Nvidia is weighing a guarantee on Nscale's unfinished data center. The question that asks how you lead before the work is scoped.
On October 31, 2026, TechCrunch reported that Nvidia is in talks to guarantee billions in debt financing for Nscale's planned Anthropic data center in West Virginia, citing The Information and noting that Bloomberg later corroborated the report. The project is enormous on paper: according to Nscale's IPO filing, Anthropic has signed a $44.6 billion contract for the first building alone. The mechanism under discussion is a residual value guarantee, under which Nvidia would "step in if Nscale defaulted on construction payments, temporarily take over the facility, and help find a buyer and new tenant," according to The Information. The arrangement is proposed, not agreed. But its shape is telling, because TechCrunch notes that Nvidia's earlier support for cloud providers "has mostly covered leases and chip purchases, not construction."
Look at what the guarantee is. It is not a plan for the campus. It is a structure for what happens if the plan fails, written before anyone knows whether it will. TechCrunch reports that Nscale had about 25,000 advanced AI chips running in August against the nearly 500,000 it plans to deploy, that runoff from the construction site has flooded nearby homes more than once, and that the project won state certification in September only after months of negotiations. The scope is contested and the delivery record is thin, yet the money has to move. So the parties are not pretending to resolve the ambiguity. They are deciding who owns which unknown, under what trigger, and what happens next. That is the discipline behind the question an interviewer asks when they want a manager to describe how they structured a team through an ambiguous, unscoped project. The question is not about the plan. It is about what you built to survive not having one.
Why they ask it
Senior roles are mostly unscoped work. By the time a project is well-defined it can be handed to someone more junior. The interviewer wants to know whether you can make a team productive before the problem is understood, and whether you did it with structure rather than charisma or overtime. They listen for three things: how you decided what was knowable now versus later, how you assigned ownership of the unknowns and not just the known tasks, and how you built in the moment when the team would revisit its own shape. A manager who can describe all three has led through ambiguity. One who can only describe the outcome has been lucky, or has forgotten how it felt.
The trap
The weak answer rewrites history. It describes the project as it looked at the end, with a crisp goal and a tidy org chart, then narrates execution against it. The ambiguity disappears in the first sentence, which is the part the interviewer wanted to hear about. A second version keeps the ambiguity but treats it as weather: requirements kept changing, so the team stayed flexible. Flexibility is not a structure. It is the absence of one. The strong answer names what was genuinely unknown at the start, says what the manager chose to lock down anyway, and explains why that choice let the team move. It treats the structure as a bet with a trigger for revisiting it, much as a guarantee has a trigger for stepping in.
Applying STAR-T
Situation. Set up the ambiguity specifically, not atmospherically. "Our platform team was asked to make the data layer ready for the next product, but the product had not been designed, the data team was mid-reorganization, and two of my six engineers were new." Three concrete unknowns beat one vague adjective. The interviewer should feel why a normal roadmap was impossible.
Task. Separate what you were accountable for from what you could not yet define. "I owed leadership a team that could absorb whatever the product turned out to need, and a credible date for when we would know enough to commit." A date for knowing is a real commitment, and it is the one available when the scope is not.
Action. This is where the structure lives, so spend most of your time here. "I split the team into two pairs and a pod of two. One pair owned the interfaces we were confident would exist whatever the product became, and shipped against them. The other pair owned the open questions: each unknown had a named owner, a hypothesis, and a two-week deadline to become a decision or an escalation. The new engineers sat in the pod, paired with me on the integration spike, which gave them context fast and gave me a direct read on the real risk." Then name the revisit point. "I told the team we would re-cut the structure at week six, when the product design was due, and I wrote down what would make us re-cut it earlier." A guarantee works because its trigger is defined in advance. A team structure should be the same.
Result. Be honest about what the structure bought. "At week six the spec arrived and half of it matched our interfaces, so that pair was two sprints ahead. Two open questions had become decisions, one was escalated and killed a feature, and one was still open. We re-cut into three feature pods and shipped the first integration a month before launch." Include the unresolved item. A result with no loose ends sounds like the trap.
Trade-off. Name what the structure cost. "Keeping two engineers on open questions for six weeks meant less visible output, and one stakeholder read the team as slow. I would make the same call, but I would publish the question backlog earlier so the work was legible." A manager who can price their own structure is more credible than one who claims it was free.
The follow-up that breaks weak answers
"What would have made you change the structure earlier than planned?" A candidate who built the structure as a bet answers immediately, because the triggers were written down: a design decision landing early, an owner missing the two-week deadline twice, a dependency falling through. A candidate narrating in hindsight stalls, because the structure in their story was never provisional. It was what they ended up with, presented as what they chose. The interviewer hears the difference in the first second of silence, and the rest of the answer rarely recovers from it.
Score your answer against the director’s bar
Q: Describe how you structured a team through an ambiguous, unscoped project.
Bank the unscoped project you led, with its triggers and its one loose end, and rehearse it until the follow-up feels easy. Try it free →