"So lets assume the drug cost 180 million to develop." The trouble is, you not only have to cover expenses for development of this, very successful drug, but a

"So lets assume the drug cost 180 million to develop." The trouble is, you not only have to cover expenses for development of this, very successful drug, but also for 50 others that you never heard of, because they were duds, but the necessary trials still cost money. Pharma is a numbers game. Most attempts fail, and you need to cover the cost of all the failures, otherwise you can't try out anything new. It is possible that AI will help here, modelling drugs in virtual environments and recommending the most hopeful molecules. If it can be done, the ratio of failures to successes could improve.