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 Infrastructure

Context

Trane has been using Danfoss Turbocor for decades. Carrier uses it too. York too. Daikin too. Turbocor is the only game in town for oil-free centrifugal refrigeration compressors. 70,000 units installed worldwide, zero competitors in 30 years. Every OEM uses the exact same heart in their highest-end oil-free centrifugal chiller lines. Hard to differentiate your product when the compressor inside it is something anyone can buy. What Trane wants out of the Garrett exclusive is exactly this: a core component nobody else can get their hands on. If this compressor works, Trane has something in its flagship chillers that Carrier and York don't have. That chip is worth the gamble for Trane.

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

What gamble? Garrett has never made an HVAC product. This is a turbocharger company. Millions of units a year. Roughly one in three turbocharged cars on the planet has their stuff in it. CEO Rabiller said he wants $1 billion in revenue outside turbochargers before 2030. HVAC compressors are one of the tracks he picked. Q3 net sales $902 million, net income $77 million. They also announced a $250 million stock buyback for 2026, and they're simultaneously developing fuel cell compressors. A company making roughly $300 million a year in profit trying to feed two brand-new business lines at the same time. Feels a bit tight to me. Q4 earnings drop February 19. That's when we'll see what management has to say about HVAC investment.

Trane is handing the core component of its next-gen flagship chiller to a company with zero HVAC experience. Garrett has no aftermarket network. No relationships with HVAC contractors. No deep knowledge of refrigerant systems. Refrigerant is not air. Liquid slugging, wet compression, refrigerant migration. These are HVAC-specific problems and Garrett has never dealt with any of them. Trane's engineers are going to have to hold their hand through all of it. On the surface this exclusive deal looks like Trane buying Garrett's technology. In reality Trane is probably putting in more than Garrett is. Engineers, test equipment, field resources, application know-how, all coming from Trane's side. What Garrett brings is compressor manufacturing capability and foil bearing production expertise.

(Source: GlobeNewswire, 2026.2.2 globenewswire.com)

Garrett's press release says their technology "represents a superior alternative to existing scroll, screw and magnetic bearing-based compressor solutions for most use cases." Magnetic bearing-based means Turbocor. Saying you're better than Turbocor before your product is even in production. That's too early. Garrett also says testing at Trane showed "over 10% efficiency improvement." But compared to what? They don't say. Danfoss's own marketing uses oil-injected screws as the punching bag, claiming "35% more efficient." Both of them are currently beating up the same straw man.

TechnologyMagnetic Bearing Systems

TechnologyFoil Bearing Heritage

My own read is Garrett will have a hard time beating Turbocor on efficiency. Magnetic bearings achieve full non-contact between rotor and bearing during operation. Losses are extremely low. Foil bearings also run non-contact, but the viscous shear losses in the gas film are not negligible at high rotational speeds. Garrett's real edge is cost. Foil bearings don't need the expensive electronic control system that Turbocor requires. Turbocor has to actively control rotor position across five degrees of freedom. The sensors, controllers, power amplifiers take up a big chunk of the BOM. Garrett needs none of that. Replace all of it with a few metal foils coated in solid lubricant. Material cost is way lower.

But this cost edge is still theoretical at this point. In the turbocharger business Garrett makes millions of units a year. Foil bearing unit cost is already spread thin. An HVAC compressor line doing a few thousand units in year one would be doing well. At that kind of volume, how much cheaper is Garrett really versus Turbocor? Hard to say. Once volume comes up the advantage will be obvious. The question is when volume comes up.

Garrett's announced product line covers 7 to 500 refrigeration tons. The 7-ton end is interesting. Turbocor's smallest TGS series starts at 40 tons. A 7-ton oil-free centrifugal doesn't exist on the market. Small commercial, distributed cooling, battery energy storage thermal management. These applications had no choice but scroll or screw before, because nobody could make a centrifugal this small. If Garrett can actually pull it off, this is a market segment Danfoss can't touch. Way more interesting than going head-to-head with Turbocor on 450-ton chiller projects.

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

The 500-ton end I doubt. Foil bearing load capacity is limited by gas film physics. Bigger cooling capacity means heavier rotors and bigger gas reaction forces. Turbocor gets to 450 tons using magnetic levitation with active control. Strong load-bearing capability. How Garrett gets to 500 tons on foil bearings, the press release doesn't say a word. Maybe multi-stage configurations. Maybe oversized bearing diameters. Maybe 500 tons is a roadmap number, not something they can do now.

Start-stop wear is the question I want answered most. Foil bearings have dry friction between foil and rotor at low speeds. They rely on surface coatings to take the punishment. Turbochargers follow the engine. Once started they blow past the dry friction zone fast, and turbocharger life only needs to match the car. A couple hundred thousand kilometers and it retires. HVAC chillers start and stop multiple times a day based on load. A single machine needs to run over ten years. Can foil coatings survive the accumulated start-stop cycles over a decade? Garrett's automotive data can't answer this. The operating conditions are too far apart. Trane's field testing exists to answer exactly this. Testing starts late 2026. Earliest preliminary data mid-2027.

Danfoss hasn't said a word publicly. 70,000 installed units and a 30-year failure mode database are Danfoss's strongest cards. But Trane is one of their biggest customers, and Trane now has a backup plan. Even if Trane doesn't cut Turbocor orders in the near term, the signal that "I've got another option" is leverage at the next price negotiation.

Commercial Building HVAC Market

Market Landscape

If Danfoss wants to fight back, pushing products into the small-tonnage end is the most direct move. Take Turbocor from 40 tons down to 20, even 10 tons. Directly block Garrett's differentiation space at the low end. The problem: as magnetic bearings go smaller, the electronic control system's fixed cost takes up a bigger and bigger share of BOM. At 10 tons, electronics might be half the bill of materials. The cost structure looks ugly. Danfoss could also optimize costs on the existing line, simplify the control system and push out a stripped-down version. But that blurs Turbocor's thirty-year brand positioning as "the best oil-free compressor" and sends mixed signals to current customers. The safest move is to do nothing and bet Garrett can't pull it off. Short-term that might be right. Garrett's HVAC experience is zero, mass production hasn't started, reliability data doesn't exist.

I lean toward thinking Garrett will eventually get a foothold in the HVAC market. The timeline will be longer than what Garrett is advertising. Production starting late 2026 doesn't mean volume ramps right away. Getting to decent shipment numbers by 2028, 2029 would be good. But Garrett's cost structure advantage is real. The small-tonnage market gap is real. And the most important thing: Trane has already put engineering resources and brand credibility on the line. Trane is not going to let this project die. My guess is Trane's investment in Garrett might be bigger than Garrett's own investment in HVAC.

Other compressor stuff from the AHR show. Copeland took two AHR Innovation Awards. Refrigeration category: a transcritical CO2 scroll compressor with DVI vapor injection. This product solves the flash tank vapor handling problem in CO2 systems. No more need for parallel compressors. System design gets a lot simpler. A key push for North American supermarkets moving from HFC to CO2. Heating category: a heat pump-optimized modulated vapor injection product portfolio. Copeland also launched the KF variable-speed scroll platform, 1.5 to 5 tons for residential heat pumps, 10:1 turndown ratio, optimized for R-454B. Copeland's strategy is clear: scroll eats residential and light commercial, CO2 eats supermarket refrigeration, stay out of oil-free centrifugal. Different battlefield from Garrett and Danfoss. (Source: Copeland AHR)

AHR Expo

Compressor Manufacturing

Building Systems

Gree's G-Storm 200Hz ultra-high-speed compressor made the AHR cooling category finals (Source: ACHR News). 200Hz is an aggressive number in the compressor world. But I'm not bullish on Gree in North American commercial. The brand recognition problem is just too big. North American HVAC engineering firms are still at the "we know you exist but we won't put you in the spec" stage when it comes to Chinese commercial chillers. This was more Gree showing face. Hitachi scroll compressors came to AHR for the second year running, showing R-290 and R-454B new models. LG brought out a DualJet reciprocating compressor and CurvedSpoke motor. Not going to say much about these. Component-level stuff, no direct impact on how the end market shapes up.

Connection to industrial air compressors. The bearing setup in Atlas Copco's ZR/ZT oil-free series comes from the same technical family as Turbocor. If foil bearings get validated by a few thousand installed HVAC centrifugal compressors, people in oil-free industrial air compressors will take notice. But industrial air compressor customers are way more paranoid about reliability than HVAC buyers. One hour of factory downtime costs enough to buy several compressors. In that world nobody wants to be the guinea pig. From Garrett getting established in HVAC to foil bearings making their way into industrial air compressors, five years minimum. And it probably won't be Garrett doing it. It'll be Atlas Copco or Ingersoll Rand kicking off their own evaluation after they've seen enough HVAC field data.

Garrett's Q4 earnings come out February 19. Danfoss is private, financials aren't public. Trane's full-year 2025 earnings came out January 29. Annual revenue $18.5 billion, Americas commercial HVAC up 8%. Going forward the only variable that matters is how Garrett's compressor performs in real buildings. Field testing starts late 2026, preliminary data by mid-2027. Until then Danfoss can afford to wait. Carrier and York can sit back and watch. Trane can't. The exclusive is signed. The next-gen chiller product roadmap already has Garrett baked in. If Garrett drops the ball Trane has no fallback except crawling back to Turbocor, and that's the same as publicly admitting they called it wrong. So Trane is going to keep throwing resources at this project until it works.

◆ ◇ ◆