Henkel Finally Puts a Number on AI — Just as the Tape Stops Paying For It
A raised top line and 6.5% organic growth mask a second-half margin squeeze, a first-ever data-center revenue disclosure, and a management that has suddenly gone silent on 2027.
HEN3.DE · Earnings Call · 2026-11-10
The Number Nobody Had
For at least two years Henkel has sold investors a story about data centers without ever sizing it. On the Q3 2026 call, Christian Faitz of Kepler Cheuvreux asked the obvious question — how big is data centers/AI actually? — and got the most concrete answer Henkel has ever given.Context matters here. Group sales in the quarter were EUR 5.7 billion; EUR 150–180 million a year is roughly a rounding error at the group level. But the shape is what management wants you to buy: CFO Marco Swoboda noted the number has already doubled since last year, split across two sub-business units, sitting inside a Mobility & Electronics division that grew 12.1% organically, driven by semiconductor packaging and advanced thermal management as global data-center capacity expands roughly 25% this year. CEO Carsten Knobel was quick to insist the segment's strength was "broad-based" and not a single-theme bet — a tell that he knows how the market now treats single-theme bets.We do expect that we see a clear double-digit growth in the coming years... it is already clearly entering double 3-digit levels of sales. So it's not EUR 20 million. It's clearly this year more around EUR 150 million, EUR 180 million.
Guidance Up, Margin Down — By Arithmetic
Headline: very strong organic growth of 6.5%, with Adhesive Technologies at 7.9% and Consumer Brands at 4.8%, price +5.4% and volume +1.1% — the fifth consecutive quarter of positive group volume. The full-year organic range was lifted to 3–4%, and Knobel flagged "a good start into Q4" with October strong, leaving "a good chance" of the upper half of the range, "not ruling out the upper end." The offset is the margin. Adjusted EBIT for the group is now guided to "around 15%" — broadly consensus, but it implies a visibly weaker second half than first. The reason is the raw-material lag that Swoboda and Knobel have been pre-announcing since the summer: direct material prices are now expected up around 10% versus the prior-year average, upgraded from the high-single-digit framing of the August call, with the pass-through hitting Adhesives for roughly three quarters of the year versus one for Consumer Brands. Knobel was blunt that oil's move back toward triple digits is the driver, and that "we are now in a double-digit mode" on material prices. This is not a new concern, it is the same concern landing on schedule: “H2 from a raw material cost perspective will be hit much harder.” — Marco Swoboda, CFO · 2026-08-06 What is worth noting is how clean Q3 was. Unlike Q2, this quarter carried no working-day benefit and no forward buying — the two effects that management spent most of the August call defusing. The prebuying reversal analysts kept asking about in August — when Swoboda conceded “so far, we haven't seen that reversal happening” — Marco Swoboda, CFO · 2026-08-06 — still hasn't shown up. That is a genuine positive: the underlying demand is real, and the pricing execution is holding. “I don't expect yet pricing in Adhesives to fade until end of the year. We need that pricing in order to also sustain margins.” — Marco Swoboda, CFO · 2026-11-10What Management Stopped Saying
The most revealing shift is rhetorical. Six months ago, Swoboda was willing to quantify the acquisition contribution to 2027 — roughly EUR 1.6 billion of annual sales on a pro-forma basis, ramping toward at least EUR 2 billion by 2030, and at least a 10% EPS contribution by then. On this call, roughly EUR 2 billion by 2030 survived, but any 2027 number did not. Instead, Knobel inserted the same caveat four separate times, in progressively firmer language, as analysts tried to triangulate whether the implied Q4 exit rate meant acceleration or deceleration:When David Hayes of Jefferies joked about having seven questions, Knobel pushed back hard. That defensiveness is the story: the guidance raise is real, but the company is deliberately refusing to let the market annualise it. In an environment where Knobel himself notes oil moved from the 70s into the 100s within weeks of the half-year call, the refusal is defensible — but it is also a step back from the transparency of August.We don't want to irritate or confuse anybody. Therefore, we have said that the exit rates which you maybe calculate based on the guidance, what we just gave that this is not irritating and making wrong conclusions.