TER — Teradyne, Inc.
Price when we picked it (Jul 23, 2026)
$362.67
Target: $420.00 (+15.8%)
P/E Ratio
67.3
P/E Forward
35.3
Dividend
0.14%
Market Cap
$56.5B
EPS
$5.39
Consensus
Strong Buy
What they do
Teradyne (Nasdaq: TER) makes automated semiconductor test equipment (ATE): the machines that take every freshly manufactured chip and put it through millions of electrical tests before it can be sold. No chip in the world — the one in your iPhone, in a car, an AI accelerator, a memory chip — reaches the market without passing that exam. And in practice, that exam is given by only two companies in the world: Teradyne (US, founded in 1960 by MIT-trained engineers) and Advantest (Japan).
The 3 businesses (Q1 2026 figures):
- Semiconductor Test ($1.11 billion in the quarter — 87% of sales) — the core. Machines that test processors and logic chips (the ones in phones, cars and data centers) and also memory — including HBM, the stacked memory that feeds AI accelerators and is especially hard to test. Its customers are the world's big chip fabs and assemblers: the TSMC, Samsung, SK Hynix ecosystem.
- Robotics ($91 million) — Universal Robots (the Danish pioneer of 'cobots': robotic arms that work alongside people without safety cages) and MiR (mobile robots that move material through factories and warehouses). This is the part that has disappointed: it peaked in 2022 and has seen declines and layoffs since, though it has now grown for 4 straight quarters.
- Product Test ($80 million) — testing of systems and circuit boards, storage drives, and wireless testing (LitePoint).
Why it's exploding now: each generation of AI chips is more complex — more transistors, stacked dies, HBM memory bonded to the processor, networks connecting thousands of accelerators. More complexity = more test time per chip = more Teradyne machines. In Q1 2026, ~70% of its revenue was already tied to AI demand (up from ~60% the prior quarter): custom accelerators for the cloud giants, next-generation HBM memory, and networking chips.
Leadership: CEO Greg Smith (in the role since February 2023). Headquarters in North Reading, Massachusetts. ~6,600 employees worldwide, with major operations in the Philippines, Denmark (home of Universal Robots), Costa Rica and Asia.
Dividend (symbolic): $0.13 per share per quarter (raised from $0.12 in early 2026) = $0.52 a year = ~0.1% at today's price. It also buys back shares, but let's be clear: nobody invests here for the income — the bet is growth.
2026 context: a record Q1 ($1.28 billion, +87%, 18% above the previous record set back in 2021), 60.9% gross margin, and a company 'target model' for the full year of ~$6 billion in revenue with adjusted EPS of $9.50-$11.00 — nearly double 2025's sales ($3.19 billion). Next report: July 28, 2026.
Why we like it
TER at $362.67 means REINFORCING a position we already owned (bought July 6 at $386.58), now ~6% cheaper thanks to a drop across the whole chip sector — not the company itself. You're still buying the toll booth of the AI chip boom: you're not betting on NVIDIA, AMD, or the cloud giants' in-house chips winning — whoever wins, their chips have to be tested on machines like Teradyne's. Specific reasons:
- Real duopoly: in semiconductor test equipment, practically only Teradyne and Japan's Advantest exist. Designing these machines takes decades of accumulated technology, and customers don't switch suppliers lightly — an extremely high barrier to entry.
- AI multiplies the testing hours: AI chips are huge, stacked, and very expensive, and each generation needs more test time (HBM memory is the extreme example). In Q1 2026, ~70% of Teradyne's revenue already came from AI-tied demand, and sales grew +87% — the best quarter in its 66-year history.
- The year is set to nearly double: the company's target model for 2026 is ~$6 billion in revenue (2025 closed at $3.19 billion) with adjusted earnings per share of $9.50-$11.00 — the kind of jump almost no company this size pulls off.
- Elite profitability: gross margin of ~61% and adjusted net income equal to 31% of sales in Q1 — when the cycle cooperates, this business prints money.
- Wall Street keeps raising its bets: 'Strong Buy' consensus (11 of 12 analysts say buy), with recently raised targets — BofA to $525, Susquehanna and Cantor to $550. Honesty: the average target (~$417) sits only ~8% above today's price, and the stock already hit $488 in June before correcting — it's not a hidden bargain, it's an exceptional business in the middle of a rally.
Key Risk
Risks:
- Post-rally valuation — the #1: the stock is up more than 4x in a year (from a low of ~$89 to a high of ~$488, $386.58 today) and trades at ~72x trailing earnings. The analysts' average target (~$417) sits only ~8% above, and the market is jumpy: in April the stock fell 19% in a single day... after reporting a record quarter. At these multiples there is no margin for error.
- Brutal cyclicality: chip test equipment is one of the most cyclical businesses in existence — when fabs freeze their purchases, sales collapse (Teradyne billed $3.19 billion in 2025, half of what it projects for 2026: that's how violent the swings are, in both directions). The company itself admits 'limited visibility' into the second half and its Q2 guide is below Q1.
- Customer concentration: a handful of giants (the TSMC, Samsung, SK Hynix ecosystem and the big Asian assemblers) account for much of the order book; losing or delaying a single big program moves the quarter.
- Advantest fights for every contract: it's a duopoly, but a fierce one — historically Advantest has kept, for example, the testing of NVIDIA's GPUs; if it wins the new AI programs, the boom goes to the rival.
- Robotics has disappointed for years: Universal Robots went from $326 million in 2022 to ~$232 million in 2025 (-28%), with two rounds of layoffs; it's growing again today, but remains an unkept promise.
- Geopolitics: nearly all its customers manufacture in Asia (Taiwan, Korea, China); export controls, tariffs or a conflict over Taiwan would hit directly.
This is educational and informational content, not financial advice. Always consult a qualified financial advisor before making investment decisions.
Vectorial Data picked TER on Jul 6, 2026 at $386.58.
Full Research
Teradyne (TER) — Full Research
Price: $386.58 | P/E TTM: ~72x | P/E fwd: ~38x | Div Yield: ~0.1% | Market Cap: ~$60 billion USD
What is Teradyne?
The American half of the global automated semiconductor test equipment (ATE) duopoly, alongside Japan's Advantest. Its machines put every freshly manufactured chip through millions of electrical tests before it can be sold — no chip reaches the market without that exam. Founded in 1960 by MIT-trained engineers. Headquartered in North Reading, Massachusetts. ~6,600 employees (10-K as of Dec 31, 2025), with major operations in the Philippines, Denmark, Costa Rica and Asia. CEO: Greg Smith (since February 2023).
The 3 businesses (Q1 2026)
| Segment | Q1 2026 sales | What it is |
|---|---|---|
| Semiconductor Test | $1,111 million (87% of total) | Machines that test processors, logic chips and memory (including HBM for AI) |
| Robotics | $91 million | Universal Robots (collaborative arms, Denmark) + MiR (mobile robots) |
| Product Test | $80 million | System and board testing, storage, wireless testing (LitePoint) |
Latest quarter (Q1 2026, reported Apr 28, 2026)
| Metric | Q1 2026 |
|---|---|
| Revenue | $1,282 million (+87% vs $686M in Q1 2025; +18% sequential) |
| Record | Best quarter in company history — 18% above the previous record (Q2 2021) |
| Gross margin | 60.9% |
| GAAP net income | $398.9 million |
| GAAP / non-GAAP EPS | $2.53 / $2.56 (expected: ~$2.08; a year earlier: $0.75) |
| AI-linked revenue | ~70% (vs ~60% in Q4 2025) |
Q2 2026 guidance: revenue of $1,150-$1,250 million and non-GAAP EPS of $1.86-$2.15 — sequentially below Q1: demand is lumpy and the company admits limited visibility into the second half. Next report: July 28, 2026.
FY2026 target model: ~$6,000 million in revenue and non-GAAP EPS of $9.50-$11.00 — nearly double 2025.
Full year 2025
| Metric | FY2025 |
|---|---|
| Revenue | $3.19 billion (+13% vs $2.82 billion in 2024) |
| GAAP net income | $554 million |
| Q4 2025 | $1,083 million (+44%) — the quarter the AI acceleration kicked in |
The AI boom thesis
- Every chip is tested before it's sold — and AI chips (accelerators, HBM, networking) are the most complex ever built: more transistors, stacked dies, memory bonded to the processor. More complexity = more test time per chip = more machines.
- ~70% of revenue already tied to AI: custom accelerators for the cloud giants (the hyperscalers' own 'ASICs'), next-generation HBM4 memory, and networking chips.
- Duopoly with Advantest: decades of accumulated technology; customers don't switch suppliers lightly. Competition happens program by program (chip by chip), not on list price.
- Elite profitability when the cycle cooperates: ~61% gross margin, non-GAAP net income = 31% of sales in Q1 2026.
- Robotics as optionality: Universal Robots + MiR — small today (~7%) and diminished since 2022, but with 4 straight quarters of growth and levered to AI-driven automation ('physical AI').
Dividend (symbolic)
- $0.13 per share per quarter (raised from $0.12 in early 2026; declared May 7, 2026, paid Jun 12, 2026) → $0.52 a year = ~0.1% at $386.58.
- It also buys back shares. But the bet here is growth, not income.
Leadership
- CEO: Greg Smith (since February 2023; a company veteran who previously ran Semiconductor Test).
- Headquarters: North Reading, Massachusetts. ~6,600 employees; ~156.5 million shares outstanding.
Anchor Fact
No chip in the world can be sold without first passing an electrical exam, and the machines that give that exam are made by essentially two companies: Teradyne and Japan's Advantest. AI chips are so complex they need far more hours of testing than a normal chip — which is why Teradyne just broke its 66-year sales record with a +87% quarter.
Top 6 Risks
- Post-rally valuation — ~72x TTM earnings; the stock rose more than 4x in a year (low ~$89 → high ~$488) and fell 19% in one day after reporting a record quarter; average target (~$417) only ~8% above
- Brutal cyclicality — 2025 billed half of what 2026 projects; the cycle goes down just as fast; Q2 guide sequentially lower and limited H2 visibility
- Customer concentration — the TSMC/Samsung/SK Hynix ecosystem and the big Asian assemblers concentrate the orders
- Advantest — the other half of the duopoly; historically kept the testing of NVIDIA's GPUs; every new AI program is fought chip by chip
- Robotics disappoints — Universal Robots: $326M (2022) → ~$232M (2025), two rounds of layoffs (~25% and ~14%); recovering, but an unkept promise
- Geopolitics — customers concentrated in Taiwan/Korea/China; export controls, tariffs or a Taiwan conflict hit directly
Analyst Consensus
- Rating: Strong Buy (TipRanks: 11 Buy, 1 Hold, 0 Sell)
- Average target: ~$417 (~+8% from $386.58); range $270-$550
- Freshly raised targets: BofA $525, Susquehanna $550, Cantor $550
- Honest read: in June the stock traded (~$488) ABOVE the average target and has corrected ~20% from that high. The most aggressive bulls see $525-550; the average sees moderate upside. This is cycle momentum, not a bargain.
One-line thesis
The toll collector of the AI chip boom: whoever wins the race — NVIDIA, AMD or the cloud giants' in-house chips — every chip has to pass through the Teradyne-Advantest duopoly's machines, and the exam keeps getting longer and more expensive.
Research date: Jul 6, 2026 | Next review: Jan 2027
This is not financial advice.
This is not financial advice. Consult a certified financial advisor.
The author may hold positions in the securities discussed.
Past performance does not guarantee future results.