Agilent: Back To The Lab - [Business Breakdowns, EP.223]

Summary

• Agilent is a razor-and-razor-blade lab equipment franchise: roughly one-third instrument sales, two-thirds consumables, servicing, and software, according to Mark de Vos. A typical chromatography instrument costs ~$100k, lasts 6-10 years, and generates another ~$100k of consumable and service spend over its life. Quality-control processes can be written into the FDA approval for a drug and may name the exact instrument and consumables, making switching expensive and risky and supporting recurring revenue.
• Market shares in this industry barely move — Waters once showed sales of a single column type flat over decades. Agilent probably holds ~2/3 of global gas chromatography, ~1/3 of liquid (Waters holds about another third), and 40-50% of spectroscopy, against a “pretty stable competitor set” of Waters ($3B revenue, 60% pharma) and Thermo Fisher ($43B, broad catalog). Published TAM is $160B versus Agilent’s ~$6.5B revenue, but de Vos notes they’re #1 or #2 in most niches that matter.
• Service is the moat’s active layer and the growth lever. Agilent employs 4,000 service technicians (75% with chemistry/biochemistry degrees, 25% with a master’s or PhD) doing 2,500 on-site calls a day, and is one of only two companies able to service competitors’ equipment in the same lab. The service attach rate has climbed from high-20s% to low-30s% and is expected to keep rising ~1pt/year — Waters, better at bundling for longer, sits in the 50s%, implying “a pretty long runway.”
• The economics compound quietly: ~5% organic sales growth since 2015, but 13% EPS CAGR, with ~$1.4B annual FCF at ~90% conversion of adjusted net income. Consumables and servicing run a 34% operating margin versus ~20-24% for instruments, and the service business is “a classic density business” that scales as local route density fills in. Instrument spend is “a rounding error” for a pharma facility, while the cost of missing a contaminant is “almost unquantifiable.”
• NASD, the oligonucleotide CDMO business, is the growth kicker born from a culture that backs its engineers. Built from in-house RNA synthesis know-how plus a tiny 2006 acquisition, orders began to pick up in 2018; management chose a $185M organic expansion over selling or giving away the asset, then broke ground on a further $725M expansion in Frederick, Colorado in 2023. With the ~$1B BioVectra acquisition, it’s now ~$470M of revenue (~7% of sales) — and “there’s no other company that has equal capacity or larger” for production-grade oligos.
• Valuation is deliberately simple: Agilent recently traded above a 4% FCF yield — near the highs of the past 10 years since the Keysight spin — equating to just under 20x earnings for a business de Vos expects to grow earnings high single digits. Upside optionality comes from testing markets that “appear out of thin air”: PFAS testing is already a $400M market growing 20% a year with Agilent the clear leader, “growing like an absolute weed.”
• Risks are cyclical, not structural: the post-COVID pharma instrument hangover appears to be ending, and China (18% of group sales, government-grant driven) appears to be improving off a down cycle. NIH-linked funding is only ~1% of sales. De Vos’s closing lessons: “not to underestimate the quality of what sounds like a boring business,” and that life science tools companies have outperformed the pharmaceutical sector “hands down.”

Deep dive

1. A fridge-sized machine that tells you what’s in the sample — and locks in a decade of spend

• De Vos’s plain-English framing: picture an instrument “about the size of a small fridge” on a lab bench whose job is to identify molecules in a sample — e.g., a pharma company pulling one aspirin tablet per thousand off the line to verify active ingredient levels and screen for contaminants. Agilent sells the instrument, then the consumables, service, and software to run it.
• The mechanics matter because they drive the razor-blade: chromatography forces a liquid or gas at pressure through a “column” — a metal tube stuffed with material that separates molecules at different rates — and columns degrade under heat and extreme pH after perhaps ~2,000 samples, so customers may replace them monthly. A ~$100k instrument with a 6-10 year life generates another ~$100k in consumables and servicing, split roughly equally.
• The footprint is broader than “biomedical”: 110 countries, 285,000 labs; pharma is ~a third of sales, academia/government ~10%, chemicals and advanced materials (battery testing, semiconductors) ~20%, food and environmental testing another ~20%. Published TAM is $160B against ~$6.5B of revenue — Matt’s wry note that “if a company publishes a TAM number, it’s probably gonna be big” — but the relevant fact is #1 or #2 positions in most key niches.

2. Born in the HP garage, refined by subtraction

• The lineage runs from Hewlett and Packard’s 1938 garage — their first product tested audio equipment Disney used producing Fantasia — through HP’s 1965 acquisition of the small business F&M Scientific, which quickly became a gas-chromatography leader and whose instruments were used for the first full-scale steroid testing at the 1972 Munich Olympics. HP spun Agilent out in 1999 as it focused more on computing.
• What followed was a period of focusing: medical sold to Philips in 2001; the semiconductor components business sold to KKR and Silver Lake in 2005 — which became Avago, then Broadcom; and in 2014 the electronic measurement business (~40% of sales, deeply cyclical) spun out as Keysight, leaving a “clean, predictable, less volatile business” focused on molecular analysis.
• The HP-way culture persists — Mark says its current HQ is about a twenty-minute drive from the original garage, managers are “generally Agilent lifers that started as engineers,” and in COVID they were “first out the door saying, ‘No one in this company is losing their jobs.’”

3. Why market shares don’t move: regulation, service density, and incumbency

• The lock-in is regulatory: QC processes can be written into a drug’s FDA approval, sometimes specifying the exact instrument and consumables — “once it’s tied in, people just use the instruments again and again and again.” De Vos’s best evidence is a Waters investor-day chart showing sales of one column type essentially flat over decades.
• Competition is a stable oligopoly: Waters (~$3B revenue, narrowly pharma-focused, strong in liquid chromatography) and Thermo Fisher ($43B, acquisition-built, selling everything from lab fridges to water baths — “a bit less of a specialist”). Agilent probably dominates gas chromatography at ~two-thirds global share, skewing to applied markets since proteins can’t be heated without being destroyed; even a great product launch might gain “a percentage point of market share” after a couple of years.
• Service is a key place the fight is won — most customers “would probably highlight service as the primary driver” when deciding whether to add another instrument. Agilent’s 4,000 directly employed technicians (75% degreed in chemistry/biochemistry, 25% master’s or PhD) run 2,500 on-site calls daily, and it’s one of only two firms that can service rivals’ equipment in the same lab. Attach rate has gone from high-20s% to low-30s%, versus Waters in the 50s% — the visible runway.

4. The numbers: modest top line, compounding bottom line, digestible cycles

• Since 2015 (the longest clean post-Keysight window), organic sales growth averaged ~5% (6% reported with bolt-ons), while EPS compounded ~13%. Consumables/service earn 34% operating margins versus ~20% for life-science instruments and ~24% in applied; FCF is ~$1.4B a year at ~90% of adjusted net income. Pricing is deliberately restrained — low single digits, “certainly not the ten, fifteen percent we’ve seen in the food companies” — protecting 30-50-year customer relationships.
• The cycle risk played out recently: COVID drove pharma to buy “well above the normal level of instruments,” then “the CFO turns around and goes, ‘We’ve got lots of instruments’” — a downturn Agilent is “just coming out of.” China, 18% of group sales and heavily government-grant led, also appears to be improving. Replacement demand is supported by wear (valves, heaters, high temperatures, “nasty chemicals”), intolerable lab downtime, and platform upgrades roughly every three years.
• The asymmetry underpinning it all: five instruments in a production facility cost maybe $1M over eight years — “a rounding error” — while a missed contaminant carries a cost that’s “almost unquantifiable. The damage alone to the brand would be huge.”

5. NASD, valuation, and the case for boring picks-and-shovels

• NASD (Nucleic Acid Solutions Department) makes oligonucleotides — lab-made DNA/RNA strands used in genetic-disease treatments and some RNA-based vaccines, including COVID-19 vaccines. Formed in 2006 from in-house RNA synthesis technology plus tiny acquisition SignPro, orders began to pick up in 2018, when management faced a fork: invest $185M or give the business away. They backed their engineers (net debt just 1x EBITDA), then broke ground on a $725M expansion in Frederick, Colorado in 2023; with the ~$1B BioVectra deal, the franchise is ~$470M of revenue, ~7% of sales. Development-phase work earns little, but commercialization volumes “really ramp, and that stays with Agilent” — only Agilent has production-grade capacity at hundred-thousand-vaccine scale. Setbacks for RNA-based drugs and changes in the funding environment are unhelpful.
• De Vos’s valuation discipline is intentionally simple — FCF yield or P/E, “trying not to overcomplicate it”: recently above a 4% FCF yield, near the highs of the past 10 years, or just sub-20x earnings for high-single-digit earnings growth. Upside comes from testing markets appearing “out of thin air” — PFAS/forever-chemicals regulation has created a $400M market growing 20% annually where Agilent leads, “growing like an absolute weed.” On takeover risk: Thermo Fisher “would, I’m sure, love to buy them” but “you’d attract competition scrutiny.”
• De Vos’s two takeaways: don’t underestimate a boring business — servicing lab instruments sounds like something “you might be forgiven for telling me to move on” about, yet it’s 33% operating margins growing high single digits — and it’s “often better to be in the businesses that are supplying tools,” since life science tools have beaten pharma “hands down.” Matt’s coda: “Picks and shovels. It’s been working for almost 200 years now.”