Primo spent $83m on capital and set aside $300m for buybacks. Which water dispense format pays when the spending stops?
By Zenith Water Dispense Team ·
Primo Brands chose a $300m buyback over more machines. Which water dispense format keeps earning when the capital stops?
I think this trade asks the wrong question about capital. Most operators ask what a new placement costs. I would rather know which part of an estate keeps paying when nothing new goes in.
Primo Brands has just put a price on that answer. The quarter threw off 274.2 million dollars of adjusted free cash flow. Capital spending came to 83.0 million in the same three months. Then the board signed off a buyback worth up to 300 million dollars. Primo's quarterly release of 5 November 2026 carries all four numbers.
A buyback instead of a bottling line
Primo is the largest bottled water delivery business in North America. Revenue for the three months reached 1.82 billion dollars, up 3.1%. Underneath that, customers took 2.6% less water. They paid 5.9% more for what they did take. Price carried the whole of the growth, and volume worked against it.
Adjusted EBITDA grew 1.6% to 411.0 million dollars. Margin gave up 30 basis points and finished at 22.6%. The company also moved its full-year profit guidance to the bottom of its published range. A business in that position can buy growth or build it. It can also decline both and hand the cash back. Primo has told the market which it prefers, and it put 300 million dollars behind the answer.
Each format asks for capital at a different moment
That choice is open to Primo because of what its revenue is made of. Almost all of it is water sold to coolers already standing in customers' buildings. No new install has to happen for that money to arrive. The three dispense formats are a long way from equal on this point.
With BWD, bottled water dispense, the cooler is one purchase and the water is another. POU means point of use. Nothing is ever delivered to it. Instant taps, or ITS, go into the counter and are usually bought outright.
In one market Zenith models in full, water sales are over nine tenths of bottled dispense revenue. Mains-fed coolers in the same market earn over nine tenths of their revenue as rent. Instant taps run the other way, with roughly four fifths of their revenue arriving as hardware changing hands. That hardware line only exists if somebody installs something this year.
The format you can stop investing in
Bottled deserves a fair hearing here, and the argument is about capital. An estate of bottled coolers already in place will earn again next year. It does not need a single new unit to do it. A mains-fed book behaves much the same way once installed, because the rent keeps running. An instant tap book does not. Most of its revenue has to be won again every year, with a new fitting job behind it.
That makes bottled the quiet funder in a mixed estate. Operators converting a bottled base to taps swap a line that renews. They take on one that must be re-sold. The conversion can still be the right call, and in most markets it is. It has to be paid for by the format it replaces. Nobody should sign a tap programme without working out what the bottles are covering while it runs.
What to put in the 2027 plan
Primo's quarter is a capital-allocation signal, and European operators should read it as one. This business read its own volume line and chose shares over machines. Smaller operators cannot buy their own equity. They still face the same arithmetic on a smaller scale.
Work out what share of next year's revenue would still arrive if you installed nothing at all. That share is what funds everything else. Almost nobody in this trade can produce it on request. Diligence teams have started to ask for it. The operators who can answer in 2027 will defend their price. The ones who cannot will be valued on unit counts instead.
📞 Deciding where the next pound of capital goes?
Half an hour on a call usually settles which part of an estate is funding the rest. Bring your format mix and next year's install plan. We will tell you what the numbers say.
P.S. A falling unit count does not have to mean a falling bill. Working out which part of an estate pays without fresh capital takes a format-by-format revenue split, and that is what the 2026 Zenith Water Dispense Market Reports put on one page. Thirty-one is the European number. If South Korea or Turkey is the market on your desk, the file is written. Whichever of Japan, Mexico and the UAE you need, the build is the same. School and campus sites are counted on a line of their own. Both the money and the machines are forecast to 2030. Although most orders are the figures alone, the long-form study can travel with them. Trusted by industry leaders since 1998. See which markets are covered.