Addtech's SEK 55 Million Confession: A Failed Deal Unit Masquerades as a One-Off Gain
Behind a 17% EBITA headline sits a shuttered, underperforming acquisition — and, quietly, one of Europe's more interesting AI-adjacent industrial exposures.
ADDT-B.ST · Earnings Call · 2026-10-22
The one-off is the story, not the noise around it
Addtech reported what the CEO called “a solid second quarter with high activity and a strong order intake” — Niklas Stenberg, CEO · 2026-10-22: net sales up 8% (2% organic), EBITA up 17% to a 16.7% margin. Strip out the noise and growth was 11% with the margin up 0.3pp — respectable for a decentralized business area conglomerate of 150-odd niche companies turning over roughly SEK 23bn.
But the +17% is not the interesting number. The interesting number is the SEK 55m "positive one-off" inside Safety — and its composition. It is a SEK 64m revaluation of earn-outs against roughly SEK 9m of restructuring cost, booked because Addtech decided to shut one unit of a recently acquired company. Read plainly: a forward-leaning purchase-price model written on an acquisition thesis that did not work. Management did not hide this — Niklas Stenberg was unusually direct:
This unit has strongly underperformed. It is a quite small part of the overall company, but here when we bought the company, we thought we would have quite strong growth in that unit, and that has strongly underperformed.
That is the honest framing. The strategic read is subtler. Addtech's entire equity story is serial acquisition of founder-led niches funded by “own cash flow” — Niklas Stenberg, CEO · 2026-10-22, with contingent consideration used as a discipline device. An earn-out reversing at SEK 64m says the discipline mechanism worked — it clawed value back — but it also says at least one 2024-25 vintage deal was underwritten too optimistically on growth. The CFO was clear that Safety's underlying margin was 15.3% adjusted, so the business is fine; the specific bolt-on was not. The market should treat the SEK 55m as a one-time gain to ignore in the model and a data point to hold against future deal pricing.
An AI trade hiding inside a Swedish distributor
The genuinely new thread this quarter is how much of Addtech's incremental growth now traces to compute infrastructure. Data centers appear in the prepared remarks as a named driver for Electrification, Safety and Industry — not as a footnote. And in the Q&A, the sharpest end-market step-up in the whole portfolio got a startlingly specific explanation. Asked why telecom jumped in the Industry segment, Stenberg said:
“It is primarily one company that has a really strong development within, let us call it data communication. So in the end, it is relating to the AI growth.” — Niklas Stenberg, CEO · 2026-10-22
And, crucially, this is not a one-quarter pop: “that we foresee to continue for quite some time.” — Niklas Stenberg, CEO · 2026-10-22
This is where Addtech meets the tape — and where the contrast becomes useful. AI data center is the single most crowded keyword in the global 360-day advancer list, carried by a 50-plus ticker complex spanning power, optics, cooling and semis. But over the last 90 days that same theme is a decliner — the price tape shows AI data center names down roughly 5-6% with 48 decliners against 6 advancers, alongside data center power and co-packaged optics rolling over. In other words, Addtech is reporting an acceleration in a theme that the market's own proxy basket has been de-rating for a quarter. For a company trading on organic growth and multiple expansion, riding a de-rating wave is a harder sell than riding a hot one. And the concentration matters: one company, one major customer, one long-dated project. That is exposure, not a franchise.
The consensus is real, though. Data centers were named by an extraordinary share of this week's reporters — CBRE on data center land sales, CCI, OVH, BE on 800V DC, NEX on high voltage, AKZO, THRM — so Addtech is neither ahead of nor behind a broader wave. It is a mid-cycle participant that has just been handed a visible, contracted tailwind. That is a better position than most of its Industrial Distribution peer group.
Energy's second half, and the comps game
The strongest forward-looking language was reserved for Energy. Order intake for electrical transmission and distribution developed strongly, and order book is described as clearly stronger than a year ago, broad-based across Sweden and the U.S., where Addtech is taking share from its Spanish subsidiary Arruti. Management committed to “a quite significant improvement the second half” — Niklas Stenberg, CEO · 2026-10-22. On the grid investment question the answer was unambiguous: “we can really not see any signs of a delay... rather the opposite.” — Niklas Stenberg, CEO · 2026-10-22
But note the repeated hedge on tough comps and its mirror, easier comps. Industry fell 7% on the anniversary of heavy sawmill deliveries; the sawmill customers remain unprofitable and are not investing, and management concedes no trend reversal yet, only "glimmers of hope." Energy's H2 improvement will also run into an exceptionally strong Q3 margin base last year. When a CEO leans this hard on comp arithmetic, it means the underlying demand picture is more mixed than the headline book-to-bill suggests.
Two risks deserve scrutiny. First, cost inflation — an explicitly named outlook risk, and one shared with SKF, Husqvarna, SK and Sonoco this week, so it is a real industrial-wide current, not a company excuse. Stenberg's answer was reassuring but soft: normal 80/20 volume-to-price has drifted to roughly 75/25, back-to-back supplier agreements protect the book, and “I don't see that there is a risk exposure in the order book that would mean that we cannot take out it on prices.” — Niklas Stenberg, CEO · 2026-10-22 Second, customers stockpiling against longer lead times — a classic pull-forward. Pressed on whether it had already happened or was still to come, the response was “impossible to answer that” — Niklas Stenberg, CEO · 2026-10-22, followed by the assertion that it is not material. The inventory build is small and short-dated, but it is exactly the kind of thing that makes a strong order intake print look better than the subsequent revenue print.
What actually changed
Not the model. Six business area verticals, four acquisitions adding SEK 500m this fiscal year plus Connectors & Switchgear at roughly SEK 130m this morning, a well-filled pipeline, return on working capital at 82%, and gearing deliberately low after the September dividend. The compounding machine is intact and the Connectors cluster alongside Allied Insulators is a sensible way to deepen the U.K. transmission position.
What changed is the composition of growth. Addtech now has a named AI-adjacent revenue line, a clearly improving Energy order book, and a freshly written-off acquisition unit reminding everyone that not every niche is a compounder. The stock's job from here is to convert that order book into second-half revenue while the most crowded theme in its own narrative de-rates around it — and to prove that a SEK 64m earn-out reversal is a footnote, not a pattern.