Himax graduates CPO from validation to production ramp — while an AI-driven supply crunch lifts the whole boat

Q2 beats on revenue, gross margin and EPS; Q3 guidance of ~34% gross margin signals a structurally better mix as automotive and co-packaged optics both accelerate

HIMX · Earnings Call · 2026-08-06

Himax Technologies reported a quarter that beat its own bar on every metric — Q2 revenue of $227.4 million (+14.2% sequential, +5.9% year-over-year), gross margin of 33.1% versus ~32% guided, and EPS of $0.114 versus $0.086–0.103 guided. The beat was broad, but the story underneath is bifurcated: the company's mature-process display IC bread-and-butter is being squeezed by an AI-driven capacity crunch, while its two long-horizon bets — co-packaged optics (CPO) and WiseEye-powered smart glasses — are finally generating concrete volume milestones.

A rare taste of pricing power

The proximate cause of the margin upside is a market-wide artifact. As Jordan Wu put it, “The ongoing surge in AI demand continues to impact non-AI applications” — Jordan Wu, President and Chief Executive Officer · 2026-08-06, and it has rippled across the broader semiconductor supply chain, producing capacity constraints at foundry, packaging and testing facilities on the mature process nodes where many of Himax's products are manufactured. That is the capacity constraint theme — a concern that climbed to #7 in this quarter's keyword trajectory, well above where it sat in recent quarters. Globally, memory cost dynamics are a fixture of the tape; for Himax, the consequence has been higher manufacturing and procurement costs, extended lead times, and pricing adjustments that began taking effect in Q2 — a rare tailwind for a driver-IC company that usually discusses pricing pressure, not pricing power. The financials reflect it: operating margin nearly doubled sequentially to 10.8%, and Q3 guidance calls for roughly 34% gross margin, implying the mix keeps improving.

CPO: the long arc finally bends

The bigger story is CPO. Himax and partner FOCI have now put both product generations into engineering production ramps — “Both our Gen 1 product supporting 1.6T and 3.2T transmission bandwidths and Gen 2 product designed for 6.4T bandwidth have begun engineering production ramps as scheduled in the third quarter” — Jordan Wu, President and Chief Executive Officer · 2026-08-06 — with “we expect our shipments in 2027 to be significantly higher than in 2026, starting to make meaningful contributions to our financials” — Jordan Wu, President and Chief Executive Officer · 2026-08-06. This is a familiar promise with a firmer grip. A year ago, management was still framing 2026 as “the first year of mass production. But we don't know and we are not going to comment exactly when” — Jordan Wu, President and Chief Executive Officer · 2025-05-08. By the November call, the view slid to “meaningful top line and bottom line contribution starting in 2027” — Jordan Wu, President and Chief Executive Officer · 2025-11-06. Now that engineering ramps are live and the next bandwidth target is 12.8T, the timeline is holding. When pressed to define "meaningful," Wu doubled down:

Meaningful means it's not going to be like low single-digit percentage certainly not for the bottom line, right? So meaningful is something we can actually — when it happens, we can categorize it and elaborate further. It's not going to be like marginally negligible addition to our financials.
Jordan Wu, President and Chief Executive Officer · 2026-08-06

The market context is friendly. co packaged optics is one of the tape's strongest 360-day advancing themes — north of +100% aggregate return across a dozen names including ANET, AVGO, MRVL, LITE and CLS — so Himax's wafer-level optics bet with FOCI, which Wu argues is differentiated from molded-glass rivals, is riding a broad adoption wave rather than pushing against one. The CPO technology keyword has been a fixture of Himax's own trajectory for a year; what is new is the verb tense, shifting from "validation" to "ramp." Tellingly, the keyword bottleneck jumped to a momentum of 207 this quarter — a direct tell that the call's center of gravity was CPO scalability, not demand.

Automotive: the steady engine

Underneath the hype, automotive remains the anchor: "well over 50%" of Q2 revenue, growing double digits, with automotive Tcon (over 10% of sales, more than half from automotive) the fastest riser. The automotive business keyword sits near the top of this quarter's trajectory, and the sustained displays per vehicle theme — now averaging more than three and rising — underpins a dollar-content-per-vehicle story. LTDI solution is in mass production across multiple car brands, with 5–10+ chips per ultra-large panel, and Wu flagged HUD Tcon (dewarping, postcard-effect elimination) as a fast-growing design-win pool. Q3's ~34% gross margin guide implies this content-value compounding continues.

Smart glasses and the new antenna

Finally, the Smart glasses theme — a top keyword in Q1 2025 and resurgent now — got its first public validation: "a leading global brand just launched a smart glasses product powered by our WiseEye technology." WiseEye technology delivers always-on outward and inward sensing at a few milliwatts, and management now cites hyperscalers "who traditionally do not offer hardware products" entering the space, with some projects poised for mass production in 2027. One concrete data point: the WiseGuard module (the security-oriented spin of WiseEye) has been adopted by a U.S. customer with mass production slated for year-end. A couple of non-operating items cushion the lumpy Q3 (a ~$13 million annual bonus, $11.7 million immediately vested): the proposed divestiture of an equity-method investee should book a roughly $23–24 million pretax gain on fourth-quarter close, and a new Taiwan policy allows ~$11 million of income tax to be deferred a year interest-free. Net: this is a name where the long-term story — CPO optics, the AI-sensing stack — is finally intersecting with near-term financials. The beat was broad, the guide is strong, and the market's own tape is already bidding up the exact themes Himax sells into. Execution risk is real — CPO still requires customer mass-production kickoff — but for the first time in years, the company appears genuinely out of the validation proving ground.