Two Industries Sitting on Exactly What the Other Needs
I attended Kornit Konnections this week and spent most of my time listening to retailers describe their own problems, mostly disconnected systems and planning calendars that had quietly become constraints rather than tools. But the elephant in the room was the problem that retail has been trying to solve for 20+ years: inventory.
There is a particular kind of organizational pain that comes from knowing exactly what is wrong and being structurally unable to fix it. Retail has been living with that pain around inventory for a generation.
On the balance sheet, finished goods inventory looks like an asset. To a lender, it is collateral. Banks typically advance between 50 and 80 percent of appraised inventory value against it. Public markets treat inventory turns as a signal of operational health. Entire finance functions have been built to optimize these numbers, and for a long time the model held together well enough. The cracks in that model are widening.
In 2023, the fashion industry produced somewhere between 2.5 and 5 billion unsold garments, representing an estimated $70 to $140 billion in retail value sitting in warehouses, on markdown racks, or written off entirely. Nike reported in 2024 that markdowns affected 44 percent of its assortment, more than double the figure from two years prior. The cost of holding unsold inventory typically runs 20 to 30 percent of product value per year so a garment worth $100 costs $25 to carry for twelve months. Across millions of units, the math becomes unforgiving.
The root cause of this is structural. Most retail planning still operates on a 12 to 18 month calendar, with buyers committing to finished goods nearly a year before a consumer ever sees them. The strategy can work brilliantly for brands that correctly predict what consumers will want six or nine months out. But for every brand riding high on the right bet, there is another working through an inventory overhang it cannot move. This planning and purchasing calendar is not just a process. It runs through organizational structure, vendor contracts, financing covenants, and incentive systems built around certainty that the market rarely provides, especially now.
Retail has been talking about fixing this problem for two decades and largely has not because of the complex change management problem embedded in how retail companies are built, capitalized, and led.
I listened to a panelist at Kornit Konnections talk about a solution. He spoke about how the longest lead time in any apparel value chain is the raw material component. The more finished a product is when you take a position on it, the worse your flexibility when demand shifts. The brands that win the flexibility game are the ones that take positions on the raw materials underlying the majority of their volume and hold inventory in its most configurable, unfinished state for as long as possible. Imagine a roll of fabric before it is printed, dyed, or decorated. This approach preserves optionality.
While I was listening to this panelist talk, I thought of the dozens of print-on-demand facility tours I’ve done over the years. POD operators have understood this plan instinctively for years.
What POD has gotten right by necessity
Print-on-demand businesses were not built around a planning calendar because they could not afford to be. With no MOQs or bulk buys, why or how could you hold onto warehouses of finished goods, all wagered on demand that has not yet materialized. The entire POD model was constructed around a single structural insight: produce only what is ordered, when it is ordered, and hold raw material in its most configurable state until the last responsible moment.
Companies like Kornit Digital have built their entire operating thesis on this principle, making the case that businesses should not hold inventory they have not yet sold. The economics are not complicated. The capital efficiency is real. The inventory risk is structurally neutralized rather than managed after the fact.
This is what retail needs to borrow. Not just the philosophy, but the operating model behind it. The two-week product drop, executed through a small nimble team or through vendors taking product to market directly on a brand’s behalf, is not a fantasy. It is already happening with brands who have punched through the change management challenge. The brands not doing it are mostly waiting for their planning process to approve something the market already validated.
What POD has not figured out yet
As I was listening to these same retailers talk, I also ruminated on what I think is a harder truth that tends to get less airtime at POD-centered industry events.
The flexibility that makes print-on-demand structurally superior is also the thing that keeps most POD businesses from becoming “real” brands. When you never fully commit to a product, you never fully commit to a point of view, which is what brands are built on. The blank t-shirt that can become anything is, in a meaningful sense, nothing in particular until a decision is made about what it should be. I think back to a conversation I had with the COO of one of the original POD-native marketplaces — she told me her organization was moving away from POD as a supply chain strategy because the products weren’t unique enough and she couldn’t hold onto her customers’ loyalty. Her brand had lost its point of view by relying 100% on a POD model.
Retail, for all its dysfunction, has produced enormously wealthy companies and operators because the model generates real brand equity, real consumer loyalty, and real pricing power. That value compounds over decades. POD has largely produced high-volume, low-margin fulfillment infrastructure that struggles to hold pricing because differentiation is thin and switching costs are low. The moment a POD operator or brand starts spending like a retailer, without the brand equity to support it, the economics unravel. The capital efficiency that makes the model attractive disappears the instant traditional retail cost structures get layered on top of unit economics that were never designed to carry them.
The financial disparity between the two industries tells this story precisely. The global retail technology solutions market reached $234.5 billion in 2023. In a single quarter of that same year, retail tech financing deals alone totaled $4.7 billion. More than 18,000 venture capital investors are active in the space. Against that backdrop, the entire print-on-demand software market is estimated at $4.5 billion in total market size in 2025. Not invested capital. Market size. The POD largest fundraising round into pure software ($260mm into Gelato) is the major VC financing deal of this decade, and that is no discredit to Printify who raised material capital around the same time.
This comparison is not a criticism of POD software but rather a precise measurement of how differently capital has flowed into these two industries, and how much of the capability gap between them traces directly back to that difference.
That capital built something specific in retail and did not just fund talented people. That money has built the data infrastructure, the demand forecasting systems, and the organizational muscle required to move from operator-driven decisions to systems-driven flows. The Harvard MBA founders and McKinsey-trained data scientists who entered retail over the past two decades tell another story of this funding, technology, and talent gap. The destination retail is sprinting toward, thanks to the large amount of money that has flowed into that world, is a business where systems handle the routine decisions, humans make the judgment calls, and the gap between sensing demand and responding to it collapses toward zero. POD is, by and large, still at the beginning of that journey.
The shift that changes everything
The transition from operator-driven decisions to systems-driven flows is the defining operational shift of this decade, and it is arriving in both industries simultaneously, though from very different starting points.
Retail built toward this transition over decades of capital investment and institutional development. The infrastructure exists. The discipline exists. What retail is now discovering is that the systems it built to optimize a calendar-planning model are capable of something far more dynamic, if the organizational will is there to let them evolve.
POD is approaching the same transition from the opposite direction. Less infrastructure. Less institutional data. Less financial backing. But a model that is already structurally aligned with on-demand, unit-level economics.
AI and autonomous workflows are an accelerant for both industries, if used correctly. AI can be a great equalizer for the POD market if we leapfrog ahead in how we build software, analyze data, and solve complex data and systems integrations. AI can help us close the distance between the instinct-driven POD operator and the analytically rigorous retail planner. The data sophistication that retail spent decades and billions building is becoming accessible in compressed time and at a fraction of the historical cost. The POD business who makes routing decisions, pricing adjustments, and replenishment calls by feel will be competing against systems that make those same decisions in milliseconds, with full visibility across every variable. That is not a threat to dismiss. It is an invitation to build.
The most important variable is still the people
I think that the POD and retail industries are missing deliberate cross-pollination of talent. The POD industry needs more operators who have run sophisticated retail planning functions and supply chains, who have sat inside organizations where Harvard MBA founders set the strategic direction and McKinsey-trained data science teams built the analytical infrastructure beneath it. People who understand what it means to manage inventory at scale, build a brand that commands pricing power, and run a business with the financial discipline the next stage of growth actually requires. Retail needs people who have built nimble, inventory-light operations from the ground up, who understand configurable supply chains, unit-level economics, and what it actually means to take a product to market in two weeks without a planning committee signing off on it.
Right now, those two talent pools barely intersect. That is one of the most consequential gaps in either industry and one of the least discussed, in my view. The conversations happening inside POD companies and inside retail organizations are almost entirely separate. The insights being generated on one side are not reaching the other.
That is the conversation worth having and Kornit Konnections made huge strides this week in starting it. I applaud Kornit for a fantastic show that recognized how much retail and print-on-demand have to offer each other. These industries have more to offer each other than either has been willing to admit, and the window to act on it, before systems make the capability gap irreversible, is open right now.
