AQ Group: The Transformer Boom's Ugly Secret — Growth Arrives With the Lowest Margin Mix

A 21% sales beat, a 10% EBIT margin, and a gross margin down three points — the AI electrification wave is real, but it is buying the heaviest, cheapest parts of the revenue line

AQ.ST · Earnings Call · 2026-10-15

Aggregate strength, concentrated growth

AQ Group is a Swedish serial acquirer of niche manufacturing plants — wire harnesses, sheet metal, and power transformers — and Q3 2026 was its strongest top line in years. Net sales rose 21% to SEK 2.5 billion, operating profit climbed 36% to SEK 259 million, and the pre-tax margin landed at 10.0%, squarely "in line with our target," as CEO James Ahrgren put it. Nine-month EPS reached SEK 6.39 versus SEK 5.54 a year earlier, and the balance sheet is a net-cash position with a 68% equity ratio. What matters is composition. Every business area grew, but this is a tale of three boats: data centers, power grid, and defense. Everything else is "single-digit growth numbers across the board," in the CEO's words. Growth arrived through inductive components — the transformers, chokes and inductors that condition power — and management was unusually explicit about what drove it: “we grew 60% in inductive components year-over-year. That is data center growth, I can say.” — James Ahrgren, Likely CEO or senior executive · 2026-10-15 The company's own keyword set for the quarter confirmed the shift, with data centers and inductive components taking the top slots ahead of the older industrial-market language.

The mix problem hiding inside the beat

The bear case on the call was the gross margin: down three percentage points year over year and sequentially, to roughly 50%. Ahrgren's explanation is mix, not competition: “it's mostly product mix in this case. When we sell more transformers, there is more material in the delivery... It is copper, aluminum, and electrical steel.” — James Ahrgren, Likely CEO or senior executive · 2026-10-15 Transformers are heavy, metal-intensive products, so the fastest-growing segment carries the most pass-through material and the lowest incremental gross margin. That is the quiet tension — the AI electrification boom is arriving with a materially lower-quality revenue mix than the headline suggests. The 10% EBIT margin only held because cost control and full-throttle U.S. and China capacity offset the traditional European industrial holiday. Pricing is a non-event. “It's maybe a small pricing effect. I wouldn't say it's in line with inflation... it is a true picture.” — James Ahrgren, Likely CEO or senior executive · 2026-10-15 Volume and mix, then, not an inflation grab. Peer copper prices flagged the same raw-material headwind on Indian electrical-equipment maker Polycab's print, reinforcing that this is a sector-wide cost story rather than company-specific slippage.

Cleveland: the distressed-factory playbook again

The strategic event of the quarter was Norlake. AQ bought the assets of a Cleveland transformer factory that had run out of cash and stopped paying employees. AQ restarted it with the former owners, lent them money to get staff back, did quick summer diligence, and closed in September. The prize is doubled U.S. "Made in America" inductive-component capacity; the existing Virginia plant will do close to $50 million this year and is "basically full." Ahrgren frames the opportunity with disarming candor:

They had run out of money... we can come in, of course, with our strong balance sheet and support them with that, but also with new sales resources and customers that they maybe didn't have as well. I think it will be a good combination.
James Ahrgren, Likely CEO or senior executive · 2026-10-15

It is the same template that worked with mdexx, the earlier buy now declared fully integrated "in line with AQ margins." But note the humility on market share: “We believe we take market share from others, but that is nothing I can really prove.” — James Ahrgren, Likely CEO or senior executive · 2026-10-15 Refreshing — and a useful caution against over-reading the 17% organic number.

The tape disagrees with the order book

Here is where the narrative gets interesting. AQ says all data-center factories are "on full throttle," has prototyped a solid-state transformer for an 800V DC data-center architecture, and expects its UPS product to keep ramping — the unit ran 150 units against a 100-per-month capacity target, with the CEO answering a flat "yes" to whether 2027 keeps accelerating. Yet the market tape is voting the other way over the last 30 days: AI data centers is a 30-day decliner, down roughly 5.7% across more than fifty tickers, with data center power and high-bandwidth memory also soft, alongside large hyperscale customers and data center capacity selling off. The AI-hardware complex has cooled even as AQ's niche — the power-conditioning hardware beneath the racks — reports a demand surge. Either AQ is late to a fading theme, or it sits in the layer of the stack that still has to get built regardless. That divergence is the watch item.

What management quietly downgraded

The real tells are in the lowlights, and there were plenty. Acquired growth came in at only about 2%, below target, and management "lost some opportunities" in the quarter; the CEO hopes for 5% in Q4. For a company whose entire model is buying two to four factories a year, the M&A pipeline looked a little gummy. Add low capacity utilization in the Mexico and New York wire-harness plants, quality and productivity problems in the India inductive-components site (leadership changed), and Time 24 — bought only last quarter — described as "okay, but not as we want," with a delivery backlog. On-time delivery slipped to 91%. There is also a backlog story: data-center and defense deliveries that should have shipped in Q3 slid into Q4, and the CEO will not quantify the magnitude. He is confident, and promises recovery "by the end of this quarter four" — but a serial acquirer whose celebrated consistency ("profit every quarter since 1994") now has four or five turnaround projects running at once is a different risk profile than the one the marketing slide implies. And with electrification demand accelerating, execution capacity, not demand, is the binding constraint.

The bottom line

AQ Group is riding a genuine multi-year wave — power-grid and data-center electrification plus European defense — and it is buying distressed capacity to capture it. The EPS story is intact: management doubled earnings per share from 2021 to 2026 and wants to double again by 2031. But this quarter revealed the seam: the growth is concentrated in lower-mix products, several acquisition turnarounds are behind plan, and the tape has recently stopped rewarding the very theme AQ is selling. The next two quarters will show whether the Q4 delivery catch-up and the Norlake ramp are real operating leverage — or just more turnaround clock.