Teradyne was started in 1960 in a Boston basement by Alex d'Arbeloff and Nicholas DeWolf, MIT classmates who couldn't get a reliable instrument to test electrolytic capacitors and decided the answer was to build one. By the late 1980s the capacitor tester had become the industry standard for testing every chip the semiconductor industry shipped — a position TER then sat on for two decades while the test-equipment cycle did what test-equipment cycles do, which is grind operators into the ground between bursts. The stock spent most of 2008-2017 under $20. The inflection wasn't a product announcement; it was a workload change. AI inference and training silicon turned every chiplet, every HBM stack, every networking ASIC into a multi-hour test event, and the Teradyne/Advantest duopoly quietly became the tax-collector on the entire AI compute buildout. Q1 2026 printed $1.282B at a 37% operating margin with the first merchant GPU production order and 70% of revenue tied to AI. The market gapped the stock -19% on April 29 reading the Q2 sequential guide as a late-cycle signal. It is reading the wrong company. The body argues with the multiple; the lede notes that the multiple is being applied to a 2009 ATE business that no longer exists.
Initiate a 3% portfolio position in TER at $340-$350 with a $290 hard stop and a structured add zone at $300-$310 (additional 1%) and trim zone at $385-$390 (reduce 25%). The thesis is a structurally winning AI test duopoly delivering 37% operating margins and a credible $6B FY26 target, bought after a positioning-driven -19% flush rather than a fundamentals downgrade. Time horizon is 3-6 months, anchored to the late-July Q2 2026 earnings print as the primary thesis-validation event and the late-October Q3 print as the kill-switch.
Teradyne is the cleanest pure-play in the AI infrastructure test layer. Q1 2026 was an inflection quarter — $1.282B revenue (+87% YoY), $2.56 non-GAAP EPS (+241% YoY), 37.1% operating margin, $200M FCF — and the operating leverage is now real and recurring. The bull case cleared 4 of 5 specificity items (catalysts, customer-segment math, capex/consolidation framing, falsifiable kill-switch); the missed item was a true M&A consolidation reference, which downgraded the natural Buy to Overweight. The April 29 -19.4% gap reflected late-cycle expectations being reset by a beat-and-don't-raise quarter, not a business break — JPMorgan's post-selloff Buy upgrade and ARK's anchored $722M position confirm institutional buyers remain. Forward P/E 36x is in line with semi-cap peers (ASML 32x, KLA 28x, Advantest 30x) and reasonable for a 28% ROE duopoly business at peak operating leverage. The risk debate converged on 2.5-3% sizing with $290 stop discipline — sufficient to participate in the +30% upside scenario (FY $6.5B revenue, EPS to $12, forward P/E compresses to 29x and stock prints $450+) while limiting drawdown to -16% if the late-July print confirms the H2 visibility weakness. The kill-switch — Q3 2026 revenue below $1.40B — is a clean, single-quarter test that resolves the thesis. Cross-ticker memory: the AMTM Overweight from April 30 underperformed in 1 day (alpha -2.5%); the lesson is to size into Overweight ratings appropriately and respect stops, which is encoded in this plan.
Quantitative Lane
Customer Concentration
Top 5 customers: 36% of 2024 revenue
Samsung Electronics: ~12.5% of consolidated revenues (direct + indirect via supply partners)
2025 disclosure: One direct customer cited at 19%; two specifiers (indirect) at 12% and 10%
Implied Top 3: ~40% of revenue concentrated in Samsung, SK Hynix, and one foundry customer (likely TSMC)
Read: Dominant single-event risk. If any one of the top three pauses capex for one quarter, Q2/Q3 2026 revenue prints take a 5-8% hit. The Q2 guidance step-down from $1.28B to $1.20B midpoint is consistent with one large customer pulling forward orders into Q1.
Backlog & Book-to-Bill — Not applicable
Teradyne discloses quarterly revenue and orders but does not publish a contract backlog or book-to-bill ratio. As a semiconductor test-equipment vendor on short-cycle PO procurement, backlog is not a meaningful forward indicator — quarterly orders growth and segment revenue trends are.
Catalyst Calendar
Catalyst nodes are color-coded by impact (red=high, amber=medium, tan=low). Events grouped by quarter in the table below.
Photon 100 (silicon photonics) launch window — New product cycle for optical interconnect test
Medium
2026-09-30
Omnyx (server board test) launch window — Adjacent expansion into board-level data center test
Medium
2026-09-30
Merchant GPU production ramp — First merchant GPU customer Q1 win — Q2/Q3 ramp confirms it
High
2026-Q4
2026-10-28
Q3 2026 earnings (estimated) - KILL-SWITCH — Must print >=$1.40B revenue to validate $6B FY
High
2026-12-15
China export rule update window — Annual Commerce Department review cycle
Low
Research Manager
Investment Plan — Overweight
Rationale:
The fundamentals here are unambiguous — Q1 2026 revenue +87% YoY, operating margin 37.1%, EPS $2.56 (+241% YoY), and a $6B FY target with $9.50-$11 EPS that management did not raise after a +22% guidance beat. The bull case is real: TER and Advantest control >90% of SoC and memory ATE TAM, AI compute test intensity is structurally higher this cycle, HBM and merchant GPU dynamics are net positives, and the forward P/E of 36x is in line with semi-cap peers (ASML 32x, KLA 28x, Advantest 30x). The bear's strongest counter — that beat-and-don't-raise plus collapsed buyback equals management seeing stock as fully priced — is the reason this isn't an outright Buy. The April 29 -19.4% gap was a positioning unwind, not a fundamentals downgrade, and JPMorgan's post-selloff Buy upgrade plus ARK's continued anchor position confirm the institutional buyer base is intact. But H2 visibility is genuinely limited per management's own words, and the implied second-half ramp (Q1 $1.28B + Q2 $1.20B + H2 needing $3.52B = $1.76B avg per quarter) requires order acceleration that has not yet been booked. The risk is asymmetric on a small EPS miss given the high TTM multiple (64x). Specificity-bar enforcement: bull cleared 4 of 5 (failed M&A consolidation item — listed product cycles instead), so natural Buy was downgraded one tier to Overweight. Overweight reflects: take exposure to a structurally winning AI infrastructure pure-play with conservative-guiding management, but size for the chance that the July/October prints reveal an order pause and the stock revisits the $260-$280 zone (the $209 200 SMA + cycle support midpoint).
Strategic Actions:
Initial position sizing: 3% of portfolio at $340-$350 — half size of a high-conviction position to reflect asymmetric drawdown risk
Add zone: $300-$310 on retest of April 29 capitulation low. Add another 1% there if print/news flow remains constructive
Trim zone: $385-$390 on a re-rating into the late-July Q2 print. Take 25% off to harvest the institutional FOMO bid
Hard stop: $290 close (above April 29 $306 capitulation low for tighter risk control)
Catalyst calendar: SK Hynix Q2 capex (~mid-July), Q2 2026 earnings (late July), China export rule changes, Bernstein/TD Cowen/Stifel/BofA conferences May 28-June 12
Kill-switch: Q3 2026 revenue (late October) must print >=$1.40B. Below that, position reduced to zero regardless of price
Trader
Transaction Proposal — Buy
Action: Buy
Reasoning: The Research Manager's Overweight maps to a half-size initial entry. Q1 2026 fundamentals are step-change positive: $1.282B revenue (+87% YoY), 37.1% operating margin, $2.56 EPS, $200M FCF in a single quarter. The April 29 -19.4% gap was a positioning event in a crowded long, not a thesis break — JPMorgan's Buy upgrade post-selloff and ARK's $722M anchor position confirm institutional buyers remain. Forward P/E 36x is reasonable for a 28% ROE duopoly business with peak operating leverage. Asymmetric drawdown risk justifies half size and disciplined add-zone discipline.
Entry Price: $340-$348 zone (current $345; willing to chase to $352)
Stop Loss: $290 close (above April 29 $306 capitulation low; ~16% below entry)
Position Sizing: 3% of portfolio initial. Add 1% on retest of $300-$310. Trim 25% above $385.
FINAL TRANSACTION PROPOSAL: BUY
Investment Debate
Bull Researcher
ROUND 1: Q1 2026 was the moment Teradyne stopped being a cyclical semi-cap stock and became the AI infrastructure tax-collector. Revenue $1.282B (+87% YoY, +18% QoQ); non-GAAP EPS $2.56 (+241% YoY); operating margin 37.1%; FCF $200M = 67% of TTM in one quarter. Full-year $6B target with $9.50-$11 non-GAAP EPS not raised after a +22% guidance beat — most conservative semi-cap mgmt putting a low fence around a number they will hop over.\n\nDated catalysts: (1) Q2 2026 earnings late July — first chance to raise/kill FY guide and quantify merchant GPU production ramp. (2) June 12 quiet-period start preceded by Bernstein (~May 28), TD Cowen (~June 4), Stifel (~June 10), BofA (~mid-June) — three weeks of bullish updates compressed into one window.\n\nCustomer-segment math: 70% Q1 revenue tied to AI; memory $203M (HBM/DRAM); SoC compute ~75% of mix; hyperscaler 2026 capex >$400B; HBM bit demand >65% YoY 2026; Teradyne+Advantest split >90% of $5.5B SoC ATE TAM + ~$2B memory test TAM. $6B revenue = ~50% combined market share, exactly where they should be.\n\nCapex/consolidation reference: NVIDIA Blackwell Ultra 2026 ramp; AMD MI400 H2; Google TPU v6, Meta MTIA v3, AWS Trainium 3 hit production test 2026. April 16 TestInsight + April 8 MultiLane JV ($165M combined) bolt onto AI-test-stack — high-speed networking + validation-to-production conversion. SK Hynix/Samsung memory expansion (Cheongju, Pyeongtaek) requires more memory ATE through 2027.\n\nKill-switch: Q3 2026 revenue (Oct print) <$1.40B = $6B FY target breaks = thesis wrong. One quarter, one number.\n\nApril 29 -19.4% drop = positioning unwind in crowded long, not business deterioration. JPMorgan upgrade post-selloff. ARK still owns 2.36M shares.\n\nROUND 2 (sharpening): On dated catalysts — late July Q2 print is the actual P&L event. Within window: (a) merchant GPU customer Q1 production order ships in Q2 per call, follow-on order visible by late July; (b) SK Hynix Q2 capex disclosure mid-July (consensus +40% YoY 2026 driven by HBM4). Two dated, testable catalysts in 6-week window.\n\nOn consolidation: NVIDIA Blackwell Ultra ramp peaks Q3 2026 per Jensen GTC commentary. Each Blackwell Ultra requires 2-3x Hopper test time due to chiplet complexity — embedded in $6B target, not story. Merchant GPU win = AMD MI400 socket = structural share gain back into market Advantest dominated.\n\nOn H2 visibility: 'limited' not 'absent.' Mgmt under-guided pattern: Q4 2025 actual $1,083M vs midpoint ~$880M (+23% beat); Q1 2026 actual $1,282M vs midpoint ~$1.05B (+22% beat). Two consecutive +20% beats means FY guide ~10% conservative. $6B becomes $6.5B; $9.50-$11 EPS becomes $11-$13.\n\nOn ATE cycle bear: prior cycles (2018, 2022) memory-led pure-cycle. This cycle AI compute-led, structurally higher attach rate (chiplet complexity, HBM stacking, advanced packaging), spending committed by hyperscaler capex pre-orders, not memory spot pricing. Different regime.\n\nOn insiders: CEO grant unlock cadence mechanical. Mercedes Johnson 10b5-1 filed sub-$100. Wrong tell.\n\nOn valuation: forward P/E 36x for 28% ROE business at peak operating leverage with duopoly market structure is appropriate. KLA 28x no AI exposure; ASML 32x slower growth. TER deserves premium.
Bear Researcher
ROUND 1: April 29 reaction was not positioning unwind — market correctly pricing a guide that does not support $6B narrative. Q1 $1.282B, Q2 guide $1.20B = sequential decline. To hit $6B: H2 needs $3.52B = $1.76B avg per quarter. Q1 $1.28B + Q2 $1.20B + H2 needs +47% sequential ramp from Q2. Doesn't happen in semi-cap without hyperscaler order surge that has not been booked. Mgmt said second-half visibility 'limited, particularly for merchant GPU and networking' = confession.\n\nSpecificity bar audit: (1) Bull gave conferences, not catalysts — sentiment events not P&L events. Real catalysts: Q3 print (Oct) + negative pre-announce from top-3 customer. (2) TAM math no source — HBM bit demand +65% whose forecast? Teradyne 10-K admits competitors with greater resources continue improvements. (3) Hyperscaler capex up does not equal ATE orders up linearly — test attach rate per GPU compressing 3 years as Advantest V93000 EXA Scale takes share at high end. Merchant GPU first production order is one customer one product not confirmed multi-quarter ramp. (4) Bull listed product cycles not consolidation. No foundry expansion pre-bid to TER. SK Hynix has been buying Advantest tools in volume. (5) Kill-switch — $1.40B Q3 testable, valid.\n\nBear case: top-5 concentration 36% with one customer ~19%. If Samsung pulls forward orders to Q1 (= $1.28B Q1 then $1.20B Q2), H2 must be net new bookings — call admitted low visibility. Insider selling steady through run $100→$400 — CEO Smith $223K Dec 2025. Mercedes Johnson $200K monthly drumbeat = informed indifference at best.\n\nBuybacks decelerated -$5.5M Q1 vs -$157M Q1 2025 and -$183M Q4 2025. TestInsight+MultiLane $165M well within historical buyback budget. Mgmt thinks stock expensive at $300+. TTM P/E 64x, P/Book 19x, EV/EBITDA 46x trailing.\n\nCyclicality is the killer. 10-K explicit on backlog cancellation. Late stage of AI ATE order cycle now 18 months old. Historical post-peak ATE corrections -30% to -50% over 2-3 quarters. H2 modest pause = TER trades $200 not $345.\n\nROUND 2: Bull's 'two consecutive +20% beats' is the trap. Beat-and-don't-raise = late-cycle signature — MU 2018, TER itself 2022. If order book supported $6.5B, mgmt would have raised; they didn't.\n\nBlackwell Ultra test-time argument selectively true. Hyperscaler ASIC test (TPU, MTIA, Trainium) attach rates structurally lower because hyperscalers do more validation in-house and use foundry test infra. As ASIC mix grows vs merchant GPU, ATE intensity per dollar of compute drops. AWS Trainium 3 + Google TPU v6 explicitly designed to displace merchant GPU.\n\nShare-gain at AMD: Advantest publicly disclosed 70%+ share at TSMC N3/N2 foundry test for advanced nodes. AMD MI400 production ~2-3M units annually — meaningful but not regime-changing. Order real; framing as 'structural share gain' overstated.\n\nInsiders 10b5-1 mechanical fine. But: buyback was $157M Q1 2025 at avg ~$130/share, dropped to $5.5M Q1 2026 at avg ~$310/share. Mgmt's revealed preference is not to buy stock at these levels. That's data.\n\nValuation: bull comparing peak-cycle multiples to other peak-cycle multiples. KLA/ASML trade where they trade because forward-cycle profiles clearer. TER forward 36x P/E assumes $9.50-$11 lands; if H2 disappoints and FY26 EPS lands $8 (still up massively YoY), forward P/E becomes 43x — 'buy at 36x' becomes 'sell at 43x' overnight. Asymmetric to downside.\n\nFinal bear position: Not a short — no fundamental break, AI structural story real. This is 'respect the chart, respect the cycle, respect the guide-and-don't-raise signal' trim. Take partial profits into May/June conference bounce, re-buy below $250 if Aug/Oct prints validate H2 ramp.
Risk Assessment
Aggressive
ROUND 1: Trader proposal too cautious. Duopoly business at peak operating leverage delivering 87% revenue growth in stock just positioning-flushed 19% on sequential guide still up 75% YoY. Textbook 'bought-bond' setup — institutional buyers will rebuild into May/June conference circuit and late-July Q2 print. Half-size at $345 leaves alpha on table. Right size 5-6% portfolio with same $275 stop, willing to chase to $360 on confirmed reclaim of $362 Bollinger upper band. ARK anchored at $87 cost not selling. JPMorgan upgraded. Asymmetry runs to upside: $6.5B FY (+20% beat pattern continued) takes EPS to $12, forward P/E compresses to 29x, stock prints $475+ on multiple expansion + EPS upside. +38% upside vs -20% downside = 1.9:1 reward-risk on thesis where bull cleared 4 of 5 specificity items. Take size. Conservative will say wait for retest at $300; by then gap fill done and we missed it.\n\nROUND 2: Conservative voice keeps invoking 'prior cycles' but data point that matters is prior cycles were memory-led pure-cyclical events at peak. This cycle AI compute infrastructure where spending committed by hyperscaler 2026 capex budgets already disclosed. Microsoft >$80B; Meta $60B+; Amazon $90B+; Google $75B+. Combined ~$300B+ compute capex with TER attached at test layer of every chip in stack. 'Late-cycle' framing assumes ATE follows memory pricing — it doesn't anymore. ATE follows compute volume.\n\nTo neutral: $390 trim fine but waiting $300-$310 retest is fantasy. JPMorgan, Goldman, desk-level bid all positioned to defend $320 (50 SMA + post-flush re-entry). Retest may not come. At minimum take half-size at $345 immediately — don't optimize for print that may not happen. Conservative position is to be in the trade, not waiting for permission.
Conservative
ROUND 1: Aggressive doing exactly what burned investors in late-2021 ARK names: confusing positioning bounce with fundamental floor. Anchor in certainty: (1) Q1 $1.282B; Q2 guide midpoint $1.20B = sequential decline acknowledged by mgmt. (2) Top customer concentration ~19% single account, 36% top-5; single capex pause from Samsung breaks H2 model. (3) Buybacks collapsed $157M to $5.5M QoQ — mgmt's own capital allocation says stock expensive. (4) TTM P/E 64x; even forward 36x assumes FY guide lands. (5) Insiders sold throughout run $100→$400; CEO December 2025 sale at $201 was discretionary timing, not pure mechanical. (6) Beat-and-don't-raise = late-cycle ATE signature — every prior cycle (2018, 2022) ended this way. April 29 -19% gap is warning shot not buying opportunity.\n\nRight action: wait. Let stock retest $300-$310 and prove floor before committing capital. Half size at $345 with $275 stop = 20% downside on position; portfolio impact at 3% sizing = -60bps tolerable until stacked against 200bps loss if stops out and restarts at -50% from April high. Size at 1.5% initial maximum, add 1.5% only on confirmed bounce off $300 with volume, require $1.40B Q3 print before adding to full 4% weight.\n\nROUND 2: Aggressive keeps citing hyperscaler capex as if converts 1:1 to ATE orders. It doesn't — call language ('limited H2 visibility') is direct evidence order book has not yet captured what bull thinks it should have. Hyperscaler capex up does not equal SK Hynix or Samsung capacity orders up immediately; 2-3 quarter lag between hyperscaler announcement and memory fab equipment booking. Some ATE order flow being delayed because customers waiting to see whether HBM4 spec freezes Q2 or slips Q3. That's actual H2 risk, not in aggressive model.\n\nTo neutral: meet at 2.5% initial sizing with tighter $290 stop (above April 29 low rather than below). -16% downside vs +25% upside to $432 retest = 1.6:1, acceptable given cycle-stage uncertainty. $390 trim correct. We do not need to be heroic; we need to be paid for the risk.
Neutral
ROUND 1: Both sides have points but both also emotional. Aggressive anchoring on bull case execution and ignoring cycle-stage risk; conservative anchoring on cycle stage and discounting structural AI compute story genuinely different from prior cycles. Truth: high-beta name (1.79) at multiple requiring execution into uncertain H2, with mgmt itself acknowledging limited visibility. Sizing at 3% is right number for position with 20% drawdown risk on stop and 30%+ upside on thesis confirmation. Aggressive 5-6% sizing assumes bull case mostly already de-risked by April 29 flush — it isn't, because late-July print could repeat same dynamic if merchant GPU follow-on doesn't materialize. Conservative 1.5% sizing too small to matter; if thesis works didn't make money, if breaks saved trivial amount.\n\nTrader's 3% with $275 stop and disciplined add zone at $300-$310 is balanced — harvests asymmetric upside while respecting cycle-stage risk. One tweak: tighten trim zone to $390 (below gap-down origin at $380 plus small premium) rather than $410 because if stock approaches $400 without confirming volume, likely getting set up for second leg down. Take profits earlier, give position less room to round-trip.\n\nROUND 2: Convergence: 2.5-3% initial sizing, $275-$290 stop, $300-$310 add zone (smaller add, 1% rather than 2%), $385-$390 trim 25%. Thesis-confirming catalyst is late-July Q2 print + SK Hynix Q2 capex disclosure mid-July. Until those land we are speculating on positioning. Position size should reflect that. Right answer: take the trade, take it conservatively, let catalyst window decide whether to scale in or scale out. Aggressive wins philosophical argument that we should be in the trade; conservative wins sizing and stop-discipline argument. Both lose if they refuse to converge.
Analyst Reports
Daily close (warm umber) with under-fill, volume bars in tan. Trailing range capped at 252 trading days.
Market Analysis
Market Report
Teradyne (TER) is the most violent chart in semiconductor capital equipment right now. The 30-day window is a textbook three-act structure: a parabolic 35% rip from $315 (April 6) to a $422.11 all-time high (April 24), a -19.4% earnings-day collapse on 13.07M shares (April 29 — 6.5x the 20-day average), and a +13% reflex bounce off $306 to $345 (May 1). The 50-day SMA at $329.80 is now the line in the sand: price closed $345.42 on May 1, having reclaimed the SMA on the bounce after losing it on the gap-down.
Trend structure. The 200-day SMA at $209.42 is far below price (+65% spread) — the long-term trend is unambiguously up; the 52-week low of $73.11 is +373% behind us. The 50-day at $329.80 is the relevant short-term pivot. The 10-day EMA on the most recent data print sat at $333, meaning the bounce has converged to short-term mean reversion rather than re-establishing trend. The 200/50 spread of ~$120 is historically wide for TER and reflects the AI re-rating that began mid-2025.
Momentum. RSI compressed from 77.6 (overbought, April 24) to 38.3 (April 29) in three sessions — a 39-point round trip rarely seen outside of single-stock blowups. Current RSI of 48.5 is neutral and uninformative. MACD turned negative on the April 29 break and has not recovered crossover on the bounce. Momentum is broken; the question is whether this is a pause inside a primary uptrend or a regime change.
Volatility. ATR was $18.37 pre-event — already elevated. Realized April 29 range was $33 ($335 high, $302 low) on a single session, so ATR-based stops anchored before earnings were instantly invalidated. Beta is 1.79 — this stock moves nearly 2x the market; sizing must reflect that.
Bollinger context. Pre-earnings, price tagged the upper band ($362) and rode it for four sessions — a momentum signature that almost always resolves with a violent mean-reversion move. The April 29 break sliced through the lower band ($277-area on extension), then snapped back. Mean reversion to the 20-period middle band ($312-ish updated) is the typical post-Bollinger-blowout playbook; price has overshot on the bounce.
Support / resistance. Key levels: $306 (April 29 low — must hold), $329.80 (50 SMA), $345 (current), $380 (gap-down origination), $400-422 (April high cluster). Reclaim of $380 reopens the gap fill toward $400+. Loss of $306 opens air to the $209 200 SMA. Risk-reward favors the bulls only if $306 holds.
Volume. The April 29 13M-share session was a capitulation-grade event: 6.5x average volume on a -19% move is a positioning flush, not a fundamentals downgrade. Cathie Wood's ARK funds reportedly hold 2.36M shares (~5% of portfolio); positioning was concentrated.
Tactical read. Chart is now in a tactical no-man's-land between $306 support and $380 resistance. A breakout above $380 with volume confirms the gap fill is in play. A break of $306 invalidates the bull thesis on momentum grounds and opens material downside.
Indicator
Value
Signal
Close (May 1)
$345.42
Above 50 SMA, below 50 EMA prior trend
50 SMA
$329.80
Reclaimed on bounce; pivot
200 SMA
$209.42
Long-term uptrend intact; +65% spread
10 EMA
$333
Short-term trend converging
RSI (14)
48.5
Neutral; reset from 77.6 in 3 days
MACD / Signal
Bearish crossover
Momentum broken on April 29
ATR
$18.37
Elevated; expect $20+ daily ranges
Bollinger Upper / Middle / Lower
$362 / $312 / $278
Mean reversion played; band expansion likely
Beta
1.79
High systemic sensitivity
Volume (April 29)
13.07M (6.5x avg)
Capitulation-grade flush
52-week range
$73.11 - $422.11
+478% off lows; -18% off highs
Social Sentiment
Social Sentiment Report
The retail/social narrative on TER is in reset. After 12 months of one-way "AI test pure-play" cheerleading that took the stock from $73 to $422 (+478%), the April 29 earnings reaction broke the spell and forced a recalibration that is still in progress.
Sentiment timeline. Pre-earnings (April 24-28): Cramer pumping the stock as a data-center build-out arcana play, Zacks #1 momentum rank on April 30, "rocketing higher" coverage on the +13.8% post-Q1 bounce attempt. Post-earnings (April 29 onward): "Why Teradyne Tumbled," "Stock Is Getting Hammered," "Expectations Issue" — the framing flipped from celebration to autopsy in 24 hours. By April 30-May 1, narrative had stabilized into "AI buying opportunity" framing led by Motley Fool and JPMorgan's Buy upgrade after the 20% selloff, which itself triggered a +6% pre-market move.
Cathie Wood / ARK as sentiment anchor. ARK holds ~2.36M shares (~$722M, 5%+ of portfolio) at an ~$87 cost basis. That's a 4x in 12 months and a still-massive unrealized gain even after the drawdown. Her position serves dual purposes: it's a sentiment tailwind on every dip ("Cathie's buying" headlines) and a sentiment tail-risk if she trims into year-end performance management. JPMorgan's flag — "Cathie Wood's robotics bet dips after 350% run" — is the kind of headline that frames the stock as crowded long.
The "expectations issue" narrative (Morgan Stanley) is the single most important sentiment frame to absorb. MS literally said it wasn't a miss — Q1 was a record, Q2 guidance midpoint of $1.2B is +75% YoY, full-year $6B target was reaffirmed — but investor optimism around AI networking had run ahead of management's willingness to raise. This is the textbook "high-bar stock" trap: results that would launch most stocks 10% sent TER down 19%.
Insider signaling. Director Mercedes Johnson has sold 625 shares every month for over a year — almost certainly a 10b5-1 plan, but a steady $200K-per-month drip. CEO Greg Smith has sold $223K (Dec 2025), $200K (Nov 2025), $161K (Oct 2025), $58-70K (Aug-Sep 2025). His sales cluster around grant unlock dates, but the cadence accelerated as the stock went vertical from $100 to $400. Officer Regan Mills sold three times in late Feb-April 2026 at prices from $290 to $342. The signal is not insiders dumping the top — sales were largely complete before the April peak — but it does undercut any "insiders see undervalued" pitch.
Retail crowd structure. The stock is a household name in WSB-adjacent forums and AI-themed subs after the 12-month run. The April 29 break was not just an institutional repositioning event — it triggered margin calls and stop-cascades visible in the volume signature. Social conviction is bruised but the "buy the dip on AI test" muscle memory remains the dominant retail reflex; the May 1 bounce reflects that.
Cathie Wood positioning risk. If ARK trims to manage portfolio concentration after a 350% run, that's a structural overhang. If she adds to the dip, that's a tailwind. Either way, the stock has become reflexively sensitive to ARK's flow, which adds intraday vol that's not in the underlying fundamentals.
Net read. Sentiment is bruised, not broken. The "AI infrastructure pure-play" thesis still anchors most coverage; the April selloff is being framed as opportunity, not regime change. JPMorgan's Buy upgrade post-selloff is the institutional capitulation signal; Morgan Stanley's "expectations issue" framing keeps the bear case soft. Sentiment is bullish-with-a-stutter — vulnerable to a single bad data point (China export news, AMAT/Advantest miss, hyperscaler capex cut) but not actively negative.
Sentiment Signal
Direction
Notes
Mainstream financial press
Mixed/Recovering
"Buying opportunity" frame dominant by May 1
Sell-side desks
Constructive
JPMorgan upgrade post-selloff; MS "expectations issue"
Cramer / TV punditry
Bullish
Active pumping pre-earnings; moderated post
ARK / Cathie Wood
Anchor long
2.36M shares, ~5% of fund, $87 cost basis
Insider transactions
Net seller
CEO + Director on regular sale cadence; no buys
Retail / social
Bullish-stutter
Buy-the-dip muscle memory intact
Earnings reaction tone
"Expectations issue"
Not a fundamental miss per consensus
Forward sentiment risk
Concentrated
Sensitive to ARK flow + China headlines
News & Macro
News Report
The macro and sector news flow over the trade window is dominated by three structural threads that all converge on Teradyne: the AI capex super-cycle continuing into 2026, a memory pricing/HBM tightness story that directly benefits TER's memory test segment, and a market regime where Fed-on-hold + elevated oil + Nasdaq-flat creates volatility around expectations rather than directional moves.
Macro tape. Fed kept rates unchanged for the third consecutive meeting (per the April 30 24/7 Wall Street note), oil prices were soaring on the same day, and Nasdaq closed virtually unchanged at 24,603. This is a "no-news-is-bad-news" tape: with the Fed pinned and crude pressing margins, single-stock surprises become the only source of P&L, which amplifies positioning unwinds (exactly the dynamic that hit TER on April 29). Nasdaq's flat print on the same session TER fell 19.4% means the move was 100% idiosyncratic — no sector flush, no macro excuse.
AI capex cycle. The structural story TER is riding remains intact. Hyperscaler capex (Microsoft, Meta, Google, Amazon) was tracking >$300B in calendar 2025 and analyst consensus has 2026 numbers higher. Jim Cramer's "data center buildout" framing of TER captures the consensus thesis: every NVIDIA Blackwell, every AMD MI400, every custom hyperscaler ASIC, every HBM3e/HBM4 stack must be tested by either Teradyne's UltraFLEX or Advantest's V93000. The duopoly economics are real. Gartner reported worldwide semiconductor revenue grew 21% in 2025 (per January 2026 release), and the AI-attached portion of that grew materially faster.
Memory / HBM tightness. Trendforce and KED Global flagged in late 2025 that Samsung and SK Hynix gross margins on memory will exceed TSMC's in Q4 2025, driven by HBM pricing. Teradyne's Q1 memory revenue of $203M (driven by HBM and DRAM test) is the direct beneficiary. Korean memory players are reportedly chasing 180T won profit on the AI memory boom, which translates to test capacity expansion at SK Hynix and Samsung — a TER tailwind running into 2026-2027.
China export controls. Notably absent from the April news flow is fresh China escalation. The April 16 TestInsight acquisition and April 8 MultiLane JV closed without obvious regulatory friction. Teradyne's ficonTEC partnership (a Chinese-owned independent unit since 2021) carries some latent risk. The semiconductor capital equipment industry remains the most exposed sector to US-China decoupling, and the next Commerce Department rule cycle is the unguarded flank for the entire group.
Sector cross-currents. The Wall Street research calls roundup on April 30 (mentioned alongside TER) covered AbbVie, Equinix, Kratos Defense, Meta, Palantir — a mix that signals capital is rotating across AI infrastructure, defense-tech, and data-center REITs. TER sits at the intersection of the first two. Equinix capex commentary will be a leading indicator for compute test demand in H2.
Tariffs / trade. The Q1 earnings call mentioned tariffs as a watch-item but not a current quarter headwind. The administration's posture on semi-cap tariffs is unsettled; any tariff escalation that targets equipment imports/exports between Korea, Taiwan, and the US would directly compress TER margins (the company has Asian manufacturing exposure for some test heads).
Robotics regime. Universal Robots and MiR (TER's robotics segment) revenue grew 32% YoY to $91M — meaningful but small (~7% of company revenue). The robotics narrative as an independent growth pillar remains the unfulfilled promise of the 2015 Universal Robots acquisition, but Q1 was the fourth consecutive quarter of sequential growth, which finally validates a multi-year reacceleration thesis. If FX cooperates (USD weakness), robotics could surprise to the upside in H2 2026.
News-driven catalysts already in motion. JPMorgan's Buy upgrade with attractive entry framing is the kind of mid-cycle sell-side capitulation that often precedes 2-3 quarters of basing. The Photon 100 (silicon photonics) and Omnyx (server board test) product launches mentioned on the call are real product cycles, not just earnings deck filler.
Net read. Macro is not the variable. The AI capex cycle is intact, memory is tightening, hyperscaler 2026 capex is up. The risks are stock-specific: expectations management on the back-half guide, China export rule cycle, and any indication the merchant GPU customer ramp slips into 2027. None of those are visible in the current news tape.
Teradyne is in the rare position of being expensive on TTM and fairly priced on forward — the bridge is the AI super-cycle being delivered in real-time. Q1 2026 was the inflection quarter. Every line item on the income statement is now operating in a different regime than the prior 12 months.
Top-line trajectory. Quarterly revenue progression: Q1 2025 $686M → Q2 2025 $652M → Q3 2025 $769M → Q4 2025 $1,083M → Q1 2026 $1,282M. That's +87% YoY in Q1 2026 and +18% sequentially after a +41% sequential Q4. This is not a rounding-error beat — it is a step-change in the demand environment that began in Q4 2025 and accelerated. The full-year 2026 target of $6B implies $3.65B over the remaining three quarters — an average of ~$1.22B per quarter, roughly in line with Q2 guidance midpoint of $1.2B. Management has not raised the bar despite the Q1 beat, which is what bruised the stock.
Margin structure. Gross profit Q1 2026 was $781M on $1,282M revenue = 60.9% GAAP gross margin (the company guides 58-59% non-GAAP for Q2, implying some reset). Operating income Q1 2026 was $476M = 37.1% operating margin — exactly inline with the 37.6% TTM operating margin from the fundamentals snapshot. EBITDA margin Q1 = 39.1%. These are software-company-grade margins on a hardware business, and they reflect the duopoly market structure (Teradyne + Advantest control >90% of SoC/memory ATE). Operating leverage from the revenue ramp is real: opex grew from $550M (Q1 2025) to $806M (Q1 2026) = +47%, while revenue grew +87%. That's a 40-point operating margin gap that drops to the bottom line.
EPS. Q1 2026 GAAP diluted EPS = $2.53 (vs $0.61 Q1 2025 = +315% YoY). Non-GAAP $2.56. Full-year non-GAAP target of $9.50-$11 implies the back half delivers $5.50-$7 in EPS, or ~$1.40-$1.75 per quarter average — roughly consistent with Q2 guide midpoint of $2.00. Forward EPS of $9.51 (per fundamentals snapshot) lands at the low end of the company range, suggesting consensus has not yet bridged the upside scenario.
Balance sheet. Total debt = $82M (essentially capital lease obligations); total equity = $3.14B; net cash position. Debt-to-equity of 2.621 in the snapshot is misleading because the "debt" is operating lease liabilities — there is no real leverage risk. Working capital = $1.16B. Current ratio = 2.15. The balance sheet is fortress-grade. The $200M debt repayment in Q1 retired the only material short-term borrowing. Tangible book = $2.58B = $16.50/share, vs price $345 = 21x P/TBV. The premium is justified only by ROE: 28.75% trailing — exceptional for a hardware company.
Cash flow quality. Q1 2026 operating cash flow = $265M; capex $65M; free cash flow $200M. TTM free cash flow per the snapshot = $298M, meaning Q1 alone delivered 67% of the trailing year. Cash conversion is improving rapidly with the operating leverage. Q1 working capital build was a $194M use of cash (consistent with revenue ramping), and inventory/receivables expansion is benign at this growth rate. Cash position dropped from $294M to $242M as TER repaid $200M of debt and paid $20M in dividends. Buyback was de minimis in Q1, freeing capital for the $165M TestInsight + MultiLane M&A spend in early Q2.
Capital allocation. Three signals worth flagging. (1) Buybacks materially decelerated — Q1 2026 share repurchase was -$5.5M vs -$157M Q1 2025 and -$183M Q4 2025. Either management thinks the stock got expensive or capital is being reserved for M&A. (2) Two M&A closings in the first 16 days of Q2: MultiLane JV (April 8) for high-speed networking test and TestInsight (April 16) for test validation/conversion software. Combined ~$165M. Both are AI-test-stack tuck-ins. (3) Debt repayment: $250M in Q1, retiring the $200M short-term borrow from Q3 2025. Capital is being deployed for strategy, not financial engineering.
Valuation. TTM P/E = 64.1x; Forward P/E = 36.3x; PEG = 1.37. P/Book = 19.3x. EV/EBITDA on TTM EBITDA of $1.16B and ~$54B market cap = ~46x. Forward EV/EBITDA on a forecast EBITDA approaching $2B (midpoint of margin guide on $6B revenue) ~27x. The forward multiple is roughly in line with semiconductor capital equipment peers (Advantest trades around 30x, KLA around 28x). The market is pricing in a successful $6B execution — not above. Bull case execution above $6B revenue with $11 EPS would compress forward P/E to ~31x; bear case revision to $5.5B with $9 EPS pushes forward P/E to ~38x. The valuation is leveraged to the FY guide, not asymmetric.
Concentration & cyclicality (from 10-K). Top 5 customers = 36% of 2024 revenue. Samsung ~12.5%. The 2025 stocktitan summary references one direct customer at 19% with two specifiers at 12% and 10% — concentration is high and concentrated in a small number of memory and foundry buyers. Combined with backlog cancellation risk (per Item 1A), the cyclicality risk is real even inside an AI super-cycle.
Metric
Value
Read
Q1 2026 Revenue
$1,282M (+87% YoY, +18% QoQ)
Step-change in demand
Q2 2026 Guide Mid
$1,200M (+75% YoY)
Sequential decline acknowledged
FY26 Revenue Target
$6,000M (~+88% YoY vs $3.19B FY25)
Aggressive but not raised post-Q1
GAAP Gross Margin Q1
60.9%
Software-grade
Operating Margin Q1
37.1%
Inline with TTM 37.6%
Q1 GAAP EPS
$2.53 (+315% YoY)
Operating leverage delivering
FY26 Non-GAAP EPS Target
$9.50-$11
Forward EPS $9.51 at low end
TTM P/E
64.1x
Expensive in absolute terms
Forward P/E
36.3x
Inline with semi-cap peers
ROE
28.75%
Exceptional for hardware
Free Cash Flow Q1
$200M
67% of TTM in one quarter
Total Debt
$82M
Net cash position
Top-5 Customer Concentration
36% (2024)
Samsung ~12.5%
Buyback (Q1)
-$5.5M
Materially decelerated
M&A Spend (early Q2)
$165M
TestInsight + MultiLane JV
Risk Factors (Item 1A)
Source: Teradyne 10-K filed Feb 20, 2025 for FY2024 (https://investors.teradyne.com/sec-filings/all-sec-filings/content/0000950170-25-023784/ter-20241231.htm)
Customer concentration. Top 5 customers = 36% of 2024 revenue. Samsung alone = ~12.5% (combined direct + indirect via TSMC/foundry channels). Loss of any major customer or a meaningful capex cut from Samsung/SK Hynix/TSMC has immediate, material P&L impact. The 2025 disclosure references one direct customer at 19% — concentration appears to have intensified.
Market cyclicality + backlog cancellation. ATE cycles are notoriously sharp. Customers can change delivery schedules and cancel backlog with limited notice. Even inside the AI super-cycle, Q2 2026 guidance step-down signals that the lumpy ordering pattern persists.
Single-source supply. Certain components are sole-sourced; any disruption from one supplier creates immediate manufacturing impact. ATE test heads have multi-quarter qualification cycles, so component swap is not a fast remedy.
Competitive pressure. Advantest is the primary direct competitor (effective duopoly in SoC and memory ATE); SPEA and emerging Asian players target lower-end and specialty test. Some have greater financial resources and continuous product introductions threaten cost/performance positioning.
Foreign currency. Robotics revenue is largely non-USD (Universal Robots is Danish-domiciled), while test is USD-denominated. USD strength compresses Robotics topline — exactly what hit FY2025 results.
Geopolitical / export control. Export restrictions on US semiconductor capital equipment to China have been the dominant regulatory risk for the entire group. Any escalation directly limits TER's addressable market in mainland China and Taiwan-China cross-strait flows.
Cybersecurity. Disclosure language is standard for the industry but ATE machines control billions of dollars of fab throughput; a successful attack at a customer site would have material reputational and operational consequences.