ASML

Current Thesis

The EUV-monopoly story is still the cleanest second-derivative play on the AI-capex super-cycle, and the narrative has gotten louder, not quieter, since the 2026-04-15 beat-and-raise. CEO Fouquet now frames chips as a "supply-limited market for quite a while" with "sporadic bottlenecks" (2026-05-20) — that is demand-visibility language, not cycle-top language. High-NA is reportedly sampling first product data on both logic and DRAM later this year (2026-05-20), which widens the TAM into memory and adds an advanced-packaging tool adjacency (the CoWoS chokepoint). Sell-side keeps climbing the wall: RBC $1,700 (2026-04-16) was topped by JP Morgan Overweight / $2,200 (2026-06-03). Institutions built in Q1 (Third Point new 12,000 sh, Appaloosa new 49,500 sh, 2026-05-15). So the narrative is ACCELERATING — but the setup is MATURING and crowded. We have DEFERred four-plus times in a row for one reason: this is archetype-2 picks-and-shovels, +41% YTD into the April print, extended above its MAs, and the entry gate (RVOL volume absorption on a pullback) has never confirmed — last read RVOL 1.19x (2026-05-14). The discipline holds: buy the pullback, don't chase the vertical into a Q2 print that guides sequentially down ~13–18%.

Bull Case

• CEO "supply-limited market" framing (2026-05-20): explicit multi-quarter demand-visibility language plus an Elon Musk chip-supply conversation anecdote — narrative still pulling forward, not rolling over.
• High-NA on logic AND DRAM later this year (2026-05-20): extends the monopoly TAM into memory; a new packaging tool layers advanced-packaging (CoWoS-bottleneck) optionality on top of the EUV core.
• JP Morgan Overweight, PT raised to $2,200 (2026-06-03): sell-side targets still climbing two months post-print (RBC $1,700 4/16 → JPM $2,200 6/3) — narrative velocity intact, upgrades clustering.
• FY2026 guide €36B–€40B, +17% implied accel (2026-04-15): raised on TSMC capex +32% YoY (~$55B) and hyperscaler 2nm pull-in; TSM ≈40% of ASML revenue = direct flow-through.
• Smart-money 13F builds (2026-05-15): Third Point new 12,000-sh stake, Appaloosa new 49,500-sh stake — discretionary funds adding the Q1 dip, not distributing.
• China chip exports doubled to $31B (2026-05-22): confirms end-demand for AI silicon is real and supply-tight, supporting the leading-edge tool pull globally.
• Management called 2027 a "growth year" (2026-04-15) despite Q2 softness — signals the Q2 dip is shipment-timing, not demand destruction.

Bear Case

• Q2 guide €8.4B–€9.0B = sequential −13% to −18% (2026-04-15): lumpy shipment cadence into a headline-risk print ~2026-07-16. Crowded long into a soft sequential print is the textbook fade.
• Nikon CEO says it will challenge ASML with low-priced gear (2026-05-28): it's DUV/low-end diversification away from Intel reliance — NOT EUV — so the leading-edge monopoly is intact, but it is a fresh competitive headline that can dent the "uncontested sole-source" frame if it gains any traction.
• China domestic-substitution accelerating (2026-05-22): US restrictions are explicitly "fueling Beijing's push toward domestic chip production." ~15% of revenue is China-exposed; this is the slow structural TAM erosion under the AI-demand noise.
• Fed-hike macro repricing (2026-05-15): CME FedWatch ~56% odds of a hike by Dec 2026; "bond market fired a warning shot at the AI rally." High-multiple semicap is rate-sensitive — multiple-compression tail.
• Trade-war chop (2026-05-15): chip stocks tumbled on Trump-Xi summit tension; export-control re-rating is binary-negative and instant.
• 13F signal is mixed (2026-05-15): one Appaloosa headline reads "Sold Entire Stake In ASML," another "Takes New Stake In ASML 49,500 sh" — net-additive read, but the ambiguity flags that institutional positioning isn't one-directional.
• Extension persists: +41% YTD into the April print, RVOL 1.19x (no thrust) on 2026-05-14 — the pullback-with-volume gate the dossier requires has not printed.

Setup & Price Structure

No live price context this cycle, so read it structurally. ASML ran +41% YTD into the 2026-04-15 beat and never gave back a clean pullback we could buy — every DEFER (4/23, 4/27, 5/14, 5/15) failed on the same two checks: edge vs SPX turned negative/extended above MA, and Rule 4 RVOL never expanded (1.19x on 5/14 vs the >1.5 thrust the entry gate wants). The TSMC High-NA snub on 4/24 wiped ~$16B and proved the tape is two-way. Since then the structure has firmed (CEO supply-limited comments 5/20, JPM $2,200 6/3, 13F builds 5/15) but the stock is still pinned near the upper rail with sell-side targets above spot — i.e., we'd be paying up into resistance with no volume confirmation. This is the actionable map: a2 entry = pullback to the 20-/50-DMA WITH a volume-absorption print (RVOL > 1.5) → that flips this to APPROVE / HIGH. The a2 scale-out line (if long) is a weekly close below the 20-week EMA = trim 50% — NOT RSI > 75 (that's the a6 rule, does not apply here). Until the pullback prints, idle is correct, but a clean retest is a fat pitch given the accelerating narrative + climbing targets.

Catalyst Calendar (next 30 days)

• ~2026-06-10 (est.) — TSMC May monthly revenue print (released ~10th). ASML's largest customer (~40% of rev); a strong May read = direct demand confirmation, a soft read = the first crack tell.
• Ongoing, June — Fed commentary / FedWatch repricing into the next FOMC; high-multiple semicap is the rate-beta expression of the "bond warning shot" (2026-05-15). Macro-tightening is a live DEFER input.
• H2 2026, no fixed date — High-NA "first product data on logic and DRAM later this year" (per 2026-05-20). No catalyst date inside the 30-day window; watch for the order/data headline.
• NEXT HARD BINARY — ~2026-07-16 (Q2 2026 earnings, OUTSIDE 30d window): the lumpy sequential step-down print. Blackout: flatten exposure 3 trading days prior (~2026-07-13). This is the binary that defines sizing risk; do not be long into it on a non-earnings thesis.

What Would Change Our Mind

• Flips us to APPROVE (up to HIGH): a pullback to the 20-/50-DMA with RVOL > 1.5 volume absorption; OR an export-control relief headline; OR a confirmed High-NA DRAM order win that opens the memory TAM.
• Invalidates the long (avoid / scale out): a weekly close below the 20-week EMA; OR FY2026 guide cut below the €36B floor at the 2026-07-16 print; OR a fresh US/NL EUV/DUV restriction on China (binary gap-down through the 50-DMA before we can react).
• Breaks the monopoly thesis outright: Nikon (2026-05-28) or a Chinese domestic tool lands a leading-edge (EUV-tier) customer — that would be a structural re-rate, not a dip to buy.

Correlation Notes

• Do NOT stack with TSM + AMAT/KLA/LRCX concurrently — same semicap/AI-capex thesis, correlation > 0.7. ASML is the highest-quality expression (sole-source EUV moat) but it is not diversification if we're already long the cluster. The dossier explicitly forbids the TSM+ASML double-up.
• Demand pull-through, not hedge: NVDA / AVGO / hyperscaler capex (Broadcom–Meta ASIC deal, 2026-04-20) is the upstream driver — positively correlated, second-derivative.
• Memory divergence: MU/HBM capex is NOT uniformly correlated (Micron's 4/16 selloff then 4/21 rip showed the decoupling); High-NA-on-DRAM is the bridge that would tighten that correlation.
• Macro beta: rate-sensitive, high-multiple — moves with bond-yield/Fed-hike repricing (2026-05-15), and gaps on US-China trade headlines. Treat export-control and FedWatch as the two exogenous risk factors that override the chart.