Waters Corp U.S. All Stars Conference

Yeah. I think firstly, thank you for having us. It is a pleasure to be here. The quarter came in ahead of expectations, as you said. On the base business, across the board, we saw acceleration, right? We saw instruments growing high single digits, service growing high single digits, chemistry yet again double digits, and if you accounted for the pull forward that occurred due to Liberation Day same time last year, it grew 16%. Geographically, we saw U.S. mid to high single digits. Europe was mid single digits, but China and India were in excess of India was in excess of 20%, and China was again double-digit growth. If you account for the pull forward, China was mid-teens.

When you look at end markets, pharma was again double digits with the replacement cycle contributing in the U.S. and Europe with large pharma. India benefiting from generics. The good news was the best coming out of China, where again, pharma grew in excess of 24%, 25%. First quarter was 50% growth in pharma in China. So pharma recovered nicely. Industrial was low single digits. Academic and government, the end market was double-digit growth. That was largely driven by nice budget release in Europe. Even in that number, there was no stimulus from China, right? So across the board, the analytical science business and the material science business did extremely well. This is a legacy Waters business. Where we were even more pleased is how the integration has started with the Becton Dickinson businesses. They both, on average, grew roughly 4%.

China had a 200 basis points headwind. If you extract it was a 6% growth, already mid-single digit plus growth. The more important thing is what was underneath contributing to it. 180-day plan to accelerate sales activity, to accelerate pricing, which was 90 basis points versus 50 in legacy BD. The growth in China that we expect to start to come in Q4, all the tactics are starting to contribute. We saw cross-selling contribute as well. Overall, from a revenue perspective, things are going reasonably well, knock on wood. The leading indicators are very good. On the cost side, what we also shared is from a cost synergy perspective, by the end of the year, we will have a run rate of $200 million as we go into next year.

Remember that $200 million was the number that we had targeted at the end of three years for cost synergies. So within nine months, we're already there, and there are a lot more initiatives to contribute there. As you finish 2026, and we're not yet even done with Q3, we think we're very well set up for 2027, where the BD businesses should accelerate. Our base business is now a sustainably high single-digit grower. So sets us up very nicely for what's to come ahead of us.

I think there are two topics that have sort of dominated conversation and tools for the last few months. One is AI, the other is reshoring.

That's the next- If the reshoring conversation gives you any evidence, we are a reluctant contributor to the discussion until we see facts.

With AI, since we knew a lot of people are interested, we sort of dug in with the teams, and I'm going to give you first a frame and then some evidence that it's having impact on Waters' business. When you think of labs can be broken into three pieces. First, there is a sample prep part where you're taking reagents and samples and putting them through a workflow. Once that workflow is perfected, you can automate it and then you can have an agent run it. Not before then, right? So just keep that in mind. So in reagents, you first have to sort of get it standardized and then only the AI model can be unleashed on it.

Second is the instrument, and in the instrument, the protocols are sort of generally refined, and different users have different protocols. Third is software. These are the three parts of a lab, right? Of an experiment in a lab. If you take that logic of reagent standardization and you go into quality control and late-stage development, that's where the variation is very little. Right? Because in quality control, the reagents that you use to analyze samples are the same if you run the experiment versus I run the experiment. That's the part of doing quality control that you want minimum variability in. Second, the instrument protocol is the same. Third, and most important, in QA, QC, and late-stage development, the software of record in pharma is Empower. Four in five molecules use Empower, right?

You cannot pull the data out of Empower and put it back into Empower and have the regulator look at it. So it's the software of record, and so that software has a certain protocol itself. Now, where we have seen evidence of the adoption of AI is on all three fronts, especially on the software side in QC, where customers are buying more seats in order to train agents on an already perfected workflow on the reagent side, on the instrument protocol, and of course then the software, right? So in QC, you start to see the adoption of agents and use of agents in labs much faster than you see in discovery. Now, you take the same logic and go to drug discovery. In drug discovery, everyone runs an experiment differently using different reagents. So first you have to standardize that.

We are seeing increasing consumption of reagents and columns in that segment as well, but not at the same pace as we are seeing in QC. Second, once you've done that, the protocols have to be standardized so the agent can run it. That also requires a bit of work. There is no software of record. There are many different softwares that are used to analyze different types of samples. Our belief and the facts suggest to us that the adoption of AI models and AI agents is much faster in QC and manufacturing than it is in drug discovery. I know the discussion's largely dominated by discovery discussions. The reagents consumption, as we compare our two businesses, in BD Biosciences, the reagents consumption is higher on the discovery side.