Belimo jumps 8% as Morgan Stanley upgrades stock on FY25 data center boost

Belimo shares jumped 8% after Morgan Stanley upgraded the stock to overweight and lifted its price target to CHF1,100 from CHF860, citing accelerating demand from AI data center cooling. The broker now sees data center revenue reaching 38% of sales by 2030, with revenue projected to rise to CHF1.31 billion in 2026, CHF1.53 billion in 2027 and CHF1.78 billion in 2028. The call highlights a structural shift from air cooling to liquid cooling and implies Belimo is positioned to benefit from hyperscaler capex growth and shorter replacement cycles.

Analysis

This is not just a single-name upgrade; it is a signal that AI infrastructure capex is tilting from compute-dominated spending toward thermal management, and that change should persist longer than consensus expects. The key second-order effect is mix expansion: as cooling architectures become more specialized, vendors with specification power in the highest-performance tier can grow faster than the data center buildout itself, which supports multiple expansion even if broad IT capex slows. That makes the opportunity less about unit volume and more about content-per-rack inflation.

The biggest winner set is not just the company named, but the upstream ecosystem tied to liquid cooling, high-pressure valves, fluid controls, sensors, and power/thermal integration. More importantly, the prize shifts away from commoditized HVAC and generic mechanical contractors toward companies that sit closest to chip-level cooling architecture, because those components are harder to redesign out once embedded in a platform. A hidden loser is any incumbent whose demand is still tied to traditional air-cooling retrofits; if the refresh cycle really compresses, legacy installed base advantages erode quickly and procurement resets can happen in under a full server replacement cycle.

The main risk is architectural disintermediation: if hyperscalers standardize coolant distribution units and bundle more functionality inside server OEM or CDU platforms, standalone component vendors lose pricing power and share of wallet. That risk is not a near-term earnings issue; it is a 12-36 month specification risk that would show up first as slower order conversion and then as margin pressure, especially after the market has already capitalized a premium growth narrative. On the other hand, the market may still be underestimating how durable the replacement cycle re-rating is, since the move from 20+ year building economics to 5-7 year server economics can structurally lift recurring demand and reduce cyclicality.

Key Decisions for Investors

• Long the thermal-content winner versus a legacy HVAC/controls peer: pair a long in the high-specification cooling beneficiary with a short in a lower-growth industrial controls name over the next 3-6 months; target alpha from mix expansion and multiple divergence, with stop-loss if hyperscaler capex guidance rolls over.
• Buy 6-12 month call spreads on the cooling beneficiary into the next large-cap cloud capex print; the trade has favorable convexity if AI infrastructure spend is revised up again and should outperform on any evidence of continued liquid-cooling penetration.
• Add to a basket of liquid-cooling supply-chain names on pullbacks, but size modestly: upside is tied to a multi-year architecture shift, while downside comes from CDU integration that can compress standalone value; use 20-25% trailing stops.
• For hedge funds with industrial exposure, short the most commoditized HVAC/MEP subcontractor basket against the high-specification cooling theme for a 1-2 quarter relative-value trade; the thesis is that pricing power migrates upward in the stack.
• Set a catalyst watchlist around next hyperscaler capex commentary and data center design updates over the next 1-2 quarters; if CDU integration becomes explicit, reduce or reverse long exposure immediately because the spec-power thesis would be impaired.