3 AI Infrastructure Stocks In Focus As Anthropic Targets $1 Trillion Compute Spend

AI infrastructure just received a stress test and a spotlight at the same time, as Anthropic maps out over $1tn of potential computing spend and eyes a huge Nasdaq listing while running heavy losses and vast pre-committed capex. That mix of ambition and risk can reshape how investors think about who really benefits from AI and who might be overextended. This article walks through 3 stocks exposed to that story and how the news could matter for each one.

The three stocks below are just a small sample of AI infrastructure enablers, while the full screen surfaced 148 more large caps with similar AI aligned business descriptions and equally compelling narratives that are not covered here. To identify your own highest conviction ideas in this space, go straight to the AI Infrastructure Enablers (Semiconductors & Data Center Providers) screener.

ASE Technology Holding (TWSE:3711)

Overview: ASE Technology Holding provides semiconductor packaging, testing and electronic manufacturing services that connect AI chips to data centers and end devices.

Operations: ASE Technology Holding generates most revenue from packaging at NT$402.7b and EMS at NT$320.0b, with testing contributing NT$92.4b.

Market Cap: NT$2,964.0b

As AI spending shifts toward high end compute and data center hardware, ASE Technology Holding sits in the middle of that plumbing, turning bare chips into usable AI accelerators and server components for hyperscalers and cloud platforms.

"ASE Technology Holding is benefiting from strong demand for AI related advanced packaging and testing, with LEAP services revenue expected to exceed US$3.5b in 2026 and management aiming to double LEAP revenue again in 2027, which directly supports future revenue growth and earnings leverage."

What really matters from here is whether one pressure on ASE Technology Holding's capital spending translates into the margin profile investors are counting on.

That margin question is exactly what the full narrative for ASE Technology Holding unpacks in detail, showing where ASE Technology Holding’s AI packaging story could be accelerating or quietly stalling.

Semiconductor Manufacturing International (SEHK:981)

Overview: Semiconductor Manufacturing International produces integrated circuit wafers and compound semiconductors as a foundry partner for global chip designers and electronics companies.

Operations: Semiconductor Manufacturing International generates about US$10.4b from manufacturing and selling integrated circuits, with roughly US$9.2b coming from customers in China.

Market Cap: HK$562.5b

For the AI infrastructure theme, Semiconductor Manufacturing International matters because it runs large fabs that can turn rising compute demand into wafer orders across data-center, AI accelerator and peripheral chip customers.

"The ongoing push for semiconductor supply chain localization, with overseas clients shifting orders back to China and SMIC indicating that its capacity is still in shortage with shipments unable to fully meet customer needs, points to sustained high utilization that can support wafer pricing and gross margins."

What investors are really watching is how one unseen pressure on Semiconductor Manufacturing International's cost base shapes the durability of that margin story.

That unseen pressure is exactly what the full narrative for Semiconductor Manufacturing International unpacks, showing where cost strains may be masking accelerating AI demand and where SMIC’s story could be decoupling from peers.

SK hynix (KOSE:A000660)

Overview: SK hynix produces DRAM, NAND flash, SSDs and high bandwidth memory that feed AI GPUs, servers and global data centers.

Operations: SK hynix generates ₩269,336,675m from manufacturing and selling semiconductor products, primarily memory used across AI servers, PCs and mobile devices.

Market Cap: ₩1,072,890,016m

SK hynix matters in this AI infrastructure screen because its HBM and high performance memory help turn large compute budgets into workable training and inference capacity.

"Accelerating demand for high performance memory solutions in SK hynix products, particularly HBM and next generation DRAM used in AI servers and inference workloads, is supported by management’s outlook for DRAM demand growth of over 20% in 2026, which points to potential revenue and earnings expansion from higher content and premium pricing."

What could materially influence results is how one capital spending decision chain affects future pricing power and the resilience of AI related demand.

That capital chain is exactly what the full narrative for SK hynix unpacks, showing where SK hynix’s AI memory demand could be accelerating or masking bigger risks and upside.

Seeking Fresh Alternatives For Your Watchlist

New themes move fast, and early buyers often catch the breakout momentum while ideas are still flying under the radar for now. Before the next wave hits and these stories get fully priced, scan these fresh lists and consider them early.

• Spot cash generative compounders before the crowd by running the 193 high quality undervalued stocks while their quality, balance sheets and pricing power still look overlooked.
• Zero in on income ideas by checking the 221 dividend fortresses while yields remain elevated and payout strength still receives limited attention.
• Track structural electrification and grid upgrade themes by reviewing the 43 power grid technology and infrastructure stocks while valuations reflect only part of the broader build out story.

Have feedback on this article? Concerned about the content? Get in touch with us directly. Alternatively, email editorial-team@simplywallst.com