Huber+Suhner: A Temporary Dip Into a Multi-Billion-Dollar Window

Huber+Suhner's H1 2026 numbers looked like a warning. EBIT margins fell 110 basis points to 9.0%. Free cash flow flipped to minus CHF 25 million. The stock dropped 6% after the August 18 results.

The only problem is that the margin decline wasn't caused by a business losing steam. It was caused by a business building a factory to serve a market that Cignal AI estimates will exceed $8 billion by 2030.

Huber+Suhner is a Swiss manufacturer of power and data connectivity components — the physical-layer hardware that keeps aerospace systems, trains, and data centers connected. It has been growing steadily, profitable, and boring in the best sense. Then it found itself with a seat at the AI infrastructure table.

Here's the mechanism. Huber+Suhner makes Optical Circuit Switches — devices called POLATIS — that route data through fiber optic networks without converting it to electrical signals first. In AI data centers, where thousands of GPUs or TPUs need to communicate at hundreds of gigabits per second, this eliminates the power-hungry optical-electrical-optical conversions that waste energy and add latency. Google has issued seed purchase orders to Huber+Suhner alongside Lumentum and Coherent, signaling a shift from Google's own internal development to a competitive merchant supply base.

The company built a 3,000-square-meter factory in Pisary, Poland, to meet demand. The existing Poland site was already at maximum capacity. The costs for staffing, equipment, and inventory buildup for that factory are showing up right now, in the H1 2026 Communication segment, which swung to a loss of CHF 1 million — down from CHF 12.3 million of profit in the same period last year.

That is the entire margin story. One segment went to the red because the company chose to invest ahead of revenue. The other two segments did the opposite. Industry — aerospace, defense, test and measurement — expanded its EBIT margin 270 basis points to 19.6%, with order intake up 42%. Transportation added 60 basis points to reach 9.0%, driven by electric vehicle demand.

But investors read the headline margin and looked away. That's the disconnect.

The full-year 2026 guidance didn't budge. Huber+Suhner confirmed its EBIT margin target of 10.5% to 12% — the upper half of its medium-term range. It also raised its organic sales growth target to at least 10%. For context, the company's FY2025 EBIT margin was 10.5% on sales that declined 3.3% due to cyclical Communication demand. The 2026 guidance asks for double-digit growth and a higher margin simultaneously. That's not a company struggling through headwinds. That's a company with a loaded backlog — CHF 517 million, a record — and a book-to-bill ratio of 1.19 telling you orders are coming in faster than they're being fulfilled.

Let's look at the valuation against that trajectory. Huber+Suhner trades around CHF 172 per share, with roughly 19.2 million shares outstanding after a share cancellation program completed in early 2024. FY2025 net income was CHF 74.9 million, or about CHF 3.90 per share, implying a trailing P/E near 44x.

That trailing multiple looks expensive. But forward on the company's own 2026 guidance, the math changes. Organic sales growth of at least 10% pushes full-year revenue past CHF 950 million. An EBIT margin of 10.5% to 12% produces EBIT between CHF 100 million and CHF 114 million. Applying the current effective tax rate of roughly 14.4%, net income lands between CHF 86 million and CHF 98 million — EPS of CHF 4.46 to CHF 5.09. At the current price, the stock trades at 34x to 39x 2026 EPS.

Still not cheap. But that forward multiple prices in a company with double-digit organic growth, a 10%+ operating margin, zero financial debt, and a 74% equity ratio — plus an optionality play in OCS that hasn't yet reached its revenue inflection point. The company also announced an acquisition of Ingun, a German test and measurement company with 400 employees and double-digit million CHF revenue, expected to close in Q3 2026. Ingun isn't included in the current guidance, meaning it represents pure incremental upside.

The valuation question isn't whether Huber+Suhner is cheap. It's whether the forward EBIT margin holds and whether the OCS revenue materializes. Those are the two variables that matter.

On margins: the company has a track record. EBIT margin was 9.7% in FY2024 and improved to 10.5% in FY2025 despite declining sales. The H1 2026 dip to 9.0% is entirely attributable to OCS ramp costs in one segment, while the other two segments expanded margins by a combined 330 basis points. Full-year 10.5% to 12% requires only that H2 Communication improves as OCS production scales and fixed costs are absorbed over higher output — which is the standard operating curve for a manufacturing ramp. The company's existing Krzeszowice facility in Poland has been running at capacity; the new Pisary facility opens bandwidth. This is a capital-deployment problem, not a business-model problem.

On OCS revenue: the market Cignal AI projects at over $8 billion by 2030, growing at more than 30% annually, is real. Near-term volume is concentrated in Google's TPU networks — the only volume OCS application today. But the next wave, Scale Up in GPU networks, is expected to begin contributing significantly in 2028. Nvidia's upcoming Rubin Ultra NVL576 architecture plans to integrate OCS, and analysts say this market can "rival and ultimately exceed" the current TPU deployment. Huber+Suhner isn't the only player — Lumentum, Coherent, and more than 20 other vendors are in this space — but it's one of three vendors that received Google's seed orders. That matters. Being on the initial vendor list of a hyperscaler is a credible design win.

The risk is concentration. Management acknowledged the OCS business currently depends on one major hyperscaler customer, with discussions with additional hyperscalers ongoing. If Google slows its rollout or shifts share to competitors, the OCS thesis loses its anchor. The Communication segment's traditional 5G and fiber markets are already weaker, so the fallback isn't strong. This is a real, not theoretical, risk.

There's also a liquidity angle. Net liquidity fell 31% to CHF 146 million in H1, and free operating cash flow went negative. The balance sheet is fortress-grade — zero debt, 74% equity — but the cash burn is real and will continue through the ramp. A CHF 146 million cushion is comfortable for a company of this size, but it's not bottomless.

The broader ownership structure adds a layer of conviction. Roughly 47% of shares are held by individual investors — predominantly Swiss family offices and legacy holders with long time horizons — and 43% by institutions. This isn't a stock that gets sold into a bad quarter because fund flows shift.

Here's the bottom line. Huber+Suhner is a compounder — the kind of business that grows 8-12% organically, holds a double-digit EBIT margin, pays a dividend, and reinvests in adjacent growth markets. Those businesses don't trade at single-digit multiples. At 34-39x forward 2026 EPS, the stock isn't cheap. But it's not expensive for what it is: a profitable, debt-free industrial platform with record orders, accelerating demand in defense and aerospace, an EV tailwind in transportation, and a verified foothold in a market that research firms project to triple from $0.56 billion in 2026 to $8 billion by 2030.

The H1 margin miss was the cost of entry. If the OCS revenue inflection arrives as the management timeline suggests — higher output in H2, significant contributions starting in 2027-2028 — the forward multiple compresses further while the growth rate accelerates. That's the setup. If it doesn't, the stock reverts to a slower-growing, narrower-margin connector business, and the current price is too rich.

The math, the backlog, and the ramp timeline line up. The question is whether one hyperscaler's demand curve is enough to carry the call.