Copilot: TXN CC "mildly positive for ASML"
Me:
Please summarize TXN CC impact on ASML NXTi & I&M tools. Any EUV mention (I WAG it's coming)?
Copilot:
Short version: TXN’s call is mildly positive for ASML’s mature-node DUV fleet (NXTi + I&M), and there’s effectively zero read-through to EUV. Your WAG that EUV mention is “coming” is fair long-term, but it’s not in this call.
1. What TXN actually said about fabs and capex
- 300mm focus, mature nodes: TXN keeps hammering the same theme—45/65 nm-class analog on 300 mm as their cost moat. That’s squarely immersion DUV (NXTi class) territory, not EUV.
- Elevated capex, still on track: They reiterated the multi-year elevated capex cycle (roughly $5B/yr range) to build out 300 mm internal capacity (RFABs, LFAB, Sherman, etc.). That implies: ? more tools per fab, more metrology/inspection, more service over time, even if near-term utilization is soft.
- Underutilization now, but by design: They’re willing to run fabs underloaded near-term to secure long-term cost and supply advantage. That means: ? tool base is already largely in place, incremental adds are slower near-term, but the installed fleet is strategic and sticky.
2. Implications for ASML NXTi (ArFi DUV)
Net: slow but solid, structurally positive.
- Node class: TXN’s sweet spot (45/65 nm analog on 300 mm) is classic ArFi/ArF DUV, not EUV. That’s ASML’s NXTi lane.
- Capex posture: They’re not slamming the brakes—just pacing spend while demand normalizes. That supports:
- Sustained DUV demand for capacity adds and line balancing
- Field upgrades (productivity, overlay, matching) rather than huge new-fab waves every quarter
- Cycle tone: End-markets are mixed, but data center and industrial are stabilizing/improving, which supports eventual higher fab loading and, later, more tools.
So for NXTi: No fireworks, but the thesis holds—TXN is building a long-lived, DUV-heavy 300 mm fleet and isn’t backing away.
3. Implications for I&M (inspection & metrology)
Quiet in the transcript, but structurally good.
- 300 mm analog = yield and matching game: When you’re running big 300 mm analog fabs at mature nodes, the differentiator is yield, parametric stability, and cross-tool matching, not bleeding-edge CD. That leans into:
- Overlay and CD metrology
- Defect inspection
- Tool-to-tool matching and fleet tuning
- Underutilization now, but more wafers later: As utilization rises, the value of I&M uptime and accuracy goes up. TXN’s whole strategy is “own the capacity, then fill it”—which is exactly the environment where I&M spend and service intensity creep higher over time.
So for I&M: Call doesn’t name it, but the long-term installed-base story is quietly strong.
4. Any EUV mention?
- Direct answer: No meaningful EUV signal in this call. TXN is still talking 300 mm, cost, CHIPS incentives, depreciation, and utilization—not leading-edge logic or DRAM nodes that would require EUV.
- Your WAG (“it’s coming”) Long-term, if TXN ever pushes into much finer geometries for specific mixed-signal or RF, you could see foundry EUV in their supply chain—but not as a direct TI-owned EUV fleet. For now, their whole strategy is:
- Internal 300 mm, mature nodes
- Cost and control, not node-race
So: no EUV yet, and no hint they’re about to flip that switch.
5. Bottom line for ASML from this TXN call
- NXTi / DUV:
- Supported by TXN’s 300 mm, 45/65 nm strategy
- Near-term: flattish to modest, with underutilization
- Long-term: structurally locked in as TXN’s cost moat
- I&M:
- Quiet in words, important in physics
- Yield, matching, and uptime become more valuable as fabs fill
- EUV:
- Not in play for TXN’s own fabs
- Any EUV exposure is indirect via foundries, not capex on TXN’s balance sheet
If you want, I can stack this next to Hynix/Samsung/TSMC to show how “TXN = DUV floor” vs “Korea/Taiwan = EUV ceiling” in one ASCII block. |
