SKF's Magnetics Business Is a Data-Center Story Hiding Inside a Bearings Spinoff

The industrial pure-play printed 16.3% margins, pulled its cost program forward six months — and quietly became a supplier to the AI build-out just as the tape turned against that trade.

SKF-A.ST · Earnings Call · 2026-10-21

The exception inside the exception

SKF spent this call doing two jobs at once: selling the market on an industrial pure-play that no longer exists as a reported entity, and quietly introducing a growth asset that has nothing to do with bearings. The headline numbers were fine, not spectacular — net sales up 5.5% with 4.3% strong organic growth, adjusted operating margin up 0.9pp to 16.3% for continuing operations, operating cash flow of SEK 2.1bn versus SEK 1.8bn. But the composition matters more than the level, and the composition is where the story sits. Bearing Solutions is still 74% of sales and 85% of adjusted operating profit, at an 18.8% margin. Specialized Industrial Solutions (SIS) is 26% of sales but only 23% of profit, at 14.0% — yet it grew 2.4pp of margin in a single quarter, with all four business units growing. Within it, Aerospace and Magnetics led, and Magnetics is the part worth staring at. It is a roughly SEK 1.5bn business — small — but management spent disproportionate airtime on it: magnetic bearings, high-speed electric motors, control cabinets, frequency drives, condition monitoring, sold as a stack into data centers, semiconductors, and energy. The pitch is energy efficiency (up to 30% savings claimed), oil-free operation, embedded sensors that "scale AI." This is not bearings. This is rotating equipment for the compute build-out, dressed in SKF's engineering vocabulary.

The AI adjacency nobody priced

Here is the tension. Globally, AI data center was one of the most powerful themes on the tape over the trailing year — and one of the most punished over the last ninety days, with a long list of decliners in data-center capacity, high-bandwidth memory, co-packaged optics, and high-performance compute. SKF's Magnetics exposure is a second-derivative way to own that theme, in a stock nobody screens for it, and it arrives at the moment the first-derivative names are de-rating. If the AI capex cycle holds, this is an unappreciated contributor; if it rolls over, SKF's exposure is small enough (SEK 1.5bn of a much larger base) that it becomes an air pocket rather than a hole. Either way, the valuation debate about SKF is not currently a debate about Magnetics. What has been dropped is arguably more revealing. In Q2, humanoid robotics and the Leaderdrive venture dominated SKF's keyword profile — a genuine company-unique theme with a joint venture, a component-supply strategy, and 80% assortment coverage claim. This quarter it gets one clause: "that initiative is progressing according to our expectation and plan." No expansion, no follow-up question in the Q&A. That silence is a signal about management's own conviction on timing, or at least about what it wants the equity story to be.

The spin is doing the talking

Vertevo — the Automotive business — has a prospectus in November, a Capital Markets Day on 17 November, and a first trading day targeted for 1 December on Nasdaq Stockholm. Management now speaks of Automotive as discontinued operations and of "two competitive and strong and fit-for-purpose businesses." The industrial remnant is being dressed for that listing by a rightsizing program that has just been pulled forward six months, to a SEK 2bn run-rate by the first half of 2027 instead of the second. It contributed about SEK 350m in the quarter. Recall that in July, the CFO was explicit about sequencing: “Moving on, now it will be on a linear basis, and it will be a limited result impact as we move along.” — Susanne Larsson, CFO or Finance Executive · 2026-07-17 The acceleration is a change of pace against that guidance, and management is entitled to claim it as execution. Less flattering is the capital allocation underneath. Full-year CapEx guidance was cut again, to SEK 2.5bn — a further SEK 0.5bn reduction on top of last quarter's cut from SEK 5bn to SEK 4bn. Meanwhile net working capital in the industrial business jumped to 38% of sales from 32%. The explanation is inventory buildup: “In this quarter, it's fully explained by the Specialized Industrial Solutions, building channel stock to cater for further growth then.” — Susanne Larsson, CFO · 2026-10-21 Building components inventory for growth while cutting capital expenditure is coherent only if the growth is real and the capacity can be flexed — SKF is doing the former and deferring the latter.

The margin promise that matters

To be brutal honest, we do not really see any real structural drivers why the business units in Specialized Industrial Solutions should operate at a lower profitability margin than the Bearing Solutions business. They should be roughly on par, and that is what we are aiming for in our long-term plans.
Rickard Gustafson, CEO · 2026-10-21

That is the single most consequential sentence on the call. It commits SIS — currently 4.8pp below Bearings — to close the gap against a 17% adjusted operating margin target for the separated industrial company. It also reverses the more hedged framing from a year ago, when SIS margin improvement was offered as evidence of "the start of the self-help and the repricing story." The claim now is structural parity, not improvement. Analyst Andre Kukhnin pushed exactly this question and got a straight answer — which is rarer than it should be on this call.

The answers that didn't come

What did not get answered was anything about 2027 input costs. James Moore laid out specifics — 100% freight rates, 30% steel, higher oil — and asked for quantification. The CFO declined, repeating that price had largely compensated. Asked about the first half of 2027, the CEO was blunt: “I have no comments on 2027 at this stage.” — Rickard Gustafson, CEO · 2026-10-21 Given the company is simultaneously raising separation-related one-off costs from SEK 2.5bn to about SEK 3bn and re-basing the whole cost structure, the refusal is understandable; it is also the second consecutive quarter where Cost inflation has been the top company keyword and the answer has been "we monitor, we price selectively." The Q4 guide was the same shape as always: sales to strengthen "somewhat," market environment similar, volatility noted. This is the identical construction used in January's "somewhat higher sales growth" guide and again in July. Analysts asked twice, three times, to quantify it; they were told it is "the best view that we can provide." That is a pattern, not a stumble — SKF guides in adjectives.

Where this leaves the equity story

SKF is asking the market to re-rate a smaller, higher-quality industrial business while handing shareholders a car-parts spin — and it is doing so with a commercial execution argument (price/mix plus 3.3pp of organic contribution), a cost argument (rightsizing six months early), and a growth argument (Magnetics and Aerospace, both tied to contract manufacturing unwinding and to the data-center capex cycle). Two of those three are verifiable and delivered. The third is a SEK 1.5bn business being asked to carry an outsized narrative weight against a tape that has spent ninety days punishing data-center leverage. The burden of proof for December is not the 16.3% margin. It is whether SIS can actually converge on Bearings — and whether management will ever put a number on 2027.