash from operations of $3.04 billion and distributed $860 million to shareholders through $440 million in share repurchases and $420 million in dividends.

“Applied Materials achieved its 13th consecutive quarter of year-over-year gross margin expansion, demonstrating the increasing value we create by enabling better chips, systems and fab returns,” said Brice Hill, Senior Vice President and CFO. “We expect continued strong revenue growth in the second half of the calendar year, particularly in DRAM as well as leading-edge foundry-logic and advanced packaging. Looking further ahead, we are making additional manufacturing capacity investments to support projected demand through the end of the decade.”

Announced a long-term joint development agreement with EssilorLuxottica to accelerate the commercialization of next-generation intelligent optical systems for augmented reality and AI-powered smart eyewear.

To ensure our supply chain and field teams can support their ramps, our largest customers are giving us longer-term commitments and rolling eight-quarter forecasts. This increased demand visibility, gives us high confidence that 2027 will be another strong growth year for Applied Materials. As customers move quickly to bring new fab capacity online, while simultaneously optimizing yield and output of their existing production facilities, we also see strong incremental demand for our advanced service solutions.

The first race is for technology leadership. AI datacenter returns are determined by the number of tokens generated per second and the total cost of ownership-which is dominated by energy consumption. Improvements in'Tokens per second per watt' are primarily driven by innovations in semiconductor devices and systems. This can be seen in the value the semiconductor and semiconductor equipment industries are generating from their most advanced technologies.

The second race is for capacity, as demand for advanced semiconductors to support AI infrastructure scaling far exceeds supply. As a result, chipmakers are intensely focused on increasing output and yields of their existing factories, while rapidly building new ones.

In the race for technology leadership, leading-edge foundry - logic, DRAM, and advanced packaging have the greatest impact on AI computing performance, power efficiency and cost. Together, we expect these areas to represent around 80% of wafer fab equipment growth in 2026 and 2027.

Advanced packaging is one of the most important areas for AI compute innovation and we see very strong multi-year growth for Applied. Applied is the overall leader in this market, with strong positions in high-bandwidth memory and 3D chiplet-stacking, and we now expect our overall packaging revenues to grow more than 70% in calendar 2026.

We are also well positioned for future packaging inflections, as the industry moves to new architectures and larger-size panel formats. We have built a broad portfolio of next-generation technologies for panel, including digital lithography, deposition, etch, and eBeam review.

In the global race to add semiconductor manufacturing capacity, our customers' ability to increase yield ind output in their existing production fabs is incredibly valuable. This creates expanded opportunities or Applied to deliver new innovations in three key areas: services, process diagnostics and control, and new products that increase wafer output per area of fab space.

The fastest growing parts of the market are the leading edge foundry logic, DRAM, and advanced packaging. We said that is ~80% of the growth in wafer fab equipment spending this year.

Q2 2026 ​

Q1 2026 ​

Q4 2025 ​

Q3 2025 ​

Q2 2025 ​

Q1 2025 ​

Q4 2024 ​

Q3 2024 ​

Q2 2024 ​

Q1 2024 ​

Q4 2023 ​

Q3 2023 ​

Q2 2023 ​

Q1 2023 ​

Q4 2022 ​

Q3 2022 ​

Q2 2022 ​

Q1 2022 ​