Start here if you're new
what it is
Cohu sells the machines and services chipmakers use to test, move, inspect, and measure chips before they ship.
how it gets paid
Last year Cohu made $453M in revenue. test handlers was the main engine at $170M, or 38% of sales.
why it's growing
Revenue grew 12.7% last year. EDGAR shows latest-quarter revenue up 162% vs. prior year.
what just happened
Revenue reached $331M, but profit still broke the story with EPS at -$1.11 in EDGAR and a miss versus Yahoo’s last consensus print.
At a glance
B balance sheet — gets the job done, barely
25/100 earnings predictability — expect surprises
23.3x trailing p/e — priced about right
11.5% return on capital — nothing to write home about
xvary composite: 44/100 — below average
What they do
Cohu sells the machines and services chipmakers use to test, move, inspect, and measure chips before they ship.
Cohu’s edge is the stuff already sitting on your factory floor. The company says recurring revenue from replacement parts, service contracts, and software supported results while equipment sales stayed soft, helping it post three straight vs. prior year growth quarters and $453M in annual revenue. Once your testers and handlers are qualified in production, switching means downtime and lower yield (yield → good chips off the line → your fab makes more money), so you keep buying the follow-on parts.
How they make money
$453M
annual revenue · their business grew +12.7% last year
test handlers
$170M
+12.7%
replacement parts, service, and software
$115M
+12.7%
inspection and metrology
$78M
+12.7%
test contactors and interface kits
$60M
4.5%
automation systems
$30M
4.5%
The products that matter
hbm test handler
Neon for HBM
$11M sales in calendar 2025
This is the growth option on the page. It produced $11M in sales in calendar 2025, which proves customer adoption exists. It also shows how early the story still is inside a $506M company.
early traction
core automation equipment
Semiconductor Test Handlers
$0.45B · 89% of revenue
This is still the center of gravity. Roughly $0.45B of revenue sits here, so when customers cut test capacity spending, this is where you feel it first.
main revenue driver
support and other sales
Services & Other
$0.06B · 11% of revenue
This roughly $60M bucket helps, but it is not large enough to smooth a real downturn in semiconductor equipment spending. You do not have a big recurring-revenue shock absorber here.
limited buffer
Key numbers
$900M
2029 revenue goal
The bull case needs sales to almost double from the latest $453M annual revenue, so you are underwriting a full recovery.
15.4%
operating margin
Operations are losing money today, which means higher sales still have to turn into actual profit.
$285M
long-term debt
Debt is 18% of capital, which gives Cohu some balance-sheet room while it waits for the cycle to improve.
$1.20
2027 EPS est
That sits far below the $3.50 forward EPS consensus, so the market is debating two very different earnings outcomes.
Financial health
B
strength
- balance sheet grade B — adequate — nothing special
- risk rank 3 — safer than 50% of stocks
- price stability 30 / 100
- long-term debt $285M (18% of capital)
- net profit margin 17.8% — keeps 18 cents of every dollar in revenue
- return on equity 14% — $0.14 profit for every $1 investors have put in
B — functional but not a standout on the balance sheet.
Total return vs. market
You invested $10,000 in COHU 3 years ago → it's now worth $7,620.
The index would have given you $14,540.
source: institutional data · total return
What just happened
missed estimates
Revenue reached $331M, but profit still broke the story with EPS at -$1.11 in EDGAR and a miss versus Yahoo’s last consensus print.
EDGAR shows latest-quarter revenue up 162% vs. prior year, but says investments in future growth and soft equipment sales are still pressuring profits. Yahoo also shows the last reported EPS at -$0.15 versus a $0.06 estimate, so the direction is clear even if data vendors disagree on the exact quarter.
$331M
revenue
$1.11
eps
15.4%
operating margin
the number that mattered
The number that mattered was -15.4% operating margin, because revenue growth means less when the core business is still losing money.
-
cohu’s business prospects appear to be turning the corner towards cyclical recovery.december-period sales for the semiconductor capital equipment maker were slightly ahead of median wall street expectations.
-
they were also up vs. prior year for the third consecutive quarter.bottom-line performance has been spotty despite system equipment sales starting to improve a bit, as there were some new product launches and a fold-in acquisition in 2025.
-
the investments in future growth have pressured profits while equipment sales are still soft.
-
recurring revenue for items such as replacement parts, service contracts, and software have been a good source of support, accounting for 60% of the top line in the december period.still, it is the system sales that really drive profitability for cohu when the semiconductor industry is thriving. the company has only recently begun to provide back-end inspection equipment to the booming ai arena. most of its offerings are for chips used in cyclical industries, such as automotive, and for consumer goods.
-
hence, its prospects are tied to the global economy.
source: EDGAR and, 2026
Get this snapshot in your inbox
This page, delivered free — plus weekly updates when the numbers change. plain english, no spam.
weekly updates
earnings alerts
plain english
no spam
What could go wrong
the key risk is simple: Cohu needs the next upcycle and the HBM product to arrive in the right order. If hardware orders slow before Neon becomes material, the math gets ugly fast.
med
Capex whiplash
About 89% of revenue comes from test hardware and handlers. When chipmakers delay equipment orders, most of the business feels it at once.
That puts roughly $0.45B of revenue directly in the path of semiconductor spending swings.
med
HBM excitement outruns HBM dollars
Neon for HBM produced $11M in calendar 2025. That's promising. It is also still small next to a $506M company.
If HBM adoption takes longer than investors want, the growth angle stays too small to offset weakness in the core business.
med
Margin improvement stalls below the target
Gross margin sits at 42.7%, and the stated goal is around 45%. That's progress, but not a dramatic mix shift.
Even if management gets there, earnings still need better volume. Margin alone is not enough to fix a weak cycle.
med
Messy data keeps the discount in place
Long-term debt is $285M, or 18% of capital, and the page's own profitability figures need footnotes because net margin and EPS point in different directions.
When the numbers need explaining, investors pay less for the story. Small-cap cyclicals do not get the benefit of the doubt.
If orders soften, roughly $0.45B of hardware revenue faces pressure while the HBM product still represents only $11M of offset. That's why this setup needs timing as much as it needs technology.
source: institutional data · regulatory filings · risk analysis
Pay attention to
trend
Neon for HBM revenue after the first $11M
The next updates matter more than the first proof point. You want to see this move from interesting to material.
metric
Gross margin versus the ~45% target
At 42.7% today, every point matters. If margin stalls below the target, one of the few clean recovery levers stops working.
calendar
The next earnings release
Here's the thing: the street expects $0.50 of fy2026 EPS after a -$0.27 year. The next report tells you whether that rebound looks realistic or too generous.
risk
Customer capex signals from memory and logic makers
Cohu sells into production lines. If large chip customers pull back on test spending, the 89% hardware mix becomes the whole story fast.
Analyst rankings
short-term outlook
below average
momentum score 4 — in human-speak, analysts think the near-term odds look worse here than they do for most stocks.
risk profile
average
stability score 3 — this sits near the market's middle, not in the safe bucket.
chart momentum
top 20%
technical score 2 — the chart looks better than the income statement does right now.
earnings predictability
25 / 100
Low predictability means estimates swing around. If you own it, expect revisions, surprises, and a market that reacts quickly.
source: institutional data
Institutional activity
77 buyers vs. 84 sellers in 4q2025. total institutional holdings: 46.3M shares.
source: institutional data
Price targets
3-5 year target range
$12
$37
$28
current price
$25
target midpoint · 11% from current · 3-5yr high: $50 (+80% · 16% ann'l return)
Want the deeper analysis?
The full deep dive: dcf model, scenario analysis, competitive moat breakdown, and quarterly tracking — everything on this page, taken further.
see plans from $5/moThe deep dive