The Two Plants
What gets preserved when the architecture stops working — and what gets rebuilt.
Two American manufacturing stories surfaced within days of each other last week. They were reported in different sections of different newspapers, by different reporters. They belong together.
In late February, Stanley Black & Decker announced it was closing its last manufacturing plant in New Britain, Connecticut. The company was founded there in 1843. The city was named, in effect, after the company. “The Hardware City” was not branding, it was geography. For 183 years, Stanley made things in New Britain. The closure ends that. Roughly 300 manufacturing jobs go. The corporate headquarters stays, with its 400 office employees. The brand stays. The making leaves.
In early May, the Wall Street Journal disclosed that Ford had been running a secret team. Ford itself called it a skunk works, borrowing the Lockheed term from the Second World War. The group has been run out of an office park in Irvine, California, for nearly four years. Seventeen engineers at the start. Now several hundred. The team had been deliberately isolated from Ford’s existing operations so that it could redesign, from scratch, the architecture by which Ford builds a vehicle. Their first product is an electric pickup truck, targeted at a $30,000 price point and a 2027 launch, to be built at a $2 billion overhauled plant in Louisville, Kentucky. The CEO, Jim Farley, has called it Ford’s Model T moment. A phrase he chose carefully, and which I will return to.
Two plants. Two companies. Two answers to the same question.
The question is this: when the architecture that built the company stops producing the outcomes the company exists to produce, what does the institution rebuild? And what does it preserve?
Stanley preserved the brand. Ford rebuilt the architecture.
The Stanley story is easy to misread. The temptation is to read it as a parable about offshoring, or about American manufacturing decline, or about the unsentimental logic of capital. None of those readings is wrong. None of them is the structural story.
The structural story is what the company chose to remove.
Stanley’s explanation for the closure was technical and unsentimental. There has been, in their words, a structural decline in demand for single-sided tape measures. Lasers measure distance now. Phones measure distance. The product the New Britain plant made for generations has been quietly replaced by tools that do not require a hooked metal blade and a coiled spring. Reading the company’s statement, you might conclude that the plant closure is simply the end-of-life of a product category. The tape measure became obsolete. The plant that made it became unnecessary. The institution adapted.
But notice what the institution did not do.
It did not ask whether the architecture that made tape measures in New Britain for 183 years could have been redirected to make something else. It did not ask whether the workforce that had spent generations developing manufacturing competence — in metal forming, in coiled-spring tension, in the small disciplines of producing thirty million tape measures a year — could have been the foundation for a different product. It did not ask whether the building itself, the equipment, the supplier networks, the muscle memory of the place, was an asset that justified a redesign rather than a dismantling.
It treated the architecture as the variable. It treated the brand as the constant.
The headquarters stayed. The logo stayed. The DeWalt and Craftsman and Stanley names will continue to appear on tools sold across the country, produced by contract manufacturers in lower-cost geographies. The company is still Stanley Black & Decker. It still measures itself by the brands it owns. But the making, the physical and structural foundation that produced the brand in the first place, has now been formally separated from the company that bears the name.
The architecture has been removed. The brand has been preserved.
This is not unusual. It is, in fact, the dominant pattern of American manufacturing over the last fifty years. The financial logic is clean. The architecture is heavy. The brand is light. Heavy assets are taxed, depreciated, exposed to labor disputes, vulnerable to geography. Light assets (trademarks, design files, distribution contracts) travel well. A company that preserves the brand and removes the architecture looks more efficient on every quarterly statement than a company that does the opposite.
The cost of this trade is invisible on the balance sheet. It is paid in a different register entirely.
The Ford story would not have happened without the Lightning failure.
This is worth saying clearly, because it is the part of the story that gets edited out of the heroic version. The Ford F-150 Lightning, launched in 2022 with enormous fanfare, was the company’s first serious electric truck. Its price climbed past $77,000 before the company effectively stopped trying to sell it. It lost billions. Last year, Ford began winding it down.
By every conventional measure, this was the moment for Ford to do what Stanley did with tape measures. Read the failure as a category exit. The high-end electric pickup is not for us. Preserve the brand. Remove the architecture that produced the loss. Return the capital. The shareholders would have applauded. The quarter would have looked better. The institution would have survived intact, smaller, and less ambitious.
That is not what happened.
Around the same time the Lightning was beginning to fail in the market, Ford had quietly stood up a small team in California. The number Ford and the Journal keep returning to is seventeen. Seventeen engineers at the beginning, working in near-total isolation from Detroit. The team was led by Alan Clarke, an engineer who had spent more than a decade at Tesla helping develop the Model S and the Cybertruck. The team’s mandate was deliberately, almost absurdly, structural: rebuild, from first principles, the way Ford designs and manufactures a vehicle.
Not the vehicle. The architecture by which the vehicle gets made.
Reading the Journal account, what is striking is not the ambition of the project but the specificity of the architectural changes. The team eliminated thousands of feet of copper wiring. They eliminated the traditional assembly line and replaced it with what they call an assembly tree: modules built in parallel and joined late in the process, rather than a single sequential line on which complexity accumulates with each station. They reduced parts by twenty percent compared to a Mustang Mach-E. Fasteners by twenty-five percent. Workstations by forty percent. Assembly time, dock to dock, by fifteen percent.
Read those numbers slowly. Each of them represents a category of decision that, inside the existing Ford architecture, had been treated as fixed for decades. The number of fasteners in a vehicle is not a question that gets re-asked every year. It is inherited from the architecture. To reduce it by twenty-five percent, you have to be willing to question the architecture itself. Every supplier contract, every tolerance, every assumption about how a panel joins to a frame.
This is what it actually means to rebuild architecture. Not to launch a new product. To re-ask the questions the architecture has been answering by default.
And the most consequential decision Ford made was not technical. It was organizational.
The team was kept outside the existing company on purpose. For two years, the engineers in Irvine had near-total isolation from Detroit. They were not subjected to the procurement reviews, the program timelines, the parts-commonality mandates, or the inherited assumptions that make it functionally impossible to redesign architecture from inside an existing operating company. The skunk works model, taken from Lockheed’s wartime aircraft development, exists precisely because a large institution cannot rebuild its own foundations while it is also running its existing business. The foundations are load-bearing. You cannot dig under them while you are standing on them.
The hard part of Ford’s bet was not the design. It was the integration. At a certain point, the isolated architecture has to come back inside the company and meet the existing one. The Journal describes this collision with a candor that is rare in corporate reporting. The Tesla and Apple veterans were not welcomed. The Ford engineers did not initially trust the outsiders. There were, in the account of one of the executives, months spent resolving conflicts down to the level of how to calculate the size of a part. The truck’s open front end nearly buckled during paint, requiring engineers to slip into a Michigan plant at three in the morning to test a section of vehicle that almost no one in the company knew existed.
This is what rebuilt architecture looks like in practice. It is slow, expensive, internally bruising, and visible only in retrospect, if it works.
It might not work. Ford has missed before. The Lightning is the most recent miss but not the only one. The bet on Louisville is a $5 billion bet, and the company has openly described it as a bet rather than a plan. There is no version of this essay that requires the truck to succeed for the architectural argument to hold. The argument is about what the institution chose to do before the outcome was known.
Stanley, faced with a product whose architecture had stopped working, removed the architecture and preserved the brand.
Ford, faced with a product whose architecture had stopped working, isolated a parallel architecture, let it cannibalize a century of manufacturing practice, and is now integrating it back into the company at considerable cost and considerable risk.
The difference is not courage. The difference is what the institution chose to treat as the constant.
When a piece of architecture stops producing the outcomes it was built to produce, there are two coherent responses available to the institution.
The first is to declare the architecture obsolete and remove it. This is the Stanley move. It preserves the brand, the headquarters, the trademark, the distribution agreements, the intellectual property. It is fast. It is legible to capital markets. It can be executed by a competent management team in a single fiscal year. The cost, which arrives later and is paid by people who are not on the call when the decision is made, is that the institution loses the capacity to ever rebuild that kind of architecture again. The competence is in the workforce, in the suppliers, in the building, in the muscle memory of the place. Once that competence has been dispersed, it does not reassemble.
The second is to declare the architecture obsolete and rebuild it. This is the Ford move. It preserves the institution’s capacity to produce the kind of outcome the architecture was originally for. It is slow. It is illegible to capital markets for the years it takes to demonstrate results. It requires the company to tolerate a parallel architecture inside itself, with all the cultural and political friction that produces. The cost is paid up front, in cash and in organizational pain. The return, if it arrives, is that the institution still knows how to make things on the other side of the disruption.
These are not equally available choices to every institution at every moment. Stanley’s plant produced tape measures whose category was genuinely contracting. Ford’s plant is being rebuilt for a category whose long-term direction is genuinely contested. The decisions are not symmetric. The arguments against rebuilding the New Britain plant are real. The arguments for the Louisville plant could turn out to be wrong.
But the structural question the two stories pose, side by side, is the one worth holding.
When the architecture stops working, what does the institution treat as the variable, and what does it treat as the constant?
Stanley treated the architecture as the variable and the brand as the constant. Ford, at least in this one consequential instance, treated the brand as the variable and the architecture as the constant. The brand of Ford’s electric vehicle program, Mach-E, Lightning, now the unnamed Louisville truck, has shifted three times in five years. The architecture is what the company is now betting will survive.
Farley’s phrase, Model T moment, is not casual. The Model T was not, in its own time, a new car. It was a new architecture: the moving assembly line, interchangeable parts, the vertically integrated supply chain that allowed a vehicle to be sold for a price most American families could afford. The Model T did not make Ford a brand. It made Ford an architecture. Every subsequent Ford for half a century was a variation on a structural foundation laid by that one decision.
The question Farley is implicitly asking, by invoking it, is whether the company still knows how to do that kind of thing. Whether the muscle is still there. Whether a 123-year-old industrial institution can rebuild itself from the inside without being destroyed in the process. The answer is not yet known. It will not be known for years.
But the asking is the architecture.
There is a version of the American industrial story in which both companies are simply doing what their circumstances require. Stanley is winding down a category that has lost its purpose. Ford is opening a category that requires a different foundation. Each, on its own terms, is rational.
That version is true. It is also incomplete.
The fuller version is that an industrial economy made up entirely of institutions that respond to architectural obsolescence by removing architecture eventually loses the capacity to rebuild any architecture at all. The competence to make things does not live in the brand. It lives in the building, in the workforce, in the supplier networks, in the small daily disciplines of doing the thing. When the making leaves, what is left is the brand. And a brand without an architecture beneath it is, in the long run, an asset that depreciates faster than its owners realize.
Ford may fail with the Louisville truck. The $5 billion bet may not produce the $30,000 vehicle. The skunk works architecture may not survive integration with the rest of the company. The category may turn out smaller, later, and harder than the strategy assumed, which is the same failure mode Honda just admitted to with its $10 billion electric-vehicle writedown.
None of that changes the structural fact that Ford asked the harder question.
When the architecture that built the institution stopped producing the outcomes the institution exists to produce, Ford rebuilt the architecture. The company is now finding out whether it can still do that.
Stanley made the other choice. The plant in New Britain is closed. The headquarters remains. The brand is intact. And the country has one fewer place that knows how to make a tape measure.
The institution that rebuilds the architecture sometimes fails at the rebuild. The institution that removes the architecture has already failed at something more durable, whether or not it shows up in the quarter.
That is the difference. The outcomes follow.