Thermal design engineer driving liquid-cooling deployment at 100+ kW/rack AI clusters.
Audience Profile
- Age / Experience: 7-15 years
- Current role: Thermal Engineer / Cooling Systems Architect (Hyperscaler / colo / OCP-contributing org)
- Top pain points: Direct-to-chip vs immersion decision data lacking; CDU sizing methodology across vendors not standardised; Server vendor liquid-cooling SKU lead times
- Top decision blockers: OEM warranty on liquid-cooled chassis; Existing air-side handler asset stranded if retrofitted; Coolant chemistry vendor lock-in
What This Segment Needs
- Information: Vendor-neutral direct-to-chip vs immersion field reliability data, CDU sizing/derating methodology, and liquid-cooling SKU lead-time benchmarks.
- Tools: CFD/thermal sandboxes (Icepak, Flotherm, Fluent), CDU sizing calculators, and coolant-chemistry compatibility matrices.
- Services: OEM liquid-cooled-chassis warranty review, retrofit stranded-asset analysis, and second-source coolant qualification support.
Top 5 Companies for You (Fit Score)
| Rank | Company | Score | Why | |------|---------|-------|-----| | 1 | Modine Manufacturing | 75/100 | Data-center sales guided toward ~$1B run-rate FY2025-FY2027 from sub-$200M base, triple-digit YoY; 'Heat Transfer Research Engineer, Two-Phase Cooling' (2026-04-28) for high-density GPU racks. Caveat: business_signals_180d empty, no backlog. | | 2 | Munters Group | 72/100 | DCT ~SEK 7B of SEK 15.5B FY2024 base, 'fastest-growing', 'sustained double-digit demand' Q3 2025; 5 DCT reqs 2026-02-27 to 2026-05-06. Field travel up to 60%; no Staff/Principal track. | | 3 | CoolIT Systems | 70/100 | Five 2026 reqs span full lifecycle incl. 'Senior Liquid Cooling Design Engineer' (2026-03-03), CDU mechanical (2026-04-08); deep in-house cold-plate/CDU IP. business_signals_180d empty - no momentum disclosed. | | 4 | nVent Electric | 65/100 | Five cooling reqs 2026-02-27 to 2026-05-08, vertically complete; one team spans direct-to-chip AND immersion, CFD required (2026-04-22). Divested Thermal Mgmt to Brookfield (~$1.7B, 2025); growth figure prior-based. | | 5 | Iceotope | 60/100 | Five reqs 2026-02-27 to 2026-05-06 incl. 'Manufacturing Operations Manager' production scale-up; ANSYS Icepak/Flotherm, two-phase + leak/endurance testing. business_signals_180d empty; Senior IC ceiling only. |
Deal-Breakers (Your Hard Preferences)
No hard preferences declared for this segment.
How to Evaluate Any Company in this Niche (Checklist)
- [ ] Check growth signals: count 2026 DC-cooling reqs spanning R&D->ops->sales; require >=4 with a two-phase/immersion research title, not sales backfill alone.
- [ ] Check comp data: none of these firms disclose comp - pull Levels.fyi/Blind for "thermal engineer" + the OEM; expect industrial-mfg bands below FAANG.
- [ ] Check learning signals: confirm a named CFD stack (Icepak/Flotherm/Fluent) AND a two-phase/immersion R&D req, not air-side-only CRAH/CRAC.
- [ ] Check stability signals: watch DCT revenue concentration (e.g. Munters DCT ~45% of group) and segment divestitures (nVent->Brookfield 2025).
- [ ] Check culture signals: ask the interviewer whether a Staff/Principal thermal IC track exists or titles top out at Senior.
- [ ] Check lock-in signals: ask whether liquid-cooled-chassis OEM warranty and coolant chemistry are single-vendor or second-sourceable.
Reverse-Hype Watch
- Capacity-expansion staffing unbacked by customer signals: 4 of 5 firms post 'production scale-up' / 'capacity expansion' reqs while business_signals_180d is empty - no customer wins, order book or backlog.
- Financials for Modine and nVent are AI priors (synthesized_from 'none'), not signal-confirmed.
- IC ceiling tops at 'Senior' across most 5-role samples; no Staff/Principal/Distinguished tier visible.
- Munters FY2025 net sales reported only as 'above ~SEK 15.5B' - DCT growth magnitude unquantified.
Under-reported for this segment: the exact decision data you need - CDU sizing methodology variance across vendors, coolant-chemistry lock-in, and OEM warranty terms on liquid-cooled chassis - is essentially absent from public hiring and earnings signals. Hiring breadth and run-rate headlines are visible; order-book confirmation and head-to-head direct-to-chip-vs-immersion field reliability data are not, so treat capacity claims as forward-looking until backed by named customer wins.