Skip to content

Dossier · ON · Dormant

ON

Last analysed · · source: watchlist_research

Current thesis

Dead EV-SiC name re-rating as an AI-datacenter-power play: Q1 (5/4) AI data-center revenue +30% QoQ, NVIDIA 800V/MGX role expanded 5/29, Mizuho PT→$150 (5/28). Theme ACCELERATING, cluster-confirmed (Navitas +346% YTD); price below PTs with no earnings in 30d. Entering post-inflection-print, so HIGH not SUPREME.

Invalidation trigger

Weekly close below 20-week EMA (~$112, est.); or Q2 print (~late-July) shows AI data-center revenue flat/down QoQ after Q1's +30% sequential; or NVIDIA awards the 800V MGX power sockets to Infineon/Navitas instead of onsemi.

Thesis status

Open commitment scored if the trigger above fires How this is scored →

Current Thesis

The dead EV-SiC name has a new growth leg: AI data-center power. The narrative we're buying is the 800VDC rack transition — as NVIDIA racks move from 54V toward 800V/1MW+, onsemi's silicon (Si/SiC/GaN) content per rack steps up sharply, and the segment is now inflecting off a small base. This is a 2nd-order AI picks-and-shovels re-rating, ACCELERATING and cluster-confirmed, not a value story. We are entering after the May 4 inflection print, so the easy leg is partly gone — but sell-side is still catching up and price sits below the newest PTs.

Bull Case

  • Q1 2026 print (reported 2026-05-04) was the inflection: revenue $1,513M beat the guidance midpoint; AI data-center business grew >30% QoQ; non-GAAP gross margin expanded to 38.5%; non-GAAP EPS $0.64; non-GAAP operating margin 19.1%. The trough is turning.
  • NVIDIA design-win is confirmed, not speculative: onsemi formally collaborated with NVIDIA on 800VDC power and expanded its NVIDIA MGX ecosystem role (announced 2026-05-29) — power FETs, multi-phase, SiC JFETs, GaN already deployed in MGX systems.
  • Content-per-rack math is mispriced: the 2026-04-20 "Wolf Pick" framed each 800V rack NVIDIA ships in 2027 as worth ~11x the onsemi silicon of today's 54V racks — and noted consensus hasn't put that into the model yet.
  • Sell-side catching up = narrative going mainstream: Mizuho Outperform, PT→$150 (2026-05-28, from [entry redacted]) on AI server demand; B of A Buy, PT→$138 (2026-05-26). Price ~$132.95 (6/3) is still below both PTs = headroom.
  • Cluster confirmation: Navitas (GaN pure-play) +346% YTD (2026-06-04); Infineon/STM/NXP/Microchip in the same 800V-power basket; new leveraged ETFs launched 2026-06-01→06-03 targeting ON/NXP/Microchip = structural retail flow.
  • Capital return at the trough: $346M buyback in Q1 (~160% of FCF) — management leaning in.

Bear Case

  • Rich and extended: P/E ~85, +184% YoY, fresh all-time high ~$129 (2026-06-02). Priced for the AI ramp to materialize; any stumble = sharp de-rate.
  • Core is still in a trough: Q1 GAAP operating margin (3.5)% and GAAP diluted loss of ($0.08). The legacy auto/industrial majority of revenue is still soft; AI data-center is a fast-growing but small slice. Blended story depends on legacy not deteriorating further.
  • Crowded socket: Infineon, STM, Navitas (GaN), Vicor all chasing the same 800V MGX power sockets — design-win share is not locked; commoditization risk on the power FET layer.
  • Late-cycle tells: leveraged ETFs being built around the theme (6/1–6/3) and SOXX/SMH record April inflows = retail going public. A broad semi pullback hits a $57B name at ATH hard.
  • No near-term fuel: next hard catalyst (Q2 print) is ~7 weeks out; momentum can stall in a tape rotation with nothing to defend the multiple.

Setup & Price Structure

  • Last ~$132.95 (2026-06-03), intraday range $126.15–$134.92; broke prior ATH (~$129.17 / close $128.64 on 6/2) → clean breakout, momentum intact.
  • +184% over trailing 12 months, market cap ~$56.7B, P/E ~84.9. Trading below both fresh PTs ($138 BofA, $150 Mizuho) → ~4–13% sell-side headroom.
  • Stretched well above moving averages (20-week EMA est. ~$112). Per the momentum playbook this extension is confirmation in an ACCELERATING + cluster-confirmed regime, not a fade signal — do not wait for a pullback that an accelerating name won't give.
  • Beginner-trap check: at ATH, rich multiple, retail going public via new leveraged ETFs. Mitigants: price below PTs, no earnings in window, segment-level acceleration is brand-new. Size to HIGH, not SUPREME — entering post-inflection-print without a fresh near-term catalyst.

Catalyst Calendar (next 30 days)

  • No earnings in the window — Q1 already printed 2026-05-04; Q2 2026 print est. ~2026-07-28 (late July/early Aug), OUTSIDE the 30-day window. Net positive: clean runway, no binary earnings risk on a fresh entry.
  • ~2026-06-04→06-18: leveraged-ETF flow ramp (ON/NXP/Microchip funds launched 6/1–6/3) — incremental structural bid.
  • Ongoing (June): NVIDIA MGX / 800VDC design-win news flow following the 2026-05-29 expansion announcement; cluster read-throughs from Navitas/Infineon.
  • Watch: summer semiconductor investor conferences (dates loose) for incremental AI-power commentary.
  • catalyst_date: none hard-dated in window → null. Next binary is Q2 (~late July).

What Would Change Our Mind

  • Q2 print (~late July) shows AI data-center revenue flat/down QoQ after Q1's +30% sequential → the growth leg is broken, narrative invalidated.
  • NVIDIA design-loss — Infineon or Navitas winning the 800V MGX power sockets onsemi is counting on.
  • Weekly close below the 20-week EMA (~$112, est.) → trend broken, exit, do not average down.
  • Auto/industrial guide-down deepens the trough → blended story can't hold an ~85x multiple.
  • Theme flips SATURATED (CNBC wall-to-wall, everyone long the leveraged ETF) with no fresh catalyst → trim into the mania.

Correlation Notes

  • High beta to the AI-datacenter-power cluster: NVDA (demand driver), VRT/Vertiv (power+cooling), Navitas/NVTS (GaN pure-play), Infineon (IFX), STM, NXP, Microchip (MCHP). Tracks SOXX/SMH.
  • Leading tells: pure-plays (Navitas, VRT) move first; onsemi follows the pack — watch them for early breaks both ways.
  • Inverse exposure to EV/auto sentiment (legacy SiC) — a fresh EV-demand scare can offset the AI bid; this name is two stories in one ticker.
  • Crowded long via the new ON/NXP/MCHP leveraged ETFs → amplified moves and harder deleveraging air-pockets on any semi-wide risk-off.