Research Log
Article
Research Log
Research Log — 2026-05-01
| Area | Status | Headline |
|---|---|---|
| Deep Dive | 🟢 | BESI / BESIY (BE Semiconductor) NEW — paradigm-validation pure-play. Three independent paradigms (autonomous trace, NuttyCLD Part 4 physics-down, picks-and-shovels conviction) converged on the same name — highest cross-thesis convergence in our entire system right now. $293.04 RSI 70-77, 30D +31.8%, 3M +51.8%, 12M +161%, -2% from 52wk high. Q4-25 orders +43% QoQ / +105% YoY (€250.4M); FY-25 orders €685M (+17% YoY); Q1-26 guide revenue +5-15% QoQ at GM 63-65%. SK hynix Kinex order March 2026 = first mass-production hybrid-bonding HBM order, ~$15M per system. Twin-ramp tailwind: HBM stacking (D2W, BESI leads ~85% share) AND BSPDN carrier bonding (W2W, EVG/TEL lead) — same toolkit family, TAM doubles without share gains. AMAT 9% strategic stake + Kinex co-development is structural. Founder-CEO Richard Blickman (1995, 30+ years). Bear discipline: forward P/E ~54x (94th percentile IT), Dutch listing premium, thin US ADR float, TSMC A16+SPR ramp slipped 2H-26 → 2027, ASMPT 2nd-gen shipping Q3 2026 (only credible D2W challenger). RSI 77 + 32% 30D = wait for $260-280 pullback before chasing. Stop $234, target $381. Watchlist symbol updated BESI (delisted) → BESIY (cache fix). → 2026-05-01-BESI.md |
| Deep Dive | 🟡 | NVTS (Navitas) NEW — $17.45 RSI 73, 30D +114.9%, 12M +751%, -11.8% from 52wk high. The small-cap GaN density bet that the 2026-05-02 blind agent-trace also surfaced (sole NuttyCLD-cohort name independently flagged — calibration win). 800V→6V single-stage GaN demoed at NVIDIA GTC 2026 at vendor-targeted ~96.5% peak efficiency; on the 14-vendor 800VDC Kyber partner list. But: $3.8B cap on $46M TTM revenue (P/S 83x), –111% FCF margin, 28% short interest, 1 buy / 63 sells insider net –$185.8M, analyst avg PT $6.68 (-62%). Q1 2026 earnings May 5 is binary squeeze event. CEO Allexandre (Sep 2025 ex-Renesas Power, led Transphorm GaN acq) replaces co-founder Sheridan (now at Empower Semi competitor). Verdict: 🔍 WATCH into earnings — do not chase RSI 73; post-earnings starter at $13-14 if thesis holds. Stop $11, target $22. → 2026-05-01-NVTS.md |
| Deep Dive | 🟡 | PLTR refresh — $144.07 RSI 50.6, -30.6% from 52wk high. Held in burry-13f mirror at -16.76% (-$11,065), the only losing 13F mirror. Editorial frame: "Burry was right and just early, OR wrong on AI software multiples." Q1 earnings May 4 is the binary catalyst — need US Commercial +130%+ to defend the narrative. Stop tightened to $115, target $187. P/S compressed 112x → 75x already. Verdict: HOLD (paper), do not add real book pre-earnings. → 2026-05-01-PLTR.md |
| Deep Dive | 🟡 | IFNNY (Infineon) NEW — $67.25, RSI 87.7, 30D +49.0%, 12M +98.8%, near 52wk high. Blackwell PMIC winner with Delta on Vertical Power Delivery (~60-70% GB200 share); FY27 €2.5B data-center target (€1.5B FY26). 300mm GaN customer samples Q4 2025 (TSMC exited GaN). EV-AI demand crossover thesis. Verdict: WAIT for $56-60 pullback or buy IFX on Frankfurt — overbought today; ADR liquidity is the silent tax. Stop $53.80, target $87.43. Closes one of nine NuttyCLD VRM-cohort deep dives queued in TASKS.md. → 2026-05-01-IFNNY.md |
| Deep Dive | 🟡 | MPWR (Monolithic Power) NEW — $1,583 RSI 75, 30D +42%, 1Y +150%, -4.7% from ATH. NuttyCLD AI Power Crisis protagonist — projected ~70% Vera Rubin VRM socket share at +60% ASP; KeyBanc PT $1,250 (already through). Q1 2026 beat ($804M vs $782M, EPS $5.10 vs $4.90); enterprise-data guide raised 50%→85% YoY. Sells the whole intersection (controller + DrMOS + module + qualification) — that's the +60% ASP. The "scarred and back with pricing power" arc fully priced. Forward P/E 52x vs 5y avg 45x. Verdict: 🔍 WATCH — entry on RSI <55 pullback to $1,300 zone; asymmetry now in peers (AOSL, VICR, NVTS, MRAAY). Stop $1,267, target $2,058. Coverage gap — open vrm-pmic watchlist. → 2026-05-01-MPWR.md |
| Deep Dive | 🟡 | VICR (Vicor) NEW — $268.36, RSI 80.6, 30D +81.0%, 12M +566.6%, -8.7% from 52wk high. The 25-year curse cracked. Q4 2025 set up the inflection (record $176.9M backlog, BTB >1.2x, Phil Davies $200-$400/XPU framing); Q1 2026 (Apr 21) detonated it — backlog $300.6M (+70% sequential), BTB >2.0x, FY 2026 guide ~$570M, Roth raised PT to $245. Founder Vinciarelli (1981) is the original FPA architect — VICR was first technically, lost first commercially in Blackwell to IFNNY+Delta; Gen 5 VPD into production with lead customer Q1, broader Gen 4→5 transition 2H 2026. Bear discipline: $218.8M insider net selling, 25 years of "almost," Vinciarelli is famously commercial-allergic. Verdict: 🔍 WATCH — entry zone $215-$245 on digestion, RSI 80 too hot to chase, stop $214, target $349. Closes one of nine NuttyCLD VRM-cohort deep dives. → 2026-05-01-VICR.md |
| Deep Dive | 🟡 | AOSL (Alpha & Omega Semiconductor) NEW — $43.08, RSI 78.3, 30D +95.7%, 3M +92.7%, -6.3% from 52wk high. The $900M-cap David that beat MPWR to GB300 by being first to OpenVReg. OVR4-22 Dec 2024 (world-first), OpenVReg16 Jan 2025 (industry-first 16-phase production), official 14-vendor 800VDC NVIDIA partner Oct 2025; Ming-Chi Kuo confirmed primary GB300 NVL72 DrMOS supplier at ~70% share. Unusual vertical integration for size: SiC + GaN + multi-phase controllers + DrMOS in-house, R&D 14-15% of sales, zero long-term debt. Founder Mike Chang (Exec Chairman, 10% owner) bought 33,727 shares ($1.5M) March 9, 2026 into the breakout — strongest insider signal in semis right now. Bear discipline: Q2 FY26 guide $160M (-12% sequential), revenue accel -14.5%, YoY -6.3%, MPWR projected ~70% Vera Rubin share leaves AOSL with little next-gen room, China supply-chain exposure ~50%+. RSI 78 + 96% in 30D = parabolic blow-off. Verdict: 🔍 WATCH / DO NOT CHASE — entry zone $34-$37 on first pullback to SMA20 area, stop $34.50, target $56. Closes one of nine NuttyCLD VRM-cohort deep dives queued in TASKS.md. Coverage-gap process improvement: cross-reference NVIDIA partner lists against our watchlists every quarter. → 2026-05-01-AOSL.md |
| Deep Dive | 🟡 | POWI (Power Integrations) NEW — $73.32, RSI 75, 30D +40.0%, 3M +58.3%, -3.8% from 52wk high. The IP/topology rentier — owns the road design, not the socket. Only volume supplier of 1,250 V / 1,700 V GaN switches (175 M+ deployed) and the most radical of NuttyCLD's 4 topology buckets — direct 800 V→12 V (1,250 V PowiGaN half-bridge) and 1,000 VDC→12 V (1,700 V InnoMux-2 at >90.3% efficiency). Established floor under the radical thesis: $443 M FY25 revenue (+6%), 21% FCF margin, dividend-paying ($0.215/Q), conservative balance sheet ($99 M liabilities vs. $772 M assets), 100% earnings beat rate. PowiGaN +40% FY25 but small base; industrial +15% is today's actual driver. Critical: sell-side consensus PT $52-53 is below spot $73 — market is ahead of analysts. Q1 2026 earnings May 7 is the binary. Bear: insider selling -$14.1M YTD (0 buys / 58 sells), CEO Balu retirement during AI inflection (no successor named yet, new CFO Nancy Erba 6 months in), Vicor 25-year curse cautionary tale, ~75% China revenue. Verdict: 🔍 WATCH — entry zone $60-$70 (preferred <$65 on post-earnings flush), stop $58.65, target $95.30. Basket allocation alongside MPWR/AOSL/VICR/NVTS, not standalone alpha. Closes one of nine NuttyCLD VRM-cohort deep dives. → 2026-05-01-POWI.md |
Related
5 eventsNo direct external sources are attached to this read.