Waters Sells Its AI Story While Quietly Retiring GLP-1
A 140bp organic raise and a first 2027 framework built around Empower and agentic labs — plus the two growth engines management stopped talking about
WAT · Earnings Call · 2026-11-03
The AI framework arrives, four quarters after "too early to quantify"
Waters has spent a year gesturing at artificial intelligence and declining to price it. On the Q3 2026 call, Udit Batra finally gave it architecture: a three-part construct across research, development and quality control, anchored on Empower — the compliance system of record that sits under roughly 80% of medicines globally and holds "terabytes of validated data for each customer, a data set that is unique in our industry."
The new machinery is concrete. AI-powered peak integration, announced the week before the call, cut manual interventions by up to 95% in a large-pharma pilot, with 76% greater accuracy and 72% lower user-to-user variability. Waters AI Link plugs instruments into Anthropic's model context protocol, collapsing agent-to-instrument connection "from weeks to hours." And the customer funnel is real: instruments placed in AI-enabled labs at 8 of the top 10 global pharma, plus dozens of AI-native biotech first-time customers in 2026 — which Batra flags as feeding directly into higher LC-MS runs and column consumption. That is the bridge from drug discovery hype to consumables revenue.
But note the verbal tic. Here is the current call:
But with all of that said, I think it's too early to start quantifying things. I mean we've grown at a sustained high single-digit level over the last 2 years.
And here is November 2025, on the very same Empower software stack:
It's too early to start quantifying exactly what that is. I can tell you that there are a significant number of customers who've already transitioned in small to midsize pharma.
Same caveat, one year apart. What changed is that the narrative now has a premium tier, a subscription model, and a partner ecosystem — so the story got bigger even as the number stayed unquantified. That is not fraud; it is a company discovering that AI is the cheapest available re-rating narrative for a business already growing organically at “10% in constant currency at the high end of our constant currency guidance range” — Udit Batra, President and Chief Executive Officer · 2026-11-03.
The two things management stopped saying
GLP-1 was Waters' number one keyword in Q4 2024, and the Q2 2026 answer was emphatic: “GLP-1s, Puneet, grew that part. GLP-1 testing part of the business grew over 40% this quarter… there are no signs of slowing down.” — Udit Batra, President and Chief Executive Officer · 2026-08-04 On this call, GLP-1 appears exactly once in the prepared remarks, inside a list of "idiosyncratic growth drivers" expected to add 200bps through 2030. No 40%. No India doubling. The theme did not die — the global tape still trades it, with LLY discussing a GLP-1 bridge program and AZN an oral receptor agonist — it simply stopped being the pitch.
The same goes for the instrument replacement cycle, which once dominated the Q&A. Now it is a single sentence: "middle innings." No 2029-2030 air-pocket discussion, no CROs-and-branded-generics-coming debate. Both retired themes share a trait: they were 2024-25 stories that had already been fully credited. The 2026 story is AI, China normalization, and price.
Where the numbers actually moved
The unglamorous paragraphs carry the quarter. Total revenue was $1.767B, with organic at $871M (+10% cc) and the acquired Diagnostic Solutions and Biosciences businesses at $896M, +1% — what Batra calls a deliberate spot landing after draining inventory-in-trade. Net pricing reached 120bps from 90bps last quarter, reagent rental compliance begins recapturing low-to-mid single-digit millions per quarter from Q4, and the $200M cost run-rate lands more than two years early — helped, pointedly, by AI tooling.
China is the cleanest inflection. Legacy Waters grew over 20% there; Biosciences posted its first positive China quarter in ten. “Overall, China taking a full step back is neutral to accretive to the overall growth of Waters, which is something that hasn't happened in a long, long time” — Udit Batra, President and Chief Executive Officer · 2026-11-03 — a claim no prior call could make.
That produced the raise: FY26 organic guide lifted 140bps at the midpoint to 9.1-9.7%, revenue +$43M to ~$6.48-6.50B. The 2027 starting point is mid-single-digit plus organic and 29.2% adjusted operating margin. Vijay Kumar's arithmetic was sharper than the headline: given synergies in hand, why only 100bps? Amol Chaubal's answer matters — “those first 6 weeks, you don't sell much, but you still have costs that are largely straight line… roughly 70 basis points headwind to the margin.” — Amol Chaubal, Senior Vice President and Chief Financial Officer · 2026-11-03 Underlying expansion is therefore closer to 170bps. The framework is sandbagged, not stretched.
A premium multiple into a de-rating AI tape
The stock is up 24.7% over 90 trading days and sits 4.2% off an all-time high of $480.07 printed on October 29 — a re-rating into the print, not a reaction to it. At 5.8x sales, up 69% year-over-year, the market has already underwritten the framework. GAAP optics are ugly and misleading: operating income of -$141M in the last filed quarter, against GAAP diluted EPS of $0.82 versus $4.10 adjusted in Q3. The balance sheet carries -$4.6B of effective net cash, and receivables at 120.8% of revenue explain the free-cash-flow question Doug Schenkel pressed, and the honest answer that 2026 conversion is dragged by integration cash costs and a ~$150M BD receivable.
The contrast with the broader tape is the interesting part. The AI-compute complex is being sold hard — data center-linked baskets are down 20-35% over 30 days across dozens of tickers. Waters is not in that trade. Its data-centre exposure is materials testing for semiconductors and build-outs, a ~$600M addressable market growing double digits, tucked inside Materials Sciences at $93M for the quarter. Waters' AI premium is priced off regulated-lab data, not GPUs — which is precisely why it can rise 25% while the compute trade falls. Whether that decoupling survives a year in which "it's too early to quantify" still stands is the whole question.