# Teradyne: a record AI upcycle, priced as if it never ends TER, NASDAQ. Published September 25, 2026. Price as of the September 24, 2026 regular session close. ## Verdict: REDUCE at $387.70 Teradyne is an excellent business at the top of an exceptional AI cycle. The shares trade 64% above the Base value, and even the Bull value is 12% below the price. The buy zone starts at $153.54. - Bear value: $171.85 - Base value: $236.22 (downside 39.1%) - Bull value: $341.67 - Expected value (25/50/25 weighting): about $246.49 - Required margin of safety: 35% - Maximum buy price: $153.54 - Mid-cycle EBITDA implied by the price at 20x: about $3.0 billion (Base $1.9 billion) - Dividend yield: about 0.1% - Business quality: 82/100 - Valuation score: 18/100 - Confidence: 20/25, grade B Every AI chip, and every stack of high-bandwidth memory next to it, has to be tested before it ships, and Teradyne makes the machines that do it. Revenue has doubled in a year and the company is gushing cash. It is one of the best businesses in this series. The problem is the price: at $387.70, the market capitalizes a cyclical peak as if it were the new normal. At $387.70 on September 24, 2026, the shares trade 64.1% above my triangulated Base value of $236.22, a downside of 39.1% to Base. Even the Bull value of $341.67 is 11.9% below the price. The required margin of safety of 35% puts the maximum buy price at $153.54. For existing holders the framework verdict is to reduce. New money should wait. Route: SF-13 Industrials and manufacturing, semiconductor test equipment plus advanced robotics. Economic classification: Quality Growth (8/12, raw 9/12, capped because the current growth is cyclical). Primary method: mid-cycle EV to normalized EBITDA. Secondary methods: normalized free cash flow after stock compensation, and a historical multiple reference. Values are present fair-value estimates, not 12-month price targets. ## What Teradyne is Teradyne designs automated test equipment for semiconductors and electronics, plus collaborative and mobile robots through Universal Robots and MiR. Semiconductor Test is the engine: it tests compute chips (SoC), memory including DRAM, HBM and NAND, and integrated systems. About 70% of Q1 revenue and more than 60% of Q2 revenue was tied to AI demand. The character of the business is improving strategically but becoming more dependent on the AI cycle. Customer concentration also increased: in Q2, 40% of revenue came from Taiwan and 20% from Korea. Robotics remains comparatively small. ## The latest quarter | Q2 2026 | Result | | --- | --- | | Revenue | $1.329 billion, up 104%, above the $1.15 to $1.25 billion guidance | | Segments | Semiconductor Test $1.122 billion, up 128%; Product Test $107 million, up 26%; Robotics $100 million, up 33% | | Memory revenue | $212 million, a record | | Non-GAAP gross margin / operating margin | 59.8% / 33.7% | | GAAP / non-GAAP EPS | $2.38 / $2.47 (prior year $0.49 / $0.57) | | Free cash flow | $378 million in the quarter; $579 million in the first half | | Q3 guidance | Revenue $1.20 to $1.30 billion, gross margin 58% to 59% | Revenue was $2.676 billion in 2023, $2.820 billion in 2024 and $3.190 billion in 2025. Consensus expects about $5.14 billion in 2026, 61% growth, and about $6.26 billion in 2027. Management says rising wafer fab equipment investment sets the stage for continued growth in 2027 and beyond. ## Why current earnings cannot be capitalized The cash quality is excellent: first-half operating cash flow of $734.3 million less $155.4 million of capex gives $578.8 million of free cash flow, and after $42 million of stock compensation still $536.8 million. Stock compensation is modest, and Teradyne's non-GAAP earnings already include it as a cost. The issue is not quality but the cycle. Q1 and Q2 were records driven by AI compute and memory. Semiconductor test spending has always been cyclical, so the framework values normalized mid-cycle EBITDA and free cash flow, not an annualized peak half year. ## Balance sheet and shares Cash and investments were about $517 million with no debt drawn. Shares outstanding are 156.34 million, down about 2.7% in a year thanks to buybacks. At $387.70, market value is about $60.6 billion and enterprise value about $60.1 billion. ## Primary valuation: mid-cycle EV to EBITDA | Scenario | Normalized EBITDA | EV to EBITDA | Value per share | | --- | --- | --- | --- | | Bear | $1.55 billion | 17x | $171.85 | | Base | $1.90 billion | 20x | $246.37 | | Bull | $2.30 billion | 23x | $341.67 | 2027 consensus EBITDA is about $2.3 billion. The Base deliberately normalizes below that, to $1.9 billion, because 2026 and 2027 look like the peak of the cycle. The 20x multiple is still a premium, reflecting the structurally higher test intensity of AI chips, but far below the current trailing multiple of about 40x. The primary Base value is **$246.37**. ## Secondary valuation: normalized free cash flow Normalized sustainable free cash flow is estimated at $1.15 to $1.30 billion rather than mechanically annualizing the first half. At the $1.225 billion midpoint, less about $84 million of annual stock compensation, hard free cash flow is about $1.14 billion. At 30x, the midpoint of a 28x to 32x range, plus net cash, that gives **$222.25** per share. Without the stock compensation deduction, it would be $238.37. ## Historical cross-check | Multiple | Today | Five-year average | | --- | --- | --- | | Price to sales | about 13.8x | about 6.5x | | EV to EBITDA | about 41x | about 20.7x | | Price to cash flow | about 46x | about 25x | AI merits a structural premium, but today's multiples are roughly double several historical norms. A historical normalized reference gives about **$235**. ## Valuation triangulation | Method | Base value | Weight | | --- | --- | --- | | Mid-cycle EV to EBITDA at 20x | $246.37 | 50% | | Normalized free cash flow after stock compensation at 30x | $222.25 | 35% | | Historical reference | $235.00 | 15% | | Triangulated Base | $236.22 | 100% | The primary and secondary methods diverge by 10.3%, well below the framework's 20% threshold. All three methods point to roughly $220 to $250. The secondary and historical methods exist only for the Base case, so the Bear and Bull values come from the primary method alone. ## What is the market already pricing in? Enterprise value of about $60.1 billion is 9.6x 2027 consensus revenue and 26x 2027 consensus EBITDA. At the Base multiple of 20x, the price implies **mid-cycle EBITDA of about $3.0 billion**, roughly 30% above the 2027 consensus peak and about 58% above the normalized Base. The market is not just assuming the AI cycle continues; it assumes 2027-level earnings, or more, are the new floor. ## Bear, base and bull scenarios | | Bear | Base | Bull | | --- | --- | --- | --- | | Probability | 25% | 50% | 25% | | Normalized EBITDA | $1.55 billion | $1.90 billion | $2.30 billion | | EV to EBITDA | 17x | 20x | 23x | | Value per share | $171.85 | $236.22 | $341.67 | | vs $387.70 | -55.7% | -39.1% | -11.9% | **Bear:** AI test orders cool, memory normalizes and earnings revert toward the old cycle; a trough does not automatically get a trough multiple, so the multiple stays at 17x. **Base:** AI raises test intensity durably, but 2026 and 2027 prove to be a peak. **Bull:** 2027 consensus EBITDA becomes the sustainable level and the market keeps a premium multiple. The probability-weighted value is **$246.49**, 36.4% below the price. ### Sensitivity: primary value per share | Normalized EBITDA \ Multiple | 18x | 21x | 24x | | --- | --- | --- | --- | | $1.6 billion | $187.52 | $218.22 | $248.92 | | $1.9 billion | $222.06 | $258.52 | $294.98 | | $2.2 billion | $256.60 | $298.81 | $341.03 | No cell in the table reaches today's price. ## How the buy price is set Confidence B starts at a 30% margin of safety. Five points are added for semiconductor cyclicality and AI concentration, for 35%. A 35% discount to the $236.22 Base gives a **maximum buy price of $153.54**. ## Dividend The expected 2026 dividend is about $0.52 per share, a negligible yield at today's price but very well covered. Buybacks matter more than the dividend. Income is not a thesis driver. ## Reasons to own TER - **A strong position in semiconductor test.** One of two global leaders in automated test equipment. - **Direct AI exposure.** Test intensity rises with every generation of AI compute and high-bandwidth memory. - **Exceptional cash conversion.** $579 million of free cash flow in the first half, with modest stock compensation. - **A clean balance sheet.** No debt drawn and a falling share count. - **Optionality.** Product Test and Robotics add growth outside semiconductors. ## What could go wrong? - **Valuation.** Multiples are roughly double historical norms, and even the Bull value is below the price. - **AI concentration.** About 70% of Q1 revenue was AI-related. - **Cyclicality.** Semiconductor and memory test spending has always gone through deep downturns. - **Peak earnings.** Current results may overstate mid-cycle economics. A cyclical surge capitalized as permanent is the main thesis breaker. - **Robotics.** It must still prove durable growth and profitability. ## Management execution Q1 and Q2 both beat guidance by a wide margin, and capital returns have historically been strong. The main question is allocation discipline while the shares trade at unusually high multiples. Management credibility scores **17/20**, good to excellent. ## Stock Analyza scorecard | Business quality | Score | | --- | --- | | Moat | 18/20 | | Economic return | 17/20 | | Balance sheet | 14/15 | | Cash flow quality | 14/15 | | Growth and runway | 11/15 | | Management and capital allocation | 6/10 | | Dilution and governance | 2/5 | | Total | 82/100 | The valuation score is **18/100**: every method and every scenario sits below the price. | Economic classification | Points | | --- | --- | | Normalized return | 2/2 | | Revenue growth | 2/2 | | FCF per share | 1/2 | | Balance sheet | 1/1 | | Unit economics | 1/1 | | Repeat economics | 1/2 | | Reinvestment runway | 1/2 | | Raw 9/12, capped for cyclicality: Quality Growth | 8/12 | ## Confidence score | Confidence | Score | | --- | --- | | Data quality | 5/5 | | Predictability | 3/5 | | Valuation robustness | 4/5 | | Accounting transparency | 5/5 | | Scenario dispersion | 3/5 | | Total | 20/25, grade B | ## What I would watch from here ### Green - Quarterly revenue above $1.25 billion - AI demand strong and diversifying across customers - Semiconductor Test and memory test growing - Free cash flow above 80% of net income - Stock compensation below 10% of free cash flow, stable or falling share count - Robotics growing with margin progress ### Yellow - Quarterly revenue of $1.0 to $1.25 billion - AI demand still concentrated - Semiconductor Test flattening, memory test normalizing - Free cash flow at 60% to 80% of net income - Share count up 0% to 2%, mixed Robotics results ### Red - Quarterly revenue below $1.0 billion - A sharp correction in AI test orders - Semiconductor Test down more than 15% sequentially - Free cash flow below 60% of net income - Share count up more than 2%, or Robotics contracting again The next review point is Q3 2026 results, expected in late October 2026. ## The three most important thesis breakers 1. AI test orders rolling over sharply. 2. 2027 expectations reset materially lower. 3. Margins reverting toward old-cycle levels. ## Adversarial review The first draft's primary valuation and sensitivity table reconciled closely. Three things were corrected. Its weighted Base was $241.75, but it was rounded up to $245 to match the primary method, which the framework does not allow. Its secondary free cash flow value did not deduct stock compensation, which free cash flow adds back; deducting about $84 million a year lowers it from about $238 to $222.25. And its net cash used cash and short-term securities of about $355 million, while total cash and investments were about $517 million. The Base moved from $245 to $236.22 and the maximum buy price from $159 to $153.54. The draft's twelve-month price-touch probabilities were dropped as not part of the framework. The verdict is unchanged. The strongest counter-argument is that AI has structurally increased test intensity per semiconductor dollar, so historical multiples and old mid-cycle earnings understate Teradyne's new economics. If 2027 revenue exceeds about $6.3 billion and growth above 20% persists into 2028 with durable margins, a much higher value is defensible. That is the Bull case, and even the Bull value is below today's price. | Robustness test | Base value | | --- | --- | | Base | $236.22 | | No stock compensation deduction | $241.86 | | Normalized EBITDA equal to 2027 consensus, $2.3 billion at 20x | $261.81 | | $1.9 billion EBITDA at 25x | $266.60 | | Free cash flow $1.3 billion at 32x | $246.70 | Even generous tests stay more than 30% below the price. The publication gate status is PASS\_WITH\_WARNING: the main uncertainty is how much of the 2026 to 2027 earnings step-up is structural rather than cyclical. ## Final verdict: reduce at $387.70 Teradyne is a high-quality company in an exceptional AI-driven semiconductor test upcycle, but the framework does not allow a cyclical peak to be capitalized as a permanent run rate. **Verdict: REDUCE at $387.70, WATCHLIST for new money. Triangulated Base value $236.22, downside 39.1%, Confidence B (20/25). Required margin of safety 35%, maximum buy price $153.54.** For an existing holder, trimming locks in part of an extraordinary move while the business is at its strongest. New money waits for $153.54 or below, or for evidence that AI test intensity makes 2027-level earnings a durable floor. The main thesis breaker is not weak fundamentals today. It is the risk that a cyclical AI-driven earnings surge is being capitalized as permanent. ## More analyses on Patreon This analysis is free. Premium members get every in-depth analysis, the full archive and a vote on which stocks get analyzed next: https://www.patreon.com/stockanalyza ## Sources Primary sources include Teradyne's [Q2 2026 results](https://www.sec.gov/Archives/edgar/data/0000097210/000119312526321933/ter-ex99_1.htm), the [June 2026 Form 10-Q](https://www.sec.gov/Archives/edgar/data/0000097210/000119312526327715/ter-20260628.htm), the [Q1 2026 results](https://www.sec.gov/Archives/edgar/data/0000097210/000119312526188706/ter-ex99_1.htm) and the [full-year 2025 results](https://www.sec.gov/Archives/edgar/data/97210/000119312526034348/ter-ex99_1.htm), plus [Q2 2026 earnings call coverage](https://ca.investing.com/news/company-news/teradyne-inc-ter-q2-2026-earnings-call-highlights-record-revenue-and-robust-ai-demand-propel--4761569). Consensus estimates, historical multiples and the market price come from third-party data providers. The reference market price is $387.70, the September 24, 2026 regular session close. Balance sheet figures are as of June 28, 2026. **Disclaimer.** This analysis is provided solely for research and educational purposes. It is not personalized investment, financial, legal or tax advice. Values are estimates, not forecasts or trading instructions. Teradyne is exposed to semiconductor cyclicality, AI demand concentration and valuation risk. Investors can lose part or all of their invested capital. Framework: SF-13, semiconductor test equipment and robotics. Economic class: Quality Growth. Engine: normalized mid-cycle EV to EBITDA with free cash flow cross-check. Confidence: B/20. Data status: PASS. Adversarial gate: PASS\_WITH\_WARNING. Version: Master v3.1.