Cohu's 157% Run: Cash-Flow Valuation Looks Rich, Sales Momentum Looks Reasonable

After a 157% rally, Cohu is being valued on proof, not just promise

Cohu still has operating momentum, but the stock is getting expensive on expectations faster than those expectations are fully proved. After a 157.2% one-year return, investors are no longer paying up simply because Cohu has an AI-related story. They are paying for a path to larger, more durable revenue. That is a much harder bar.

Revenue improvement is real, and so is the repricing

Cohu just reported Q2 sales of $149.0 million. The company also returned 157.2% over the last year, which puts the current debate in context: this is no longer a sleepy turnaround being overlooked. It is a stock the market believes can become something more.

That is why the setup feels split. On sales, Cohu looks close enough to fair value. On cash-flow terms, the same analysis suggests the shares are richer. So the real question is not whether the story has improved. It is whether the improvement is already reflected in the price.

Why the bullish case still deserves respect

The bull case is not based on a vague AI label. It is based on a business that appears to be moving into a larger, more strategic revenue bucket.

A bigger addressable market helps the thesis

Cohu lifted its view of the opportunity to ~$750 million AI-driven compute TAM. It also raising FY26 high-performance computing revenue outlook to ~$80-100 million. That does not require perfect execution. It only requires that AI compute becomes a more meaningful part of the mix over time.

Recurring revenue and margin quality matter

In the first quarter, Cohu posted First quarter net sales $125.1 million , approximately 60% recurring Gross margin of 46.3%. That mix matters because recurring revenue at that margin level suggests the business is not purely dependent on one-off hardware cycles.

Then the trend accelerated. Cohu followed with second-quarter net sales of $149.0 million and second-quarter 2026 income of $14.1 million. That is the core bull mechanism: revenue is moving up, and profitability is starting to show through.

Why the bear case still matters

The bear case is less about weak demand than about timing. Even with a stronger second quarter, Cohu still reported GAAP loss was $12.2 million or $0.26 per share for the first six months of 2026. That keeps the stock in a proof phase rather than a fully mature profit-ramp phase.

If demand cools before the earnings model stabilizes, the market can cut the premium quickly. That is the main valuation risk after such a strong run.

Why "reasonable on sales" is not the same as a bargain

The distinction here is mechanical, not moralistic: a stock can look fair on one valuation lens and still not be cheap on another. On sales, Cohu looks close enough to fair. Our model points to a fair P/S of 5.3x, while the shares trade around 5.0x sales. But on cash-flow terms, the screen is different: using latest twelve month free cash flow of about $37.3 million, the DCF implies an intrinsic value of about $40.58 per share, which the same analysis says screens as around 26.9% overvalued versus the current market price.

That gap is the whole issue. Revenue credibility does not automatically translate into cheap equity when the cash-generation base is still modest.

Why the signals diverge

The revenue story is easy to admire. After the rally, Cohu still reported second-quarter net sales of $149.0 million, and the quarter also showed improvement in profitability. That is enough to keep the narrative alive.

But price is paid for cash flows, not narrative alone. The P/S fairness check looks closer to the market only because it implicitly accepts the growth and quality of those sales. The DCF is stricter because it starts from a smaller current cash-flow base and asks how much of the AI premium has already been earned rather than merely anticipated.

What investors are really pricing

After a move of this size, psychology matters. Recent buyers may be more willing to pay a rich multiple because they fear missing out, while earlier holders may be reluctant to sell into what they now see as a new regime. That helps explain why enthusiasm around AI-driven semiconductor testing can keep the stock supported even when the valuation inputs are not perfectly aligned.

If sales keep rising but cash conversion lags, the multiple can still compress without any quarter turning outright bad. That is why calling Cohu "cheap" simply because the sales multiple looks near fair can be misleading.

What would keep the story on track

The evidence base does not support every recent phrasing in the draft, especially claims about guided sales, test utilization, customer momentum, design wins, and balance-sheet strength. With the supplied sources, the clearest way to frame the setup is through Cohu's published operating improvement and its larger AI compute opportunity.

The proof points investors should watch

The bullish case remains credible if Cohu keeps building on second-quarter net sales of $149.0 million and ~$750 million AI-driven compute TAM. Investors should watch whether revenue growth continues, whether profitability stays healthier than the first-quarter baseline, and whether management keeps expanding the commercial footprint around high-performance compute.

The cleanest risk is simpler: if demand cools before the earnings model fully stabilizes, the market may stop paying a growth premium even if the company is still improving. That is the tension defining Cohu after a 157.2% one-year return: the operating story is getting better, but the stock now needs sustained proof.