Semiconductor Automatic Test Equipment (ATE) is the set of machines that electrically validates every chip before it ships. The global ATE market generated roughly $7.5B in calendar 2024 and an estimated $9.3B in 2025 (+23%), per SEMI. est. Two companies — Japan-listed Advantest (58% share) and US-listed Teradyne (TER, ~37% share) — together control approximately 95% of the market. est. The remaining ~5% is split among smaller players including Cohu (COHU), which makes test handlers and inspection equipment rather than the testers themselves. ACM Research (ACMR) is not an ATE company — it makes wafer cleaning and wet-processing equipment; its financials are shown for completeness but it has zero ATE revenue.
Every AI accelerator, HBM stack, and advanced-packaged chiplet must pass an ATE system before it ships. AI chips are larger, hotter, and have more I/O pins than traditional chips, increasing test time and system complexity per device. Teradyne trades at 73× trailing earnings ($61.5B on $554M FY2025 net income). Cohu is losing money (-$74M FY2025) while ramping into the HPC handler opportunity.
ATE systems apply electrical signals to finished semiconductor devices and check whether the responses match expected patterns. Every chip must be tested — untested chips cannot ship. The ATE supply chain has three layers:
Revenue model: equipment sales (systems + upgrades) plus recurring service/spare parts. Teradyne's FY2025 service revenue was approximately $525M (~16% of total). Cohu reports recurring revenue at ~60% of total (Q1-2026), reflecting handler spare parts, refurbishment, and inspection services. Customers are chip manufacturers (Intel, Samsung, TSMC test floors), fabless companies' outsourced assembly/test partners (OSATs like ASE, Amkor), and increasingly hyperscalers testing custom AI silicon.
Source: Teradyne FY2025 earnings; Cohu Q1-2026 earnings call; Advantest investors guide.
Source: SEMI Equipment Market Forecast (Dec 2025); Teradyne Q1-2026 earnings; Cohu Q1-2026 earnings call; Advantest Q3 FY2025 results; Teradyne blog (Apr 2025).
Source: Advantest investor presentation (Jan 2026); RCR Wireless (Jan 2026); Teradyne FY2025 10-K; Cohu Q1-2026 earnings call.
| Measure | Demand side | Supply side |
|---|---|---|
| Global ATE market (SEMI) | $9.3B (2025), $9.8B (2026F) | Advantest + Teradyne combined ~$8.5B shipped (FY2025) est. |
| Advantest production | "Sold out through next year" | 3,000→5,000 units/yr (2026), targeting 10,000 |
| Teradyne Q1-2026 run-rate | $1.28B quarterly ($5.1B ann.) | ~$4.4B semi-test capacity (Q1 ann. × semi-test %) est. |
| Cohu HPC handler orders | +211% YoY computing orders | ~14 week lead time; $80–100M HPC rev (2026 guide) |
| Pricing direction | ASPs rising — both Advantest and Teradyne see higher system configs for AI; Teradyne gross margins stable at 58–61% | |
Pricing dynamics: ATE has a relatively fast supply response (quarters, not years). The real constraint is that the two dominant players ramp capacity at the pace they can maintain quality and engineering support. Teradyne's gross margins have been stable at 57–61% across cycles; Advantest's operating margins hit 29% in FY2025 (ended Mar 2026), up from historical mid-teens.
Contrast with front-end equipment: ATE grows proportionally with chip shipments, not with fab construction. When TSMC builds a $40B fab, it buys billions in litho/deposition/etch equipment during construction but buys test equipment continuously as production ramps. Test is ~8% of total semiconductor equipment spending ($9.3B of $125.5B in 2025). est.
Source: SEMI (Dec 2025); Teradyne Q1-2026 and FY2025 earnings; Advantest FY2025 results; Cohu Q1-2026 earnings call.
| Company | Ticker | Role in ATE | FY2025 Rev | Semi-Test Rev | Op. Margin | Net Income | Mkt Cap |
|---|---|---|---|---|---|---|---|
| Teradyne | TER | SoC + memory testers (#2 global, #1 US) | $3.19B | $2.33B | 20.4% (GAAP) | $554M | $61.5B |
| Advantest | ATEYY (ADR) | SoC + memory testers (#1 global, 58% share) | ¥780B (~$5.3B) | ~$4.9B (9mo ann.) est. | 29.3% | ¥161B (~$1.1B) | ~¥7.5T (~$51B) est. |
| Cohu | COHU | Test handlers + inspection (not testers) | $453M | N/A (handler co.) | -15.4% | -$74M | $2.7B |
| ACM Research | ACMR | Not ATE — wafer cleaning equip. | $901M | $0 (no ATE) | 12.1% | $94M | $6.4B |
Teradyne (TER) is the only large-cap US-listed pure-play on ATE. Semiconductor test was 73% of FY2025 revenue. The rest is Robotics ($358M, restructuring underway with 400 layoffs) and legacy product test. Q1-2026: $1.28B total revenue (+87% YoY), $1.11B from semi test (+148% YoY). Teradyne repurchased $702M of stock in FY2025 and carries $200M short-term debt against $448M in cash + securities. Free cash flow ~$450M/yr. est.
Cohu (COHU) makes the handlers that feed chips into Teradyne/Advantest testers, plus inspection equipment. Revenue peaked at $638M (FY2022), fell 37% to $402M (FY2024) in the cycle downturn, and recovered to $453M (FY2025). The company has lost money for two consecutive years (-$70M FY2024, -$74M FY2025) and raised $288M in convertible debt in 2025. Cohu's Eclipse thermal handler is winning multi-unit HPC orders, and the HPC segment could reach $80–100M in 2026.
ACM Research (ACMR) makes wafer cleaning and wet-processing equipment for semiconductor fabs, primarily in China. It has zero ATE exposure. Its $901M FY2025 revenue and $94M net income belong in the semiconductor equipment: cleaning category, not test. Included only because the scan listed it.
Source: Teradyne FY2025 and Q1-2026 earnings; Cohu Q1-2026 earnings call and FY2025 10-K; ACMR FY2025 earnings; Advantest FY2025 results; market cap data Jun 2 2026 (stockanalysis.com).
| Metric | TER | COHU | ACMR (not ATE) |
|---|---|---|---|
| Share price (Jun 2 2026) | $392.62 | $57.59 | $91.92 |
| Market cap | $61.5B | $2.7B | $6.4B |
| Shares outstanding | 156.5M | ~47M (basic) / 52.6M (dil.) | 69.1M |
| Trailing P/E (GAAP) | 73× | N/M (loss) | 70× |
| FY2025 revenue | $3.19B | $453M | $901M |
| Q1-2026 revenue (annualized) | $5.13B | $500M | N/A |
| Price / FY2025 revenue | 19.3× | 6.0× | 7.1× |
| Price / Q1 ann. revenue | 12.0× | 5.4× | — |
| FY2025 net income | $554M | -$74M | $94M |
| FY2025 FCF | ~$450M est. | $11M ($32M ops - $21M capex) | Not disclosed |
| Net cash (debt) | +$248M | -$188M | +$846M |
| Total debt | $200M (ST, repaid Q1-2026) | $296M ($288M convert.) | $288M |
| Book value / share | $17.87 | $16.82 | $27.93 |
| Price / book | 22.0× | 3.4× | 3.3× |
Teradyne arithmetic: At $61.5B, reaching a 20× P/E requires ~$3.1B/yr in owner earnings (5.6× FY2025 net income). Q1-2026 GAAP net income was $399M, annualizing to ~$1.6B. If that run-rate holds, forward P/E drops to ~39×. Q2-2026 guidance ($1,150–$1,250M revenue) implies the pace roughly holds.
Cohu arithmetic: At $2.7B market cap on -$74M/yr net income, the price embeds a bet that HPC handler revenue ($80–100M in 2026) inflects into a larger, profitable business. If Cohu reaches $600M revenue at 15% operating margins (roughly its FY2022 profile), that is $90M operating income — $2.7B = 30× operating income. The $288M convertible adds ~5.6M diluted shares at unknown conversion price. est.
ACMR arithmetic (not ATE): $6.4B on $94M net income = 68× earnings. Guided 21–30% revenue growth to $1.08–$1.18B in 2026. Not relevant to ATE exposure.
Source: stockanalysis.com (Jun 2 2026); Teradyne FY2025 and Q1-2026 earnings; Cohu FY2025 and Q1-2026 earnings; ACMR FY2025 earnings.