Home kitchens and coding agents are opening a new band of viable small enterprise

A home-based ice cream brand in Hopkins, Minnesota is running at six to seven figures in annual revenue on 40 to 50 percent profit margins. Sam Parr cites it as one of the largest ice cream brands in the state. Parr adds a second case: a woman earning one million dollars a year baking cookies from her own kitchen. Both numbers sit well outside what most people assume a home-based food operation can achieve.

The margins Parr cites matter as much as the revenue. Forty to 50 percent on a food business is, as he puts it, “unheard.” Commercial kitchens, licensing costs, and distribution overhead typically compress food margins far below that level. A home operator avoiding most of that overhead means the revenue translates to income in a way that a comparably sized conventional operation often does not.

The software world is surfacing a structurally parallel claim. Martin Casado argues that coding agents are opening the same kind of access at the software layer. “Now with coding agents you can build a software product that generates $100,000 of revenue a year, a million dollars of revenue a year,” Casado says. The implication is that people without engineering backgrounds will be able to build and operate software businesses in revenue ranges that were previously gated by the cost of technical talent, in much the same way that home food operators are reaching revenues that once required institutional overhead.

Now with coding agents you can build a software product that generates $100,000 of revenue a year, a million dollars of revenue a year. Martin Casado

The connection between these two patterns is structural. In both cases, a gatekeeping requirement has been reduced or removed: the commercial kitchen and its associated cost structure in one domain, the engineering team in the other. What follows in both cases is that meaningful revenue becomes achievable by a single person or a household operating well beneath the threshold that attracts institutional capital or attention. These businesses are too small for venture capital and, at the margins Parr describes, profitable enough not to need it.

That combination has always existed in theory. What the cases in evidence suggest is that the ceiling is moving. A home food business generating six to seven figures at 40 to 50 percent margins is not a side income supplementing a primary livelihood. It is the livelihood. Casado’s framing for software points in the same direction: $100,000 to $1,000,000 in annual software revenue, built and run without an engineering team, occupies a band of enterprise that the startup model was never designed for and that traditional small business infrastructure is too expensive to support.

None of this constitutes a settled pattern. The cases Parr names are specific and real, but it is not clear how replicable the highest-revenue examples are across operators with different products, locations, or distribution access. Casado’s claim about coding agents is forward-looking and has not yet been tested at scale. What the evidence does suggest is that two separate enabling shifts, one in home food production and one in software development, are creating viable small enterprises that fit neither the startup model nor the conventional small business template. Whether that band widens significantly or remains occupied by unusually capable individual operators is the question worth watching.