New Capacity Takes Years. Recovering It Doesn’t.

By Author_Posts in Articles, Dairy, News Posted August 25, 2026

You may not be able to build meaningful new whey-processing capacity in time for this year’s demand, but you may be able to recover capacity from the assets already installed.

The fastest way to increase capacity for a dairy processor today may not be a new plant. It may be inside the membranes already running.

U.S. whey protein inventories have fallen roughly 50 percent since 2023, and the industry’s response has been to build. Saputo, DMK, and FrieslandCampina have all announced new or expanded whey-processing capacity. Fonterra’s Studholme expansion, Wisconsin Whey Protein’s own expansion, and Idaho Milk Products’ new blending facility are expected to begin contributing supply in 2026. Larger projects, including Glanbia’s additional WPI capacity and Tirlán’s premium-whey facility, won’t contribute until 2027 or later.

None of that solves a shortage this year. Converting liquid whey into WPC and WPI requires specialized membrane and processing equipment that takes years to design, permit, and commission, and recent reporting points to bottlenecks at several stages: membranes, pasteurization, evaporation, drying, and CIP turnaround.

For a processor who can’t wait two years, the opportunity is what’s already on the floor.

What the permeate stream is carrying away

Dairy Reporter posed the question directly this July, asking how a processor boosts whey output without expanding. Part of the answer sits inside the UF and MF systems already running.

In ultrafiltration, membranes retain protein in the concentrate while water and smaller solutes, including lactose and minerals, pass into the permeate stream. Well-performing UF membranes minimize that protein loss, but performance drifts slowly enough that it’s easy to miss without a closer look. Microfiltration plays a different role upstream and downstream of UF, defatting, clarifying, or fractionating depending on pore size, and drift matters there, too.

In field assessments of dairy processors’ UF membranes, Solecta has identified recoverable protein yield of 5 to 15 percent, protein that would otherwise pass into permeate but can be kept in the retentate instead. Turning that recovery into finished product still depends on downstream capacity. In one anonymized Wisconsin case, a microfiltration train recovered 16.2 additional pounds of protein per operating hour, about 292,000 pounds a year across three parallel trains at the plant’s operating schedule. Results like these are site-specific and can vary depending on membrane age, feed composition, and downstream processing capacity.

None of that required a new line, a new building, or a gallon of additional milk. It required converting to higher-performing membranes in the trains already on the floor.

That recovery doesn’t automatically become more finished powder, though. Evaporation and drying capacity can each become the binding constraint even where membrane performance improves, so the real question is whether the rest of the plant can absorb what the membranes recover.

Public conversation about membrane performance tends to focus on cleaning cost and water use, both easy to measure. Protein yield gets less attention because it never shows up as a cost, only as revenue that was never captured. On a plant processing 1 million pounds of whey a day, a 2.5 percentage point yield improvement recovers roughly 25,000 additional pounds of protein daily. At $12.50 to $14 a pound, that’s $312,500 to $350,000 a day, on one plant, at that throughput and price. Not a small number to leave in a waste stream.

Why the economics of an element matter more than its price tag

Every membrane element reaches what Solecta’s applications engineers call a flip point: the moment operating performance has degraded enough that the protein lost to permeate costs more than a new element would. Where that point falls is site-specific, depending on flux, protein passage, cleaning recovery, downtime, and how the plant values the protein it’s processing, not a fixed number of months that applies everywhere.

That’s why the replacement decision is a lifecycle economics question, not a purchase price question. In Solecta’s lifecycle analyses, the purchase price is often small relative to yield loss, downtime, cleaning, energy, and labor over an element’s operating life.

The element that costs the least to buy is not always the one that costs the least to run.

That math matters more as protein prices climb and margins tighten. Nearly 70 percent of U.S. dairy companies reported flat or shrinking margins in 2025, up from 66 percent in 2024, and about 65 percent of U.S. respondents ranked cost management among their top three priorities. Getting more from equipment already installed and paid for fits squarely inside that pressure.

An overlooked part of the plant, doing more than it gets credit for

Knowing where a plant’s membranes sit relative to that point takes data, and collecting it consistently is its own challenge. U.S. dairy leaders continue to report difficulty filling and retaining skilled maintenance and process engineering roles, along with succession-pipeline concerns as experienced operators retire.

Most plants already collect flow rates, pressures, CIP history, and permeate composition analysis. What’s often missing is tracking that data against a consistent baseline long enough to catch a decline before it gets expensive. Every dairy plant running ultrafiltration or microfiltration is sitting on some recoverable protein yield, whether or not it shows up as a line item today. Finding it starts with an honest look at how performance has drifted since the last change-out, and what that drift is costing, measured against the plant’s own baseline.

Solecta applications engineers spend a good part of their week walking through exactly this question with dairy processors. If it’s a conversation worth having, we’re glad to have it.