Katerra Burned $2 Billion Before Anyone Asked If It Actually Worked
Michael Marks had already run Flextronics and sat briefly in Tesla's top seat when he decided the construction industry
Michael Marks had already run Flextronics and sat briefly in Tesla’s top seat when he decided the construction industry was next. His pitch to investors was blunt: buildings should be assembled the way iPhones are, on a production line, not stitched together by dozens of subcontractors who’d never worked with each other before. SoftBank liked it enough to put in more than $2 billion. Six years later, that company, Katerra, filed for Chapter 11, walked away from apartment blocks mid-build, and shut its doors without paying staff their final leave.
The Katerra collapse gets told most often as a cautionary tale about construction being harder to disrupt than software. That’s true as far as it goes, but it misses the more useful lesson. Katerra didn’t fail because construction resisted technology. It failed because it tried to become too many companies at once, before it had proven it could be any single one of them well.
Build a skyscraper like an iPhone
Founded in 2015 by Michael Marks, formerly the CEO of Flextronics and briefly Tesla’s interim chief executive, alongside Fritz Wolff, Katerra’s pitch was vertical integration. Rather than subcontracting out design, manufacturing and building, it would own every piece: architecture, engineering, factory-built wall and window components, project management, even its own building materials. The logic wasn’t crazy. Construction genuinely is fragmented, inefficient and resistant to standardisation, and a $12 trillion global industry with that little technological change in decades looked, on paper, exactly like the kind of market ripe for a platform play.
Investors agreed. By 2018, SoftBank had put in $865 million, a stake that eventually grew past $2.4 billion. Katerra opened a 200,000 square foot factory in Phoenix to manufacture entire prefabricated walls, complete with wiring and plumbing already installed. The Wall Street Journal ran a piece asking why you wouldn’t want to build a skyscraper the way you build an iPhone. For a while, Katerra looked like the future of the industry.
Then Greensill collapsed too
The trouble wasn’t a single bad decision. It was the accumulation of several, most of them variations on the same mistake: adding complexity faster than the organisation could absorb it. Katerra acquired dozens of smaller firms, architecture practices, factories, software teams, and tried to integrate all of them into one operation at the same time it was still working out how any individual piece should run. It drifted from its “Salesforce of construction” focus into software and Internet of Things projects that had little to do with the core business. Leadership churned, with a rotating cast of chief executives making it hard for any single strategic direction to actually stick.
By the time the pandemic hit in 2020, the ground beneath what would become the Katerra collapse was already fragile. Its factories struggled to fill orders, and general contractors and developers, the customers it needed to actually win over, remained sceptical of handing their projects to a single vertically integrated vendor rather than the subcontractor relationships they already trusted. A board and SEC investigation into the company’s accounting practices added another problem entirely. When Katerra’s lender, Greensill Capital, itself collapsed in early 2021, the company lost access to the capital it needed to keep going, and SoftBank’s final rescue attempt wasn’t enough to save it.
More zeroes, same mistake
Strip away the construction-industry specifics and what’s left is a familiar shape. Katerra scaled its ambitions, its headcount and its acquisitions far faster than it scaled its ability to actually run what it had built. Every new factory, every new acquisition, every new product line added operational weight the company hadn’t yet proven it could carry, and it kept adding weight instead of pausing to test whether the foundation underneath could hold it.
This is the same trap that catches businesses at every scale, just with more zeroes attached. A ten-person agency that takes on a client requiring the output of thirty people is making a smaller version of the same bet Katerra made when it tried to run manufacturing, software, architecture and project management as one integrated machine from year one. The difference isn’t the mistake. It’s the size of the cheque required to keep it going before the mistake becomes visible.
Prove it before you scale it
The Katerra collapse is often used to argue that some industries just resist disruption, and there’s something to that. But the more useful reading is about sequencing. Katerra tried to prove every part of its model simultaneously rather than proving one piece, getting it genuinely working, and only then adding the next. A founder with a fraction of Katerra’s funding faces the identical temptation on a smaller scale: take on the bigger client, open the second location, launch the adjacent product line, all before the first version of the business has actually demonstrated it can run smoothly under pressure.
The businesses that avoid a Katerra-style collapse tend to treat proof as a prerequisite for expansion rather than something to sort out after the fact. They ask whether the current version of the business, at its current size, is actually working, before adding the next layer of complexity on top of it. Katerra had the capital to paper over that question for six years. Most companies don’t get anywhere near that long, which makes the discipline to ask it early considerably more valuable than the ability to raise the next round.cks winners and losers whether a company means it to or not, and most companies never sit down and choose that outcome on purpose.more useful. A confident consensus is not the same as a correct one, and the more unanimous the certainty, the fewer people are actually checking it. A man once had to ask permission simply to attempt something everyone expected him to bungle, and he built a career, and an empire, out of being underestimated by people who never did the work to find out if they were right.
