Estimation question

How many computers does Google own?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Estimation for an unusual, large scale infrastructure question: can you decompose by known business units rather than guessing a single figure.

Break Google down by what needs computers: employees (laptops, monitors), and data centers (servers, which are a different category of computer at massive scale).
Estimate employee devices, assume roughly 180,000 employees, each with 1 to 2 devices, giving roughly 250,000 to 350,000 employee computers.
Estimate data center servers separately, since this dwarfs employee devices, assume Google operates on the order of several million servers globally to run Search, Cloud, YouTube, and other services.
Combine the two categories but note they answer very different implied questions, employee laptops versus server infrastructure, so the total depends heavily on which the interviewer means.
State the answer is dominated by servers, likely several million, versus employee devices in the low hundreds of thousands.
Flag the ambiguity explicitly and say you would confirm with the interviewer whether they mean office computers or infrastructure servers.

Identifies that the question is genuinely ambiguous between office devices and data center servers, and addresses both.
Uses a known reference point, Google's employee count, to ground the office device estimate.
States clearly which category dominates the total, showing scale judgment beyond just adding numbers.

Answering only for employee laptops and ignoring the vastly larger server infrastructure.
Giving a single number without flagging the ambiguity in what counts as a computer here.

Chapters of the AI PM course that teach what this question tests.