AMD vs. ASML: The $1 Trillion Semiconductor Cases, Compared

AMD and ASML are two very different ways to invest in the AI semiconductor buildout. AMD is trying to become a major alternative supplier of complete AI compute systems. ASML supplies the lithography equipment required to manufacture many of the world’s most advanced chips. That distinction matters: AMD’s route to $1 trillion depends on gaining share and delivering earnings; ASML’s depends more on preserving its technological bottleneck position while customers keep expanding advanced-chip capacity.

The trillion-dollar semiconductor benchmark

The number of semiconductor-linked companies above the $1 trillion line changes with daily market moves and with how broadly “semiconductor” is defined. Recent reporting identified five companies directly tied to the industry that had exceeded the threshold, while also noting that SK Hynix had subsequently fallen back below it. Nvidia, TSMC, Broadcom, Micron Technology and SK Hynix are the names commonly included in that broad framing. 21

The useful takeaway is not the exact club membership on a particular day. It is that the market has assigned trillion-dollar valuations to companies positioned at critical points in the AI infrastructure stack: accelerators, foundry capacity, networking and memory.

AMD: a bet on AI-system adoption and earnings delivery

AMD’s investment case is increasingly centered on data centers rather than its legacy PC-cycle exposure. In the second quarter of 2026, it reported $11.536 billion in total revenue and $6.7 billion in data-center revenue, up 107% year over year. 35 That growth shows that EPYC server processors and Instinct accelerators are already meaningful contributors, not merely future promises.

Why Helios is central to the thesis

Helios is AMD’s rack-scale AI offering, combining Instinct MI455X GPUs, sixth-generation EPYC CPUs, Pensando networking and ROCm software. AMD says the platform can deliver up to 30% more inference tokens per dollar than a leading competing platform; that is a company performance claim, not an independent benchmark. 46

The commercial question is whether customers convert platform evaluations and announced commitments into repeatable, high-volume deployments. Reporting has linked Helios activity to OpenAI, Anthropic, Meta and Microsoft, while Anthropic is reported to have committed to deploy up to 2 gigawatts of MI450-series GPUs. 41 If those deployments scale, AMD could capture revenue beyond standalone accelerators by selling a more integrated rack-level system.

AMD management has also framed the opportunity as exceptionally large: its CFO said the AI-chip total addressable market could reach $2 trillion to $3 trillion by 2030. That is management’s market estimate, not an industry consensus forecast. 36

What must happen for AMD to reach $1 trillion

A $1 trillion valuation requires more than another quarter of fast growth. AMD must demonstrate that Helios can scale economically, maintain competitive performance, secure supply, and build a software and customer-support experience that makes switching from Nvidia practical.

The valuation hurdle is substantial. One recent analysis used approximately 1.63 billion shares outstanding and estimated that a share price above $613 would place AMD above a $1 trillion market capitalization. 47 Another report cited expectations for earnings per share to rise from $2.65 to $15.61 and a roughly 66-times forward-earnings valuation. 37 Those estimates illustrate the core issue: the bull case relies on a major increase in earnings, not simply a higher valuation multiple.

AMD does not need to displace Nvidia to succeed. But it does need enough accelerator, inference, CPU and rack-scale-system share for its present growth to become durable earnings power.

ASML: a lithography moat with a timing question

ASML’s case is more structural. Its lithography systems are used to produce leading-edge logic and memory chips, putting the company upstream of many AI hardware beneficiaries. The key question is not whether AI infrastructure demand matters to ASML—it clearly does—but how quickly customers convert that demand into orders for its highest-value tools.

ASML reported €9.3 billion in second-quarter 2026 net sales and raised its full-year 2026 net-sales outlook to €43 billion to €45 billion. 1 Its own second-quarter release also outlined third-quarter net-sales guidance of €11 billion to €12 billion and a projected gross margin of 55% to 57%. 12

Capacity expansion supports the core thesis

ASML plans to add 30% to its 2026 low-NA EUV manufacturing capacity of roughly 65 systems in 2027, while investigating another 30% increase for 2028. It announced a similar plan for immersion DUV capacity, starting from about 130 systems in 2026. 12

That capacity plan is important because it is evidence of customer demand and a potential ceiling on future revenue. It also means the ASML thesis is not solely dependent on High-NA EUV: the company continues to benefit from demand for existing EUV and DUV lithography systems.

Analysts have responded positively. Morgan Stanley raised its price target to €1,660 and retained an overweight rating, citing greater confidence in EUV shipment capacity. Bernstein later raised its target to €2,500 and maintained an outperform rating after ASML’s second-quarter results, highlighting top-line, capacity and margin developments. 1411

High-NA EUV: proof point and constraint

Intel’s use of High-NA EUV for a subset of Core Ultra Series 3 processors is an important production-readiness milestone for ASML. 6 But High-NA adoption is not moving evenly across customers.

TSMC has said it has no current plans to use High-NA EUV in chip production through 2029, citing the cost of systems priced above €350 million apiece. 49 This does not erase ASML’s existing low-NA EUV or DUV opportunities. It does, however, reduce confidence that High-NA will be the near-term revenue and margin inflection that some investors expect.

Reports of more than one million High-NA EUV wafers processed with Intel were not substantiated by the stronger sources available here, so that figure should not be used as an investment premise.

Which company has the stronger $1 trillion case?

ASML has the more defensible business-positioning case. Its lithography franchise is embedded in advanced semiconductor manufacturing, and its raised outlook plus capacity plans point to strong current demand. The principal risks are the semiconductor capital-expenditure cycle, export controls, supply-chain execution and the investment decisions of a concentrated customer base.

AMD has the greater operating leverage—and the greater execution burden. Its data-center growth and Helios commitments establish a plausible path to a much larger AI business. But its valuation depends on customers translating commitments into sustained purchases and on AMD competing effectively against Nvidia’s hardware, software and ecosystem advantages.

Risks that could delay either outcome

• AI spending normalizes. A slowdown in hyperscaler or AI-lab capital expenditure would affect AMD’s system demand and ASML customers’ equipment budgets.
• AMD’s deployments do not scale. Announced commitments matter less than repeat orders, software adoption, margins and reliable supply.
• High-NA adoption takes longer. TSMC’s decision to hold off through 2029 highlights the cost sensitivity of ASML’s most advanced platform. 49
• Semiconductor cycles remain cyclical. Even strategically essential suppliers can face volatile customer spending, inventory corrections and delayed fab projects.
• Valuation expectations are high. Fast growth does not guarantee positive stock returns when future earnings assumptions are already demanding.

The clearest way to separate the two stories is simple: AMD is a high-growth AI-share-gain thesis, while ASML is a picks-and-shovels thesis built on lithography leadership. AMD could reach $1 trillion more quickly if Helios becomes a scaled AI-platform alternative. ASML may offer the more durable path if leading-edge logic and memory investment stays resilient—but a $1 trillion valuation by 2028 is still an outcome to be earned, not assumed.