What it costs to make power cheap

Start with what Dell is rejecting. For most of a decade, home energy was sold as a premium product: rooftop solar, smart panels, backup generators, bought by people who wanted them and priced at a margin that reflected that want. Dell calls those product companies rather than infrastructure companies, and says their whole business model was oriented around a high-gross-margin green premium. His claim is that electricity is a commodity, that the best version of a commodity is the cheapest one, and that the moment you accept this the business you are in changes. Commodity businesses do not win on desirability. They win on a cost advantage that compounds, which in his framing means buying that advantage with technology and vertical integration and then extending it.

Vertical integration is sequenced rather than total, and he gives a crisp test for the sequencing: if you are buying five percent of the world’s supply of something, there is a good argument for making it yourself; at five ten-thousandths of a percent, there is not, absent some specific operational reason. Base is not mining lithium and is not in the cell business. It designs almost everything and does final assembly, test and pack. What the integration is really for is the coordination problem. Dell’s unit of optimization is the fully landed cost, which means the hardware design, the parts bill, the manufacturing, the logistics, the installation, the financing and the tax credits all move together. Solve for the cheapest parts and you may buy them somewhere that costs you a tax credit. Solve for the cheapest manufacturing and you may push cost into installation. This is why the company is organized by function rather than by profit-and-loss line: one finance team, one supply chain team, one hardware team, working across every market, so that a single group can see the whole equation at once. He is candid that this is a guess they may revise, and that it costs them something real: it demands people who can run at 120 percent, and everyone is in the office every day.

The same logic runs through the capital structure, and this is where the reasoning turns most sharply against intuition. The cost of capital is an input to the cost structure, which is an input to the price a customer pays, so lowering it is not a finance-department concern but a product one. The tempting move is to limit equity and limit dilution, and Dell says plenty of companies have tried that and not made it. His argument is the reverse: you have to accept that capital will be expensive for a while and raise a great deal of it anyway, because only scale drives returns up and the cost of capital down, and the cheaper capital is what eventually shows up on the bill. He is explicit that this is risky rather than clever. Raising $2.5 billion in three and a half years means you now have to turn it into something like ten, and if you do not, you have let a lot of people down. The path runs through instruments in order: venture debt early, tax credit monetization, then the debt capital markets as returns improve.

Residential batteries are where all of this gets applied, and the reason is speed rather than elegance. Dell argues they are the most scalable way to add capacity to the grid today, full stop, because the homes are already interconnected so there is no queue to wait in, and because the deployment itself can be run as a factory. There is a factory across the street making hardware and, as he puts it, a factory in the field installing it. On sourcing he is deliberately flat. Domestic manufacturing capacity has been a strategic advantage from day one, and depending on third parties is risky whether they sit in China or in Indiana. Last summer’s tax law tightened restrictions on Chinese components for credit eligibility, and his response is arithmetic rather than moral: parts from China are not inherently bad and are often cheap and good, but if buying them forfeits incentives available elsewhere, the spreadsheet says buy elsewhere. He offers a personal view that a global economy is a better economy, then sets it aside as explicitly not his to decide.