Lam Research's comments about 2027 supply constraints bode well for semicap subsystem suppliers Ultra Clean, MKS Instruments and Ichor, which have substantial cyclical torque. The three trade at 7-12x estimated 2027 EPS. x.com 

Low multiples give AI infrastructure good risk/reward as demand outruns supply and agentic workloads have runway. TSMC at 17x 2027E EPS appears to have all the accelerator and CPU demand it can handle this year and next; Ultra Clean is at 8x, MKS 12x, TTM 13x, Nvidia 14x and Celestica 15x. Kioxia at 3x and Samsung preferreds at 2.5x carry a big discount on pricing and cycle fears. Multiples imply a 2027 capex peak and meaningful drop; 2028 capex growth is expected, but below 2027's rate. x.com  … Show more

ASML may have topped, since semicap thrives on shortage and would be hit hardest in an overbuild. Then earnings acceleration from memory and fab projects stops or reverses; the labs' leading-edge chase may be at its apex, positioning is bullish and the moat is threatened. Bullishness on Korean, US and Taiwan fab projects makes sense for now. But if the cycle top is confirmed and labs change how they chase the frontier, or token deflation kills some labs, the shortage can end sooner. x.com  x.com  … Show more

Semiconductors could stay undersupplied for a long time as more capable AI per dollar creates more uses and drives compute demand. Micron and NVIDIA trade at low forward P/Es because investors fear peak cyclical earnings, yet current earnings show no slowdown. Shortages may worsen as AI adoption spreads, though the industry could remain cyclical; a reversal of the forced rollover trade is hoped for, not assured. x.com  x.com  x.com 

Semiconductors could capture scarce spot compute's value if its forward curve holds. Compute trades in deep contango, while chip stocks price in little of that value; near-term moves are unpredictable. That makes the sector very bullish. The earlier lab-funding risk looks less acute: Anthropic should hold nearly $200B cash after its IPO and is operating-cash-flow profitable. Even with 2029 capex cuts, chips look cheap if future buyers do not default. x.com  x.com  x.com  x.com  x.com 

Financing strain threatens smaller AI-related plays. Elon Musk says debt markets have nearly closed to firms without substantial non-AI-data-center revenue. Neoclouds, energy, HVAC, electronics and builders will be a much smaller field in a year; its 'guppy' plays can be shorted here toward zero. CoreWeave, IREN and Vertiv look troubled, perhaps Nebius; Ouster, Applied Digital, Ultra Clean, Cipher Digital, WhiteFiber, TeraWulf and IPG Photonics too. x.com  x.com 

SanDisk CEO David Goeckeler's report of customer signals beyond fiscal 2030 adds storage to the long-duration AI hardware-demand case. He says those signals come through new-business-model partnerships; Micron and SK hynix in memory and Lumentum in photonics were already cited as signs of demand beyond 2030. x.com 

ASML's planned 2027-28 output cannot meet customer demand, supporting a Buy rating and €1,830 target. Long-term customer down payments rose from €3.6bn to €9.3bn; customers seek 2029-30 orders that ASML has not accepted. Critical-tool price rises begin in 2027, with full effect in 2028, but are not meant to recur annually. Q3 EPS excluding a €509.7m investment gain still beat consensus. x.com  x.com  x.com  x.com  x.com 

Excess data-center capacity makes Applied Materials look cooked because it is mainly tied to DRAM-manufacturing equipment. x.com 

ASML's upgrades and parts capture the chip shortage before its long-lead-time systems do. Q3 installed-base revenue reached €3.3B versus ~€2.9B guided, a ~€0.4B beat after Q2's ~€0.3B beat. System prices are locked in years ahead; upgrades add wafers per hour to tools already running while chipmakers wait for new machines into 2028. The Elec reportedly says Samsung and SK hynix accepted a 10% parts increase from January, though that report was unconfirmed. x.com