Bulten Sold Its Factories and Bought an AI Story

A manufacturer that just sold its manufacturing

The quarter itself is almost the least interesting thing in Bulten's Q3 call. The story is the surgery: over three weeks the Swedish fastener group handed its European manufacturing operations to Maelir on October 1, its Tianjin plant in China on October 16, and will transfer the Hallstahammar real estate in December. What remains is a distribution-led business — and management is explicit that the risk profile has changed, not just the revenue base.

Now we are in a pure trading model with very low fixed assets, which means that we are much more resilient to volume fluctuations and have a much more attractive business model with less risk today than what we used to have before the divestment of the factories.

That is the crux of the divestment: Bulten is trading a vertically integrated, fixed-asset-heavy model that punished it every Q3 — when automotive customers idled, the margins fell hard — for a low-capital trading model. The factory footprint going away is genuinely new for this company; the prior calls were about running those plants, not shedding them. The accounting is ugly up front — discontinued operations hit equity by roughly SEK -1,200 million, mostly impairment — but the expected Q4 cash effect is around SEK 545 million, which is large relative to a company with a market cap of about SEK 1.6 billion. The Maelir handover and the China exit are the largest pieces, and management says there is "nothing major in the transformation box that remains."

An 8% base, and a 48% optics problem

Continuing operations delivered SEK 831 million of net sales and SEK 67 million of EBIT, an 8.1% margin — up sequentially from 7.6% in Q1 and 8.0% in Q2. The more meaningful line for a business that has bled volume for years is that the top line is finally ticking up: “we have had a few years now of declining volume, so it's nice to see that the volume started to tick up again.” — Axel Berntsson, President and CEO · 2026-10-21

But the headline growth number is contaminated, and management knows it. The 48% year-on-year jump in "other revenue," with other industries reaching 33% of sales, is flattered by the base effect of last year's cyber attack at a large customer. Asked how much of the 48% is underlying, Axel Berntsson was candid: “A quite large portion of that is driven by the cyber attack. I think we still have a good growth number underlying, but it is no 40% growth, of course, that we are seeing.” — Axel Berntsson, President and CEO · 2026-10-21 So the underlying non-automotive growth is real but smaller than the print — a nuance worth holding onto when the divestment comps roll off.

The AI kicker — arriving late to a de-rating trade

What is genuinely fresh for Bulten is its AI space language. The CEO is careful to define it as infrastructure rather than models — energy supply for data-center buildouts and semiconductor-adjacent work — and says Bulten has won "some quite large new projects" there while broadening its verticals away from automotive. The India bet is the physical manifestation: a new Chennai factory for micro screws and micro fasteners, equipment installing in Q4, operational early next year, aimed squarely at consumer electronics and AI infrastructure production. The customer logic is simple: “When they grow, we typically grow.” — Axel Berntsson, President and CEO · 2026-10-21

Here is the contrast that matters. Globally, AI data center was one of the strongest themes of the last 360 days — 41 advancers against 13 decliners — but over the last 90 and 30 days it has flipped to a decliner, with the 30-day list showing 48 down names against just 6 up. The market has spent the past quarter selling exactly the exposure Bulten is now marketing. And data centers is no longer an exotic theme: this week alone it surfaced at a specialty chemicals company, several regional banks, a software infrastructure firm and a European cloud provider — consensus, not differentiation. Bulten's AI line is a legitimate operational fact, but as an equity narrative it is a small-cap arriving late to a crowded, fading trade.

The margin gap, and the targets that do not exist yet

The sharpest exchange came on the competitive gap. Bufab and Bossard earn materially higher gross and EBIT margins, and the CEO did not hide behind the question: “I do not like that question, but it is a good question... I think the main difference is the mix.” — Axel Berntsson, President and CEO · 2026-10-21 His answer — that non-automotive carries favorable margins, that Bulten can go "toe to toe" on service, that it can reach Bufab's levels but "will we do it? I do not know" — is honest, but it is an ambition, not a plan with numbers attached.

And that is the gap in the story: there are no financial targets. The board discussed new targets at its meeting the same day, and a Capital Markets Day is promised, but the CEO would not commit to levels: “until we have any financial targets, I will not speak about any kind of levels where that will end up or what ambition we have for that.” — Axel Berntsson, President and CEO · 2026-10-21 A name change is also in approval limbo. So investors are being asked to underwrite a transformed, factory-free, AI-adjacent industrial with a lower fixed-cost base — and no quantified framework for where margins or growth land. The C parts and distribution acquisition pipeline is the swing factor: management wants to buy distributors at the front end and selected make-to-spec suppliers at the back end to stop suppliers setting the margin. That is the right strategic instinct — but it is the kind of story that needs the numbers the board has not yet published.

Bottom line: the divestment is the real change and it is largely done — the Q4 cash and the cleaner, more resilient P&L are genuine. But the growth narrative around it is a blend of a solid underlying non-automotive business, a comp-flattered 48%, and an AI-infrastructure kicker whose global tape has already turned. Until the new targets and the CMD arrive, the market is being asked to price a transformation on adjectives and a pending board decision.