Why a single drug costs more than a building | AI Learning Guide

The key idea here is that the headline cost of a drug is really the cost of all the drugs that failed alongside it. A company funds several candidates at once because it cannot know in advance which one will work, and the one that succeeds has to pay for the rest. Notice the three drivers: the timeline is long, the spending is committed early and cannot be taken back, and the odds at each stage are low. That is why making molecule design faster only trims one early line in a very large budget.

0:00 / 0:00

Auto-play

The cost is the cost of the failures

A company cannot know in advance which candidate will work, so it funds several at once. The approved drug must, in effect, pay for the ones that did not make it. This is why the price of a successful drug looks disconnected from the cost of making that particular molecule.

What makes the total so large

Time: money is committed years before any revenue, and the clock keeps running through every phase.
Irreversibility: preclinical work, manufacturing scale-up, and early human trials must be paid for before the drug's fate is known.
Low success probability: most candidates fail, so each success carries the cost of its failed siblings.

Cheaper, faster molecule design shrinks one early line in the budget. It does not shrink the trials, the manufacturing, or the failures — which is why the economics can absorb an AI breakthrough and still look unchanged from the outside.