Dossier · GTX · Dormant
GTX · Garrett Motion Inc. · Stock research
Last analysed ·
Current thesis
Turbo cash-cow re-rated (~8–9x → ~18x) on a datacenter/industrial oil-free cooling narrative; the parabolic Feb→June leg (~$20→$36) is spent, shares faded ~11% to the 50-day near $32.30, and the ~late-July Q2 print is the first chance to attach real cooling revenue. Fully-priced, MATURING, a wait-for-the-print name.
Invalidation trigger
A weekly close below $29 loses the May–June breakout shelf and the rising trend off the February base. Secondary: an FY26 sales guide cut below the $3.6B floor, or a Q2 print (~late July) with no datacenter/industrial-cooling revenue line, breaks the re-rate thesis.
Thesis status
Invalidated resolved published trigger fired How this is scored →Latest analysis and events for GTX —
As of 2026-08-08, orbyd's latest analysis for Garrett Motion Inc. (GTX): Turbo cash-cow re-rated (~8–9x → ~18x) on a datacenter/industrial oil-free cooling narrative; the parabolic Feb→June leg (~$20→$36) is spent, shares faded ~11% to the 50-day near $32.30, and the ~late-July Q2 print is the first chance to attach real cooling revenue. Fully-priced, MATURING, a wait-for-the-print name.
Invalidation trigger: A weekly close below $29 loses the May–June breakout shelf and the rising trend off the February base. Secondary: an FY26 sales guide cut below the $3.6B floor, or a Q2 print (~late July) with no datacenter/industrial-cooling revenue line, breaks the re-rate thesis.
Current Thesis
The datacenter-cooling re-rate that carried Garrett from a ~$20 February base to a 52-week high of $36.23 has stopped paying. On 2026-07-29 the company delivered the print the story needed — Q2 EPS $0.53 against a $0.45 estimate, net sales $976M against $965.45M, record adjusted EBIT of $152M at a 15.6% margin — and lifted every line of FY26 guidance. Deutsche Bank upgraded to Buy with a $36 target the next morning. The tape absorbed all of it and gave it back: $29.92 close on 2026-07-28, +3.73% premarket to $31.035 on 07-29, $29.85 close on 2026-08-07. What the call actually clarified is that the cooling narrative has no 2026 P&L: CEO Olivier Rabiller put first E-Cooling production shipments in 2027 and datacenter-specific applications "between the end of 2027 for data centers and beginning of 2028." The narrative is SATURATED — everyone who wanted the story owns it, good news no longer produces a bid, and the next proof point is more than a year out. Price now sits on the $29–30 shelf with RSI(14) at 42.5, mid-range, with no base built.
Bullish and bearish views on Garrett Motion Inc.
The model's bull view on Garrett Motion Inc. (GTX), in brief: Q2 CY2026 (2026-07-29) beat on both lines. The bear view: The best print of the cycle was sold. Seven sessions after a beat-and-raise and an upgrade, the close ($29.85, 2026-08-07) is below the pre-print close ($29.92, 2026-07-28). When a name stops rewarding its own good news, the marginal buyer is already in. The raise mostly lifted… Both cases follow in full.
Bull Case
- Q2 CY2026 (2026-07-29) beat on both lines. Net sales $976M (+7% reported, +5% constant currency) vs $965.45M consensus; EPS $0.53 vs $0.45; net income $101M; record adjusted EBIT $152M at 15.6% margin, up 200bps; adjusted FCF $122M; cash $158M.
- Guide raised across all four measures. FY26 net sales $3.7–3.9B (from $3.6–3.9B), net income $330–360M (from $300–360M), adjusted EBIT $560–600M (from $520–600M), adjusted FCF $385–475M (from $355–475M). Management framed the midpoint as $3.8B sales / $580M adjusted EBIT / $430M adjusted FCF.
- Datacenter exposure moved from press release into guided revenue. Industrial turbo sales for FY26 were taken to approximately $200M from a prior "$100M-plus," with more than $80M already booked in H1, and Garrett disclosed a major production award for MEG turbos used in datacenter gensets. This is the first datacenter dollar figure inside a guided number rather than a partnership announcement.
- Capital return continuing. $28M repurchased in Q2, $115M year-to-date against the $250M 2026 authorization, $135M remaining; $0.08/sh dividend payable 2026-09-15; 186.8M shares outstanding at 2026-06-30.
- Every named sell-side mark sits above the tape. Deutsche Bank upgrade to Buy, PT $36 (2026-07-30); Stifel Buy $36 and JPM Overweight $33 (both 2026-05-14) against a 2026-08-07 close of $29.85.
Bear Case
- The best print of the cycle was sold. Seven sessions after a beat-and-raise and an upgrade, the close ($29.85, 2026-08-07) is below the pre-print close ($29.92, 2026-07-28). When a name stops rewarding its own good news, the marginal buyer is already in.
- The raise mostly lifted the floor. The sales range went $3.6–3.9B → $3.7–3.9B; the top end did not move, and the $3.8B midpoint essentially matched the $3.785B consensus already on the tape.
- Cooling revenue is a 2027–2028 event. First E-Cooling production shipments guided to 2027, datacenter applications to late 2027/early 2028 (2026-07-29 call). Nothing between here and then converts the story into reported revenue.
- Insider distribution into the highs, no disclosed offsetting buying. CFO Sean Deason sold 110,000 sh at $31.93 (2026-06-05);
- Repurchase pace decelerated as the stock fell. $87M in Q1 versus $28M in Q2 — the buyback did not step up into the drawdown.
- Structure has rolled. -17.6% from the $36.23 52-week high, +3.2% over three months, RSI(14) 42.5. The 50-day stood at $31.63 on 2026-07-10; the 2026-08-07 close is well beneath that level.
Setup & Price Structure
Life-cycle: SATURATED, dated 2026-07-29/30 — the beat-and-raise and the Deutsche Bank upgrade landed inside 24 hours and produced one premarket pop with no follow-through, while coverage of the cooling angle is now broad and repetitive. The February-base uptrend is not broken (the 200-day was $20.96 on 2026-07-10, far below price), but the intermediate structure has turned over: three consecutive lower highs since the late-June $36.23 peak and a drift into a $29–30 area marked by the 07-28 and 08-07 closes.
Crowding and positioning observables, stated as observables: three named sell-side targets at or above $33 with no published sell rating; a beat-and-raise that produced no closing gain; insider sales at $31.93 and $33.00 with no disclosed open-market purchases; a buyback that shrank from $87M to $28M quarter-over-quarter; and $135M of authorization left, which is a finite bid rather than a standing one. RSI at 42.5 is the uncomfortable middle — not oversold enough to mark a flush, not strong enough to signal accumulation.
Inference, not measurement: the $29–30 area is where the pre-print sellers and the post-print disappointment meet, so it is the level that carries information. A weekly close that holds it while the industrial-turbo line compounds toward the guided ~$200M would argue the drawdown was multiple compression rather than thesis failure.
Catalyst Calendar (next 30 days)
- 2026-09-15 — $0.08/sh quarterly dividend payable, declared with Q2 results. Confirms the capital-return cadence; not a narrative event.
- ~2026-10-28 (est.) — Q3 CY2026 results. Not yet formally scheduled; Q2 landed 2026-07-29. First read on whether industrial turbo tracks to the raised ~$200M FY26 figure and whether FY26 sales guidance holds at $3.7B+.
Elapsed catalysts
- No scheduled company event inside the next 30 days. Q2 was reported 2026-07-29; the next company-controlled print is roughly a quarter out. (passed 11d ago)
- 2027 (est.) — first E-Cooling production shipments per the 2026-07-29 call; datacenter-specific applications late 2027/early 2028. The date that actually settles the cooling thesis sits outside any tradeable window from here. (passed 11d ago)
What Would Change Our Mind
The structure that has to fail is the $29–30 shelf that has absorbed the post-print supply. A weekly close below $29 loses it and the rising trend off the February base, and would date the re-rate as fully unwound; a secondary confirmation would be FY26 sales guidance cut back below $3.7B at the Q3 print, or that print passing with E-Cooling still absent from guided revenue and industrial turbo tracking short of ~$200M.
What would argue the other way, with equal specificity: a weekly close back above $33 — reclaiming the JPM mark and the July trading area — on the back of a named datacenter E-Cooling order with dollars attached, or an industrial-turbo figure at the Q3 print running ahead of the ~$200M FY26 guide. Absent one of those, the name is a wait-for-a-base situation and fresh entries into the current drift have no near-term event to resolve them.
Correlation Notes
Garrett re-rated on a datacenter-cooling narrative but earns its revenue from cyclical light-vehicle and commercial turbochargers. That gap means headline risk now arrives from the datacenter power/cooling complex — Vertiv, Modine, Trane, Ingersoll Rand (the 2026-05-12 partner) — while reported results depend on global vehicle build rates and the industrial genset cycle. The two can move in opposite directions, and the July–August drawdown occurred through a window in which the company's own beat-and-raise was the dominant news, so it reads as name-specific de-rating rather than sector beta.
Within the guided P&L, the datacenter linkage that exists today runs through MEG turbos sold into backup gensets, which ties to hyperscaler datacenter build schedules and the large-bore engine makers, not to liquid-cooling capex. Any read-across from a cooling-hardware peer's order book is therefore an inference about 2027–2028 optionality, not about the FY26 $3.7–3.9B guide.
Notes
- Datacenter E-Cooling revenue is not in FY26 guidance: first production shipments guided to 2027, datacenter-specific applications late 2027/early 2028 (2026-07-29 call).
- Core P&L remains cyclical light-vehicle and commercial turbochargers; today's guided datacenter exposure is MEG genset turbos inside the ~$200M FY26 industrial turbo line.
- Buyback authorization is finite: $135M left of the $250M 2026 program as of 2026-06-30, after $87M in Q1 and $28M in Q2.
- Quarterly dividend is $0.08/sh; the Q3 declaration is payable 2026-09-15. Q3 CY2026 results are not yet formally scheduled.
Related · shared themes
HPE
Hewlett Packard Enterprise Company
Five straight up sessions took HPE to a $53.22 close on 2026-08-07 with the 2026-08-04 gap floor at $50.24 still untested; Trefis put the five-day move at +19.8% against +5.6% for the S&P 500, so this leg is no longer pure sector beta. Forward P/E 13.79 versus 49.66 trailing. The still-unconfirmed ~2026-09-02 Q3 print against the $11.5–12.1B / $0.88–0.93 guide is the binary.
MRCY
Mercury Systems Inc
Defense-electronics turnaround still re-rating on record bookings and a promised FCF turn, but the stock has given back ~20% from its $128.45 ATH and now chops near the analyst median (~$103) into a binary ~Aug 10-18 Q4/full-year print. Consolidation, not acceleration — the move is maturing into the report.
SIMO
Silicon Motion Technology Corporation
Controller-mix decoupling got its first hard datapoint on 2026-08-06 — SIMO +8.42% to $267.43 while Western Digital fell 16% and SanDisk 11% on beat-but-guide-light prints — then gave most of it back on 08-07 (-4.23% to $256.12, $18.87 off the high, thinnest volume of the run). Six sessions after a beat-and-raise the ADS sits $1.02 above the 07-30 reaction close; the 50-day at $286.82 is unreclaimed.
XMTR
Xometry, Inc.
Q2 on 2026-08-04 cleared every self-set number — revenue $229.3M (+41%), marketplace +45%, Adj EBITDA $14.1M — and the new CEO raised FY2026 growth to 33–34% from 27–28%. Price took three sessions to agree, closing $93.15 on 08-07 back above the $90.65 50-day, still under the $106.08 July high.