GLP-1s will gross nearly three times more revenue this year than every frontier AI lab combined.
Pharma is still the most hated industry in America. AI is on its way to becoming one. Same playbook - love the product, resent the company.
Different math underneath.
About 90% of clinical-stage drug programs fail. Tufts puts the all-in cost of one approved drug at roughly $2.6B, most of it paid for molecules that never reached a patient. The profit people resent is a survivor-bias premium on a decade of failures.
Pharma will spend more on programs that fail this year than every AI startup has raised this year.
The cycle that gets lost: that premium is what funds the next class. Statins, HIV antivirals, mRNA - every category we now take for granted was hated at peak before it became life-saving, low cost, and widely available. The same money the public resents in year five is what builds the standard of care in year fifteen.
GLP-1s sit exactly there.
The interesting question is what bends the 90%.
The only thing that has a chance of meaningfully changing that failure math is AI inside the drug discovery cycle. Not AI sold separately as a productivity tool. AI fused into target selection, trial design, and patient stratification.
The two industries getting profit-shamed in 2026 are also the two that have to merge for the math to actually move.
What's the first drug AI helps make that we'll all eventually take for granted?
13.4K views · 113 engagements
View on LinkedIn