Cohu May Be Rich on Cash Flow, Cheap on Sales

Mixed valuation signals sit between cash-flow caution and sales resilience

Cohu now looks less like a plain semiconductor-equipment story and more like a valuation-method mismatch. After a 157.2% one-year return, the stock clearly reflects optimism around future AI test demand. A cash-flow model suggests the shares may be about 26.9% rich on current free-cash-flow assumptions, while the sales multiple looks closer to fair value. That gap is the core debate.

What investors are really deciding

The key question is not whether Cohu has an AI story. It does. The question is whether the market is paying too much for current sales before execution fully catches up with the narrative. Bears can argue that market-based multiples already look consistent with present fundamentals rather than an exaggerated growth scenario. Bulls can counter that recent operating improvement was real: Cohu reported second-quarter net sales of $149.0 million, gross margin of 45.4%, estimated test cell utilization increased sequentially to 80%, and raising annual AI-driven compute opportunity pipeline to approx. $850 million.

My working thesis is that Cohu is not obviously expensive on sales, but it may be expensive on hope. If the company keeps converting that pipeline into utilization, orders, and revenue, today's expectations can hold. If conversion slows, the first risk is less likely to be a collapse than multiple compression.

Cohu's sales story has some real operating support

What keeps the bull case alive is not just sentiment. Cohu's recent results show a few qualities investors usually want before trusting an AI growth story: acceleration, reasonably stable margins, and signs of a real market bottleneck.

Q1 to Q2 improvement looks meaningful

The move from first-quarter net sales of $125.1 million to second-quarter net sales of $149.0 million matters because it suggests Cohu is doing more than enduring a flat semiconductor cycle. Q1 also included approximately 60% recurring content and Gross margin of 46.3%. Q2 brought gross margin of 45.4% as sales climbed. If growth were coming mostly from low-quality or one-off business, the margin profile would likely look less stable.

The AI test bottleneck is becoming part of the story

The most important AI link is testing capacity. Management said estimated test cell utilization increased sequentially to 80% at the end of June, while the annual AI-driven compute opportunity pipeline to approx. $850 million moved higher from the earlier Increasing AI-driven compute addressable market estimate to ~ $750 million. That change matters because it shows management is updating the story as demand signals improve, not simply defending an old target.

If test infrastructure is tight and customers still want Cohu's solutions, current sales are less a ceiling than evidence that the company can capture part of a scarce resource.

Orders are starting to broaden the narrative

There is also fresh evidence that demand is spreading beyond a single platform or customer niche. In early May, Cohu said a leading semiconductor manufacturer has placed multiple orders totaling approximately $5 million for testing next-generation GaN power devices for AI data centers. The company also raised its FY26 high-performance computing revenue outlook to ~ $80-100 million. Bulls are not guessing blindly; the pipeline has been moving upward, and recent order activity suggests that motion is still intact.

The next move depends on conversion, not momentum

That leaves the setup conditional. After mixed valuation signals and second-quarter net sales of $149.0 million, the market is no longer paying for a concept. It is paying for conversion. The practical test is whether management can turn recent strength, stable margins, tighter test-cell capacity, a growing AI compute pipeline, and the leading semiconductor manufacturer has placed multiple orders totaling approximately $5 million into a clearly advancing forward narrative.

What would validate the story

• Management advances its FY26 high-performance computing revenue outlook to ~ $80-100 million to a higher range.
• Order intake and booking cadence show that utilization and pipeline gains are converting into visible revenue.
• GaN and high-performance computing activity start to look like a repeatable pattern rather than a single validation headline.

How to size it

• If management clears that execution bar, a starter position makes sense because the story now has a measurable checkpoint behind it.
• If the update falls short, patience is reasonable. After mixed valuation signals, this is not a trade that demands blind momentum adds.

What could break the thesis

• Management does not move meaningfully beyond its current HPC outlook.
• Pipeline commentary becomes less concrete while the market starts valuing the stock more on hope than demand.
• Utilization improves, but order conversion and booking cadence do not follow.

Near-term timing

The next clear decision point is the July 30, 2026 Cohu Reports Second Quarter 2026 Results. Until then, this remains more of a watchlist setup than a fully actionable trade.

Fairly valued on sales, but not yet a margin-of-safety story

Cohu is not cheap today. Even with mixed valuation signals after a 157.2% one-year return, the stock looks more fairly valued on sales than obviously discounted. The catch is that fair value does not mean the story is finished. Investors are paying for a business that is starting to work, not one that has fully proved it can compound from here.

Bulls have a real case because the operating picture has improved in ways that matter: broad-based improvement across end markets, estimated test cell utilization increased sequentially to 80%, and a raised FY26 high-performance computing revenue outlook to ~ $80-100 million. Bears will counter that strong expectations do not automatically create a margin of safety. That is the right debate, and it comes down to this: Cohu deserves attention, but not an unearned premium.