Trane, Garrett, and the Oil-Free Centrifugal Compressor Shakeup

Trane Technologies signed an exclusive agreement with Garrett Motion on February 2. Garrett built an oil-free centrifugal HVAC compressor using foil bearings instead of magnetic bearings. Trane got exclusive rights for the commercial HVAC space.

Most industry media reported this as a product launch. I think they got it backwards. That compressor is still a prototype. Mass production hasn't even started. What's actually worth paying attention to isn't the product itself. It's why Trane was willing to sign exclusive at this stage.

Commercial HVAC system in a large building

Commercial HVAC Infrastructure

Context

Here's the thing about Turbocor. Trane has been buying from Danfoss Turbocor for decades. So has Carrier. So has York. So has Daikin. Everybody buys from the same shop. 70,000 units installed worldwide, and for thirty years nobody has put up a credible alternative for oil-free centrifugal refrigeration compressors. Think about what that means if you're Trane's product team: the most critical component in your highest-end chiller line is the same part Carrier can walk in and order next Tuesday. Hard to sell differentiation when your competitor's machine has the same heart transplanted from the same donor. That's the real reason Trane jumped on this deal before Garrett even had a production line. A core component nobody else can source. If the compressor works, and that's still a big if, Trane's flagship chillers suddenly have something Carrier and York can't copy by picking up the phone. That chip alone justifies the risk.

(Source: Trane Technologies website, 2026.2.2 trane.com)

And there is real risk. Garrett has never touched HVAC. Not once. This is a turbocharger company, through and through. They ship millions of units a year. Roughly one in three turbocharged cars on the road has their hardware spinning inside. CEO Rabiller has been talking publicly about wanting $1 billion in non-turbocharger revenue before 2030. HVAC compressors are one of the bets he's placed. Q3 numbers: net sales $902 million, net income $77 million. They also announced a $250 million stock buyback for 2026, and they're simultaneously pouring money into fuel cell compressors too. So a company generating roughly $300 million a year in profit is now trying to feed two entirely new business lines at the same time. I keep looking at those numbers and it feels tight. Q4 earnings drop February 19. That call is where we'll hear whether management is actually backing the HVAC line with real money or just talking about it.

Now here's what bugs me about the deal structure. Trane is handing the core component of its next-gen flagship chiller to a company that has never shipped a single HVAC product. No aftermarket network. No relationships with contractors. No intuition for what happens when refrigerant does things air doesn't do. Liquid slugging. Wet compression. Refrigerant migration at shutdown. I've watched turbocharger engineers discover these failure modes for the first time. It's a steep learning curve, and Garrett is starting from zero. Trane's engineers are going to be babysitting this thing for years. On the surface the exclusive looks like Trane buying technology. Practically speaking, Trane is probably contributing more to this project than Garrett is. Engineering hours, test rigs, field access, decades of application know-how, all flowing from Trane's side. What Garrett actually brings to the table is precision manufacturing at scale and the foil bearing production know-how. That's it. That's the trade.

(Source: GlobeNewswire, 2026.2.2 globenewswire.com)

Garrett's press release rubbed me the wrong way. They describe their technology as "a superior alternative to existing scroll, screw and magnetic bearing-based compressor solutions for most use cases." Magnetic bearing-based means Turbocor. They're out here calling themselves better than Turbocor before their product has even seen a production line. Bit early for that. Garrett also claims testing at Trane showed "over 10% efficiency improvement." Compared to what, though? They don't say. And Danfoss's own marketing does the same trick. They claim Turbocor is "35% more efficient" but the punching bag is oil-injected screws. Everybody's beating up the weakest kid on the playground and calling themselves champion.

Close-up of precision bearing components

TechnologyMagnetic Bearing Systems

Turbocharger cutaway showing foil bearing assembly

TechnologyFoil Bearing Heritage

My honest assessment on the efficiency question: I don't think Garrett can beat Turbocor there. Magnetic bearings achieve full non-contact levitation during operation. The rotor floats. Losses are tiny. Foil bearings get to non-contact too, eventually, once rotational speed builds up and the gas film develops. But there's viscous shear happening in that film, and at high RPM the drag adds up more than people seem to acknowledge. Where Garrett actually has a shot, and what nobody in the trade press seems to be picking up on, is cost. Turbocor's magnetic bearing setup needs active rotor position control across five degrees of freedom. Sensors, controllers, power amplifiers, a whole electronics package that eats a serious chunk of the bill of materials. Garrett throws all of that away. Replace it with a few metal foils and a solid lubricant coating. Mechanically elegant. And cheap. Well, cheap in theory.

Because here's the catch. That cost advantage only shows up at volume. Garrett runs millions of turbochargers per year, so foil bearings cost them almost nothing per unit on that side. An HVAC compressor line that ships a few thousand units in year one would be doing well. At those numbers, is Garrett actually undercutting Turbocor on price? I genuinely don't know. Give them a few years of ramp and the math will be obvious. Getting there is the hard part, and the in-between period is where the whole thing could stall out.

Testing starts late 2026. Earliest preliminary data mid-2027.

The product range Garrett announced runs from 7 to 500 refrigeration tons. Let me take the ends separately because they tell very different stories.

The 7-ton end is the part I find genuinely exciting. Turbocor's smallest offering, the TGS series, bottoms out at 40 tons. A 7-ton oil-free centrifugal compressor simply does not exist on the market today. Small commercial spaces, distributed cooling setups, thermal management for battery energy storage systems, all stuck with scroll or screw because nobody could make a centrifugal that small. If Garrett can actually deliver here, they're not fighting Danfoss. They're walking into a room Danfoss can't even enter. I care about this a lot more than whether Garrett can win a 450-ton chiller bid against a product that's had thirty years to prove itself.

The 500-ton claim I have a harder time with. Foil bearings support the rotor through gas film pressure, and gas films have physics-imposed load limits. Bigger cooling capacity means heavier rotating assemblies, bigger reaction forces. Turbocor handles heavy loads at 450 tons by throwing more current at electromagnets. Brute force, but it works. How Garrett plans to get foil bearings to support a 500-ton class rotor, the press release says nothing about. Maybe they're thinking multi-stage. Maybe oversized bearing surfaces. Maybe 500 tons is an aspirational number on a PowerPoint slide and not something they can build today. I'd love to ask their engineering team directly.

The thing I most want to see data on is start-stop durability. This is where I think the real landmine is hiding. Foil bearings have a fundamental problem: at low speed, before the gas film develops, there's metal-on-metal contact between the foil and the rotor shaft. They coat the surfaces with solid lubricant to survive this, and in turbochargers that's fine. A turbo follows the engine. Once you start the car it spins up past the friction zone in milliseconds, and the whole assembly only needs to last as long as the vehicle. A couple hundred thousand kilometers and everyone's happy. HVAC is a completely different ask. Chillers cycle on and off multiple times per day based on building load. One machine needs to run for ten, fifteen, twenty years. The accumulated wear from all those start-stop transitions is orders of magnitude beyond what any turbocharger foil bearing has ever experienced. Garrett's automotive durability data tells us nothing here. Completely different operating envelope. This is exactly what Trane's field testing program needs to answer, and it's why I'm circling late 2026 and mid-2027 as the dates that actually matter.

Danfoss has been quiet. No public statement, no press release, nothing. I'd argue 70,000 installed units and a thirty-year failure mode database are the strongest cards Danfoss holds. Nobody else has that depth of reliability evidence. But they have to be nervous. Trane is one of their biggest customers, and Trane just told the world "we're developing an alternative." Even if Trane doesn't cut a single Turbocor order this year or next, the message at the next price negotiation is unmistakable: I've got options now.

Rooftop chiller installation on a commercial building

Commercial Building HVAC Market

Market Landscape

So what does Danfoss do about it? If I'm sitting in their strategy meeting, the most aggressive move is to push Turbocor down-market. Take the current floor at 40 tons and drive it to 20, maybe even 10. Go right at Garrett's small-tonnage differentiation before Garrett can establish themselves there. The problem is economics. Magnetic bearing systems have a fixed cost floor. The sensors, the control electronics, the power stage. That floor doesn't shrink as the compressor gets smaller. At a 10-ton unit, electronics might be half the total bill of materials. The unit cost looks terrible. Danfoss could alternatively strip down the control system on the existing line, push out a cheaper version, sacrifice some performance margin. But that undercuts thirty years of "Turbocor is the best oil-free compressor" positioning and confuses the install base. The third option, and frankly the one I think Danfoss will actually pick, is to sit tight and bet that Garrett can't execute. In the short run that bet might even be correct. Garrett's HVAC experience is nonexistent, production hasn't started, and there isn't a single hour of real-world reliability data.

I think Garrett gets into HVAC eventually. It's going to take longer than their press materials suggest. "Production starting late 2026." Sure, maybe, but that doesn't mean units are shipping in any meaningful volume. Getting to real commercial quantities by 2028, 2029 would be a solid outcome. The underlying cost structure advantage with foil bearings is genuine. The small-tonnage gap in the market is genuine. And the thing that seals it for me: Trane has already committed engineering headcount and staked its brand on this. Trane isn't going to walk away. If anything, I suspect Trane's total investment in making this compressor work, counting engineering time, lab resources, and field test infrastructure, might exceed what Garrett itself spends on the HVAC program.

I want to mention a few other things from AHR since they're related, even if they're not part of the Trane-Garrett story directly.

Copeland walked away with two AHR Innovation Awards. In the refrigeration category they showed a transcritical CO2 scroll compressor with DVI vapor injection. This is actually a meaningful product because it solves the flash tank vapor handling headache in CO2 systems. You no longer need parallel compressors to deal with flash gas. System piping gets simpler. For North American supermarkets in the middle of switching from HFC to CO2, this matters. In the heating category they won with a heat pump-optimized modulated vapor injection portfolio. They also launched the KF variable-speed scroll platform: 1.5 to 5 tons, residential heat pumps, 10:1 turndown ratio, designed around R-454B. Copeland's playbook is pretty clear if you step back and look at it. Scroll owns residential and light commercial. CO2 scroll goes after supermarket refrigeration. And they stay completely out of the oil-free centrifugal fight. They're playing a different game than Garrett and Danfoss. (Source: Copeland AHR)

AHR Expo

Compressor Manufacturing

Building mechanical room with chiller equipment

Building Systems

Gree brought their G-Storm 200Hz ultra-high-speed compressor, which made the AHR cooling category finals (Source: ACHR News). 200Hz is a genuinely aggressive number for a compressor. But I just can't get excited about Gree in North American commercial. The brand recognition gap is enormous. Engineering firms that spec commercial chillers in the US and Canada know Gree exists but they won't put them in a specification document. Not yet, maybe not for a long time. This felt like Gree planting a flag more than making a real market play. Hitachi was at AHR for the second year with scroll compressors, showing R-290 and R-454B variants. LG showed a DualJet reciprocating compressor and CurvedSpoke motor. Component-level stuff. Interesting to engineers, irrelevant to how the market actually reshapes.

One more thread worth pulling on. Atlas Copco's ZR/ZT oil-free air compressor series, the industrial side, not HVAC, uses bearing technology from the same lineage as Turbocor. Same family tree. So if foil bearings eventually prove themselves in a few thousand installed HVAC centrifugal compressors, you can bet people in the oil-free industrial compressed air world will start paying attention. But industrial air compressor buyers make HVAC people look reckless by comparison. An hour of unplanned downtime in a semiconductor fab or pharmaceutical plant costs more than the compressor itself, sometimes multiples of it. Nobody in that world volunteers to be the first adopter of anything. Realistically, from the point where Garrett establishes credibility in HVAC to the point where foil bearings start showing up in industrial air compressors, I'd say five years at minimum. And it probably won't be Garrett doing it. It'll be Atlas Copco or Ingersoll Rand running their own internal evaluation programs after they've seen enough HVAC field data to feel comfortable.

Garrett Q4 earnings come out February 19. Danfoss is private so we won't see their numbers. Trane's full-year 2025 earnings already landed January 29: annual revenue $18.5 billion, Americas commercial HVAC up 8%. None of the financial details change the fundamental picture, though. What actually determines how this plays out is simple: does Garrett's compressor work in a real building? Field testing starts late 2026. Preliminary data by mid-2027. Until those results come in, Danfoss can afford to wait. Carrier and York can watch from the sidelines. Trane can't. Trane signed the exclusive. The next-gen chiller roadmap already has Garrett written into it. If Garrett falls short, Trane's fallback is going back to Turbocor cap in hand, which amounts to a public admission that they bet wrong. That's not something Trane's leadership wants to explain on an earnings call. So Trane will keep pouring resources into this project until it works. They don't really have a choice anymore.

◆ ◇ ◆