Chroma ATE
Designs and assembles automated test systems and instruments; sells capital equipment plus test solutions/software and service. Asset-light assembler (outsourced fabrication) with high gross margins; revenue is order/insertion-driven and cyclical, now riding a secular AI capex wave.
Earnings, margins, COGS & capex
Record FY2025 (revenue NT$28.31B +31%, net income ~NT$11.7B +121%, EPS NT$27.7 +122%) followed by an even sharper Q1 FY2026 that broke normal seasonality: revenue NT$11.8B (+73% YoY), EPS NT$9.12, gross margin 63%, operating margin ~40%, net margin ~33%. Note FY2025 reported net income/EPS include a ~NT$3.6B one-time gain from a Q3 residential-asset sale, so normalized earnings are lower than the headline. The step-change is AI-driven — SLT for AI/HPC silicon and AI-server power test — layered on a diversified test franchise (semiconductor, EV/battery/ESS, photonics, general instruments).
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~39¢ is cost of goods and ~21¢ operating expense, leaving ~40¢ of operating profit (~33¢ net).
Revenue trend
Margins
rising — second-highest quarterly GM on record in Q1'26, mix shift to high-value SLT/power test
expanding on operating leverage
expanding; FY2025 net income +121% (flattered by a one-time asset-sale gain)
up sharply
COGS structure
~37-39% of revenue (implied by 61-63% gross margin). COGS is components/instrument hardware and assembly labor; fabrication largely outsourced, keeping the model asset-light and margins high relative to a heavy-equipment OEM.
Capex
Not disclosed as a discrete figure. Two near-term draws: second-phase headquarters construction (funded partly by new long-term debt) and tooling for co-packaged-optics (CPO) test at manufacturing Stages 3-4. Historically modest capex intensity for an assembler.
Latest earnings
Beat — 'crushed seasonality' with record Q1 revenue vs. a normally soft first quarter
Constructive: management expects continued 2026 growth led by semiconductor/SLT and AI-server power; 'seriously considering' an upward SLT forecast revision; CPO ramping into an incremental line
- Q1'26 revenue
- NT$11.8B (+73% YoY / +38% QoQ)
- Q1'26 EPS
- NT$9.12 (+81% YoY)
- Q1'26 gross margin
- 63%
- Q1'26 operating income
- ~NT$4.8B (+122% YoY); ~40% operating margin
- Q1'26 net income
- NT$3.9B (+83% YoY)
- FY2025 EPS
- NT$27.7 (+122%, incl. one-time gain)
- FY2025 dividend
- NT$19.5 (~70% payout)
Growth drivers
- AI-server power test — the test-instrument/ATS line grew ~105% QoQ / ~145% YoY in Q1'26 (more than doubled sequentially); tied to the shift to high-voltage direct-current (HVDC) power architectures in AI datacenters, with orders placed for the full year
- System-level test (SLT) for AI/HPC chips — the top three AI/HPC customers (NVIDIA, AMD, Google) have adopted Chroma SLT; AMD increasing cycle times and orders; management 'seriously considering' an upward SLT forecast revision
- Co-packaged optics (CPO) test — emerging incremental driver; POs received on three of four insertions, fourth (4E) in pilot from June 2026, with larger volume expected from 2027
- Semiconductor test (burn-in, functional, SLT) broad recovery on AI/HPC demand (Q1'26 semiconductor-test revenue ~NT$3.4B, +19% QoQ / +31% YoY)
- EV/battery and energy-storage (ESS) test, plus grid/HVDC — structural electrification demand
- Photonics/optical and general electronics instruments — diversified base
Bull & bear
Chroma has become an indispensable, high-margin test supplier to the two hottest AI bottlenecks at once — testing the compute (SLT for NVIDIA/AMD/Google silicon) and the power (HVDC AI-server power) — with CPO test as a third emerging insertion, all on a net-cash balance sheet compounding earnings triple digits.
- Q1 FY2026 broke seasonality: +73% YoY revenue, EPS NT$9.12, 63% gross margin — evidence the AI ramp is accelerating, not peaking
- Designed into all three top AI/HPC buyers for SLT with AMD raising orders; management may revise the SLT forecast up
- AI-server power test revenue more than doubled sequentially (~+105% QoQ), with orders booked for the full year — a distinct driver most test peers don't have
- CPO test opens a genuinely new insertion category with POs already in hand on three of four insertions
- 61-63% gross margins, ~40% operating margins, net cash, ~70% dividend payout — quality compounding, not just growth
- Diversified base (EV/battery, ESS, photonics, instruments) gives a second leg if AI cools
A sub-$1B-revenue, Taiwan-only capital-equipment maker priced at ~100-110x normalized earnings (~62x trailing), whose growth now leans on a concentrated AI-server/SLT order wave that is inherently cyclical and squarely in the crosshairs of far larger rivals Teradyne and Advantest.
- Valuation (~100-110x normalized P/E, ~62x trailing, ~NT$832B / ~$28B market cap on <$1B annual revenue) demands flawless, sustained hypergrowth — and FY2025's headline EPS is flattered by a one-time asset-sale gain
- The incremental growth is concentrated in a few AI/HPC programs and hyperscaler power orders — a capex pause hits the fastest-growing line first
- Teradyne (TER) and Advantest (ATEYY) dominate high-end test and are expanding SLT; a determined push can compress Chroma's niche and pricing
- Capital-equipment orders are lumpy and cyclical; 'crushing seasonality' cuts both ways when the cycle turns — the ~NT$400-2,795 52-week range shows how violently the stock re-rates
- Taiwan concentration adds cross-strait geopolitical, FX and supply-chain tail risk
- No US listing means thinner US-investor access, liquidity and disclosure cadence than the US-listed comps
What it is worth
Peer-relative multiples (vs. TER, ATEYY, KEYS, COHU) cross-checked against growth/margins; TWSE:2360 quote
An AI-capex digestion or Teradyne/Advantest SLT encroachment stalls the top line; a high-multiple stock de-rates hard on any miss, with FX and cross-strait risk amplifying the drawdown.
AI-test demand stays strong but decelerates from Q1'26's +73%; earnings grow briskly while the multiple compresses toward a still-premium 40-60x as growth normalizes — total return driven by EPS, muted by de-rating.
Sustained strong AI-test growth (SLT forecast revised up, AI-server power more than doubling, CPO scaling) keeps earnings compounding fast enough to grow into and beyond the current multiple; margins hold at 60%+ gross / ~40% operating.
Market cap ~NT$832B (~$28B USD) on <$1B FY2025 revenue implies ~25x TTM sales (~29x FY2025) and a normalized P/E near ~100-110x — vs. a trailing P/E of only ~62x, the gap being FY2025's ~NT$3.6B one-time asset-sale gain that inflates reported EPS. The premium is defensible only if SLT + AI-power-test + CPO compound for several years; any AI-capex air-pocket or duopoly encroachment justifies a sharp de-rating, and the ~NT$400-2,795 52-week range shows the stock already re-rates violently. US-first note: this is context, not a buy/own recommendation; US investors have no ADR and only foreign-ordinary access. Not financial advice.
SWOT
Strengths
- Designed-in at the top three AI/HPC chip buyers (NVIDIA, AMD, Google) for SLT — a sticky, insertion-based revenue stream
- Rare dual exposure to both AI-compute test (SLT) and AI-power test (HVDC/server power) — most peers touch only one
- 61-63% gross margins with ~40% operating margins and a net-cash, effectively debt-free balance sheet
- Diversified test franchise (semiconductor, EV/battery/ESS, photonics, instruments) cushions any single end-market cycle
- Strong capital returns — ~70% payout, NT$19.5 dividend on record earnings
Weaknesses
- Cyclical capital-equipment demand — orders can air-pocket when customer capex pauses
- Customer concentration risk in the AI-server/SLT surge — a handful of hyperscaler/AI programs drive the incremental growth
- Sub-$1B revenue base — small vs. Teradyne/Advantest/Keysight; less able to absorb a large program push-out
- Extreme valuation (~100-110x normalized P/E, ~62x trailing) prices in sustained hypergrowth — little margin for a miss
- No US listing / no ADR — US investors access only via foreign-ordinary or funds; lower liquidity and disclosure cadence than US peers
Opportunities
- Upward SLT forecast revision as AI/HPC volumes ramp
- CPO (co-packaged optics) test as a brand-new test insertion category tied to next-gen optical interconnect
- HVDC / AI-datacenter power test as power architectures change with GPU density
- EV/battery/ESS and grid electrification as a second secular leg
- Content-per-chip growth as SLT coverage expands beyond flagship AI parts
Threats
- Teradyne and Advantest (the ~80%-share test duopoly) pushing deeper into SLT and moving down into Chroma's niches
- AI-capex digestion or a hyperscaler spending pause hitting the highest-growth line first
- Taiwan-centric operations expose it to cross-strait geopolitical and supply-chain risk
- FX translation (TWD strength/weakness vs. USD) on reported and USD-converted results
- A high-multiple de-rating if AI test demand normalizes faster than consensus expects
Moats, dependencies & bottlenecks
Moats
Moderate-to-strong sticky once qualified into a program, but re-competed each new chip generation Qualification into NVIDIA/AMD/Google SLT flows is hard-won and switching mid-program is costly, but not permanent.
Breadth across test domains (semiconductor + power + EV/battery + photonics + instruments) Diversification is a structural cushion rather than a pricing moat.
EV/battery/ESS, HVDC, CPO) First-mover in emerging insertions like CPO test buys a lead, but large rivals can invest to catch up.
Dependencies
Demand / customer concentration The incremental growth engine; a spending pause hits the highest-margin, fastest-growing lines first.
Secular demand driver Ties Chroma's power-test upside to datacenter power-design roadmaps it does not control.
Asset-light assembler dependent on component availability and pricing.
Geopolitical / concentration Cross-strait risk and single-geography manufacturing footprint.
Affects reported results and USD-translated comparisons for foreign investors.
Advantages
- Simultaneous exposure to AI compute-test (SLT) and AI power-test (HVDC) — a dual-driver few peers share
- Very high gross/operating margins for a capital-equipment maker (asset-light assembly model)
- Net-cash, debt-free balance sheet funding expansion and a ~70% dividend payout
- Qualified relationships with the three largest AI/HPC chip buyers
- Diversified end-market portfolio spanning semiconductor, EV/battery, ESS, photonics and instruments
Weaknesses
- Small revenue base (<$1B) versus the Teradyne/Advantest/Keysight peer set
- Growth concentrated in a few AI programs and cyclical capital-equipment orders
- Extreme valuation leaves no room for a demand miss
- Taiwan-only footprint and no US listing/ADR limit investor access and add geopolitical risk
Bottlenecks
- Test-tool build and delivery capacity to meet the AI-server power and SLT order surge without lengthening lead times
- Skilled test-engineering and applications talent to qualify tools into new customer programs
- Component/instrument-part supply availability during the AI demand spike
- CPO-test tooling investment and ramp timing across manufacturing stages 3-4
Top signals & trends
Top signals
Demand is accelerating ahead of the usual cyclical rhythm.
Guidance bias to the upside on the core AI driver.
Second distinct AI growth vector inflecting.
New insertion category converting from pipeline to orders.
Valuation and capex/leverage bear watching against execution.
Competitive encroachment into Chroma's growth niche.
Trends
More AI silicon and more complex parts mean more system-level test insertions.
Directly expands Chroma's power-test TAM.
Creates a brand-new optical-test insertion category.
Second secular demand leg beyond AI.
Amplifies both upside and downside; orders are lumpy.
Tail risk to a Taiwan-concentrated operating base.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Electronic components & instrument-part vendors Semiconductors, power components, precision instrument parts feeding Chroma's asset-light assembly model.
Mechanical and subsystem fabrication that Chroma assembles into finished test systems.
AI/HPC silicon leader; adopter of Chroma SLT for AI/HPC chips.
AI/HPC customer increasing SLT cycle times and orders.
Top-three AI/HPC buyer using Chroma SLT (e.g., custom AI accelerators).
Back-end test/assembly houses and chipmakers buying semiconductor, burn-in and SLT tools.
ESS and power-electronics manufacturers Buyers of Chroma EV/battery, energy-storage and power test systems.
US-listed leader in semiconductor ATE and SLT; roughly half of the ~80%-share test duopoly. Far larger; expanding SLT and system test.
The other half of the high-end test duopoly, called the 'ASML of test'; V93000 and T2000 SLT platforms target AI/HPC — direct overlap with Chroma's growth niche.
US-listed test handling, contactors and test/inspection; overlaps in back-end test and SLT-adjacent handling.
US-listed broad electronic test & measurement; competes in general instruments, power and EV/battery test.
US-listed wafer-level burn-in and test; overlaps in burn-in/SLT for AI and power devices.
NI, now part of Emerson, provides modular test/measurement platforms competing in general and EV/battery test.
US-listed; instruments and power/electronics test overlapping Chroma's general instrument lines.
Private (Italy) functional/board and SLT test — niche competitor in system-level and mixed-signal test.