Mistral raised €3 billion in one round. The decision question that follows the money.
On September 8, 2026, TechCrunch reported that Mistral had raised €3 billion at a post-money valuation of €21 billion, in a round led by Samsung Electronics alongside the EQT-managed Scaleup Europe Fund and PSG Equity. The company said the capital would let it "scale its compute capacity, build infrastructure, accelerate commercial growth, and expand its international footprint," including a plan to build 1 GW of compute in Europe by 2030.
Strip the numbers and look at the shape. A company has just been handed several years of runway and a mandate to spend it in four directions at once. Every one of those directions is defensible. None of them can be done first without the others waiting. This is what a raise actually buys: not certainty, but a much larger set of decisions that all look reasonable, made faster than the evidence arrives.
Which is why funding news is interview news. Money moves the decisions from "what can we afford" to "what do we choose", and choosing under that kind of pressure is the thing senior interviews are built to test.
Why they ask it
It reads like a throwaway. It is not — it is a question about whether you have a method or a mood.
Most candidates answer with adjectives. "I'm data-driven." "I gather input, then I decide." Those describe a disposition, and every candidate claims the same one. The interviewer is listening for something structural: what you do when the data is thin, how you know a decision is reversible, who you involve and at what point, and what makes you stop deliberating. A method can be repeated by the team around you. A mood cannot, which is why it does not scale past the person who has it.
The question is also deliberately unbounded. There is no situation supplied, so how you narrow it is itself the first signal. A candidate who asks "important how — a hiring call, an architecture call, a call about a date?" has already demonstrated more decision hygiene than one who starts talking.
The trap
The frameworks trap is the common one. Naming a matrix — one-way and two-way doors, RAPID, a weighted scorecard — earns nothing on its own, because the interviewer has heard the vocabulary and cannot tell whether you have used it. Frameworks are only evidence when attached to a decision where the framework changed the outcome.
The second trap is the consensus answer: "I get everyone aligned and we decide together." Alignment is not a decision procedure. It is a description of the comfortable subset of decisions, the ones where everyone eventually agrees, and it quietly leaves out the ones that make the job hard. Senior roles exist for the decisions where alignment does not arrive.
The third is the story where you were right. A decision that worked out is weak evidence of good process, because good decisions and lucky ones look identical in the retelling. If you have a story where you followed your process and the outcome was still bad, and you would run the same process again, that is the stronger one to tell.
Applying STAR-T
Situation. Choose a decision that was genuinely close. If the right answer is obvious in your own retelling, you have picked a decision that made itself, and the question is unanswered. "We could keep the vendor for another year at triple the price, or migrate mid-quarter with the team we had" is a real fork.
Task. Say who owned the call and what the deadline was. Decisions without a clock are opinions; the clock is usually what made it hard.
Action. Show the machinery, briefly and in order. What you would need to believe for each option to be right. Which of those you could actually test in the time available, and what the test returned. Whether the door was one-way — and if it was, what you did to make it less so. Who had a veto and who had input, decided before the discussion rather than during it. Then the stop rule: "I gave it four days and said I'd choose on Thursday whether or not the load test finished." Naming the moment you stopped gathering is the single most senior thing in this answer, because indefinite gathering is how most real decisions actually fail.
Result. State the outcome plainly, and separate it from the process. "We migrated, it took three weeks longer than planned, and it was still the right call at the price we were quoted."
Trade-off. Name what you gave up and what you would keep. "Deciding on Thursday meant shipping without the full benchmark. I would do it again — the extra week of data would not have changed the price, which is what the decision actually turned on."
The follow-up that breaks weak answers
Expect "when did you last get one of these wrong?" It is a trap only for candidates whose method exists to produce correct answers.
The answer that works names a decision, names the flaw in the process rather than in the luck, and names the change: "I decided on incomplete information because I set the deadline before I understood what the test would take. Now I size the test first and set the date second." That is a person improving a method. "I trusted the wrong person's estimate" is a person assigning blame, and it tells the interviewer nothing they can transfer to their own team.
Score your answer against the director’s bar
Q: How do you make important decisions?
Bank a decision story where the outcome was mixed and the method still held, and rehearse it in L8 Loop until the stop rule is the clearest line in it. Try it free →