Keysight Technologies Inc · Moat analysis

Report previewKEYS.USReport date Last updated
← Back to reports
Report contents
01

Industry landscape

Where the company competes: a handful of firms at the top, a crowd below

Keysight Technologies sells the instruments engineers use to prove a signal behaves the way the design says it should. The catalog runs from oscilloscopes, vector network analyzers (instruments that measure how a component responds across a range of frequencies), signal generators and spectrum analyzers to semiconductor test systems and network-visibility software. Buyers sit in chip design and fabrication, network equipment and handsets, aerospace and defense, automotive electronics, and wireless operators. Revenue therefore follows customers' R&D budgets, qualification schedules and production ramps, not the unit sales of the products being tested. The company's 10-K reports two segments, communications and electronic industrial, and the communications side has been the larger and the more cyclical of the two, tied to telecom capex and handset programs.

Upstream, the performance ceiling of a high-end instrument is set by high-speed data converters and by custom silicon. Keysight designs its own ASICs (application-specific integrated circuits, chips built for one job), which is unusual among instrument makers and lets the company reach new specifications before off-the-shelf parts get there. Precision mechanics, microwave components and software supply the rest. Converter suppliers carry some leverage, but none is large enough to dictate terms to a buyer of Keysight's size.

Downstream the balance tilts toward Keysight. Customers are few and large, and the engineer who specifies an instrument writes it into a test procedure; replacing it later means re-qualifying a design, re-buying calibration and rewriting automation code. Procurement can press on price at renewal, yet switching mid-program is rare. That workflow stickiness, more than brand, is where pricing power comes from.

Market position: strong, but unevenly so across sub-arenas. The core arena is precision radio-frequency and high-speed digital test, sold globally: instruments ship anywhere, while the customers who buy top-end models cluster in the US, Taiwan, China, Japan, South Korea and Europe. Service and calibration are the exception, since they must be local. That local footprint is itself a barrier — a customer with labs in five countries wants one vendor's calibration certificate in all five.

The arena is not one market, and this is where a casual read of Keysight goes wrong. Bench RF and microwave test is contested by Rohde & Schwarz, Anritsu and Tektronix alongside Keysight. General-purpose and mid-range instrumentation brings in National Instruments, now part of Emerson, plus a long tail of regional vendors. Semiconductor automated test equipment (the machines that test finished chips at the factory) is a different fight, run by Advantest and Teradyne. Network test and visibility has been a three-way contest with VIAVI Solutions and Spirent Communications, and Keysight's agreed purchase of Spirent in 2025 consolidates that corner. Leadership in one of these does not transfer automatically to another; the buying center, qualification cycle and rival set all change.

Apply the count-on-one-hand test to the top of the RF bench, and the structure passes. Four suppliers are credible for the most demanding measurements. Getting there takes measurement-grade accuracy, traceable calibration, custom silicon and customer qualification that runs for years. At the bottom of the market, basic oscilloscopes, bench supplies and teaching equipment, the supplier count runs to dozens and Chinese vendors such as RIGOL and Ceyear compete hard on price. The structure is barbell-shaped: an oligopoly at the top, something close to commodity competition below.

Two facts push against the comfortable version of that picture. First, the customer base is concentrated in a small number of very large technology and defense programs, so orders swing sharply when one program pauses; concentration among buyers offsets some of the concentration among sellers. Second, the mid-range is where volumes and learning accumulate. Chinese instrument makers have been adding bandwidth and features year by year, and the open question is whether they reach the specifications that govern 5G and 6G qualification and high-speed digital compliance. So far the evidence shows them competing mainly in price-sensitive, lower-bandwidth segments, which leaves the high end intact but not permanently sealed.

What would settle it: sustained share gains by low-cost vendors in instruments above roughly the mid-band, or evidence that Keysight and Rohde & Schwarz keep holding specification leadership generation after generation. The first would compress the oligopoly from below; the second would confirm that the barrier is engineering depth and calibration infrastructure, which take longer to replicate than a product spec sheet.

02

Competitive advantage

Does the company have a moat? Yes, but goodwill dilutes the protection

Keysight's instrument franchise earns the returns of a protected business. The consolidated company, at least since fiscal 2024, does not. The case for protection is narrow and hard to fake: gross margin stayed between 62.1% and 64.6% in every fiscal year from 2021 through 2025, including fiscal 2024, when revenue fell 8.8%. A supplier without pricing power concedes gross margin in a downturn, because that is when orders are scarce and discounts are the fastest lever. Keysight conceded almost none.

The case against a wide moat sits below the gross margin line. Estimated after-tax return on invested capital (ROIC: NOPAT divided by equity plus debt minus cash) fell from about 26.7% in fiscal 2021 to 11.6% in fiscal 2025. Nearly all of that came from the denominator. Invested capital rose from $3.56 billion to $6.59 billion, goodwill reached $3.44 billion by fiscal 2025, and NOPAT slipped from $0.95 billion to $0.76 billion.

The two facts fit together. A moat protects the price of an instrument; it does not protect the price paid for an acquisition. Keysight has spent years buying into adjacent test markets, and the goodwill from those deals now sits in the capital base that has to earn a return. The open question is not whether Keysight can hold price against Rohde & Schwarz or Anritsu, since the margin record says it can, but whether the revenue it bought earns anything close to the returns of the franchise that bought it. Two years near 12% do not settle that.

High margins alone would not prove a moat, because good execution gets copied. Concentration plus persistence does. The top of this market has a handful of credible suppliers, and a gross margin above 60% that survived a double-digit revenue decline is what entry barriers look like in a P&L: incumbents hold price because a new entrant cannot yet meet the specification, and customers cannot easily leave.

Market position: top tier in RF test, priced accordingly

Market position: Fairly Strong. Keysight's revenue sits overwhelmingly in the demanding end of the market, where the supplier count is small, and its pricing reflects that. A vendor ranked second or third in a market where buyers can shop freely would not post 62% gross margins for five straight years while revenue was falling. Rank itself is not measurable here: no public share series covers fiscal 2021 through 2025, so a five-year share trend cannot be confirmed. Position rests on the structure of the arena and on price, and both point the same way.

Stability: pricing held, share trend unconfirmed

Stability: Moderate. Revenue ran $4.94 billion in fiscal 2021, $5.46 billion in fiscal 2023, $4.98 billion in fiscal 2024 and $5.38 billion in fiscal 2025, a dip and recovery that looks cyclical. Gross margin gave up 2.5 percentage points from its fiscal 2023 peak, far less than an 8.8% revenue decline would normally force. The nine months to July 31, 2026 are harder to read: revenue of $5.17 billion was 31% above the $3.96 billion of the same three quarters a year earlier, and goodwill rose by about $1.0 billion in the quarter ended October 31, 2025, alongside $1.63 billion of investing cash outflow, so an acquisition accounts for part of that growth. One year of recovery is not a trend.

Profitability: above the benchmark before the acquisitions, below it after

Profitability: Fairly Strong. NOPAT here is operating income times one minus the effective tax rate; invested capital is year-end equity plus debt minus cash. Both are estimates.

Fiscal year, USD bn20212022202320242025
Revenue4.945.425.464.985.38
Gross margin62.1%63.7%64.6%62.9%62.1%
Operating income1.081.121.360.830.95
NOPAT (est.)0.950.981.060.590.76
Invested capital (est.)3.563.954.015.136.59
ROIC (est.)26.7%24.8%26.4%11.5%11.6%

Through fiscal 2023 the company cleared the 15%-25% band that marks a protected business. Fiscal 2024 is the turning point: NOPAT fell 44% while invested capital rose 28%, and returns have sat near 12% since. Strip the $3.44 billion of goodwill out of fiscal 2025 capital and the remaining $3.15 billion earns about 24% (estimated). That adjustment flatters the figure, since it treats acquisition spending as free. The honest reading is that the operating assets earn protected returns while the prices paid for the acquired ones have not yet been justified.

Leverage does not distort the picture. Return on equity ran 25.3%, 28.3%, 24.0%, 12.6% and 15.4% across fiscal 2021-2025, and the balance sheet held $2.60 billion of cash against $2.57 billion of debt at July 31, 2026, so ROE and ROIC tell the same story. Cash conversion is not the constraint: free cash flow was $1.28 billion in the nine months to July 31, 2026, matching the whole of fiscal 2025.

Margins: 62 cents of gross profit, 18 cents of net

Margins: Strong. Gross margin was 62.1%, operating margin 17.7% and net margin 15.7% in fiscal 2025; the quarter ended July 31, 2026 reached 65.9% gross and 21.5% net. The roughly 44 cents of every revenue dollar between gross and operating margin pays for sales coverage, applications engineering, the service and calibration network, and R&D. Fixed costs of that size explain both the 24.9% operating margin of fiscal 2023 and the 16.7% of fiscal 2024, and they are the reason scale matters here. Net margin ranged from 12.3% to 20.7% over the five years.

Overall rating: Fairly Strong, with the doubt in the capital base

Overall: Fairly Strong. Gross margin behavior is the decisive evidence: five years, one downturn, a 2.5-point band. Estimated ROIC cleared the benchmark in three of five years and fell short in the two that carried most of the goodwill. Market position and margins carry the weight, stability is unproven, and profitability is the swing factor. Two years at 12% are a warning about what Keysight paid for growth, not yet proof that the instrument franchise has lost pricing power.

Where the advantage comes from and how long it can last: the qualification file outlasts the spec sheet

In this sub-industry the durable assets are a customer's signed-off test procedure and a calibration network that reaches every lab that customer operates. Both take years to build, and neither can be bought by a rival in a single order cycle. The gross margin is where their value shows up.

Customer captivity: the instrument is written into the test procedure

Demand-side advantage: Fairly Strong. An engineer picks an instrument while designing a chip, a radio or a satellite payload, then writes that measurement into the qualification plan that goes for internal sign-off and, in regulated work, to a customer or a certifying body. Changing the box later means re-correlating results, re-buying calibration and rewriting the software that drives the measurement sequence. The switching cost is measured in engineering weeks and schedule risk, not in the price gap between two instruments, which is why a 10% discount rarely moves a program mid-flight. Captivity has to show up in price, and here it does: a gross margin that moved 2.5 points while revenue fell 8.8%, and 65.9% in the quarter ended July 31, 2026.

The annual worth of that grip is easy to size. On fiscal 2025 revenue of $5.38 billion, one point of gross margin is about $54 million. Holding price through a downturn instead of conceding five points is worth roughly $270 million a year (estimated). That is what the captivity pays, and it is why the margin line is the first place erosion would appear.

Two limits keep this short of Strong. Procurement officials do extract concessions at renewals, and multi-vendor benches are routine below the top tier, so the customer is captive to a workflow rather than to Keysight. What would break it: a rival that matches the top-end specification and absorbs the requalification cost, or a customer generation that standardizes on modular hardware and vendor-neutral software, at which point the instrument stops being the lock.

Scale: about 44 cents of every revenue dollar is fixed cost

Scale advantage: Moderate. The 44-point spread between gross and operating margin in fiscal 2025 is the share of revenue consumed by sales coverage, applications engineers, the calibration and service network, and R&D. Those costs do not move with a single order. Fiscal 2024 shows the arithmetic: revenue fell $0.48 billion and operating income fell $0.53 billion, so each lost dollar of revenue took about $1.10 of operating income with it (estimated from reported figures and possibly distorted by one-time items). A small rival carrying the same fixed base on a fraction of the revenue cannot price to match.

The leverage cuts both ways. It is why a pause in customer programs hits earnings so hard, and it is why scale is a barrier only when captivity holds: without captivity, an entrant can buy volume with price and spread the same costs later. The acquisition history complicates the picture further, because the operating base grew with the deals and is now larger than the standalone revenue would justify. What would break it: a shift toward modular, software-defined instruments built on merchant hardware, which lowers the volume a vendor needs to be efficient and lets a smaller competitor skip the service network.

Proprietary silicon and calibration traceability: real, and partly copyable

Supply-side advantage: Moderate. Keysight designs its own signal-conversion and processing chips, which lets it reach a new specification before off-the-shelf parts get there. That is a lead, not a wall: converter vendors sell to rivals, and know-how moves with people. The stickier asset is traceable calibration, meaning an accredited lab network, audit records and certificates that a customer's own customer will accept. Replicating that takes years and capital, and it is renewed every calibration cycle. Brand recognition among engineers adds little on its own; an installed base matters only where calibration and automation keep it locked. What would break the supply-side edge: a rival reaching the same bandwidth generation on merchant silicon at a comparable price, which turns a two-year lead into no lead at all.

Durability: the substitute is cheaper verification, not a better instrument

The threats that matter do not come from a superior box. Chinese vendors have been climbing from the price-sensitive low end, and their progress decides whether the top tier stays a four-supplier arena or narrows into a three-supplier one with a cheap alternative below it; the evidence so far puts them under the specification threshold for the most demanding work. A slower but larger risk is verification migrating into simulation and design software, which would shrink the number of measurements that need a physical bench at all. Export controls on sales into China are a revenue risk rather than an erosion of the moat, since they remove a market without changing the competitive position.

So far the margin record supports durability at the top of the market and leaves the acquired adjacencies unproven. The single measure that tests the whole case is gross margin through the next revenue swing: a figure that holds above 60% while revenue grows would say the captive base is being defended, and one that slides while revenue rises would say it is being bought. Whether the acquired businesses can earn the core's returns on capital remains open, and it is a different question from whether the instruments still command their price.

Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.

Continue the comparison

Continue reading the full report

  1. Comparison and summaryHow Keysight Technologies Inc compares with its closest rivals, and our moat rating

Free to sign up · 20 research points · unlocking this report uses 10

20 points cover 2 full reports. No card needed.

Research reflects its generation date. Preview the first two modules.

More on Keysight Technologies Inc