AstraZeneca's insitro deal puts 2.7% of its disclosed value up front

The other 97.3% of the disclosed total, $1.46 billion out of about $1.5 billion, is payable only if development and commercial milestones are met ^[9]. We think that makes the deal an option on target discovery. AstraZeneca puts up $40 million ^[1]; insitro uses its Virtual Human platform to identify causal disease biology and take AstraZeneca-nominated targets to development-candidate stage; and only then does AstraZeneca run and fund the clinical trials ^[3]. Until a target reaches that stage, there is no trial for AstraZeneca to fund under this agreement.

The counter-reading is that any early-stage discovery deal would look like this, because no target exists yet to pay for. Daphne Koller, insitro's founder and CEO, told Bloomberg that insitro has made considerable progress on the MASH side and expects to identify targets on a relatively near-term timeline, while the cardiovascular program will need more time ^[6]. That makes the liver program the first test of the structure and the heart program the slower one.

Whether big pharma still structures AI deals this way is a question about a pattern, and one deal is one observation. Insitro's expanded Bristol Myers Squibb partnership in ALS and its Eli Lilly work on models of small-molecule properties get a line each in the Reuters account ^[5].

The total itself is contested. Bloomberg described the deal as worth up to $1.86 billion ^[11], while the figures the companies disclosed add up to about $1.5 billion ^[12], a difference of roughly $360 million ^[13]. Measured against $1.86 billion, the $40 million would be 2.2% of the deal ^[10], against 2.7% on the companies' own figures ^[8].

AstraZeneca has reasons to buy early. Bloomberg reported that Wainua failed to show benefit in a heart muscle disease trial earlier this year and that the drugmaker pulled a mid-stage MASH candidate last year ^[7]. Sharon Barr, AstraZeneca's executive vice president of biopharmaceuticals R&D, gave Reuters a reason built on novelty. She told Reuters: "By the time we read about a new pathway or target (in major scientific journals), it is no longer novel. It's in the public domain and we're already a few steps too late." ^[4]

What would show we have the timing wrong is the milestone schedule. The reports do not break out how the $1.46 billion divides between development and commercial milestones ^[1]. If early development-candidate milestones hold most of it, the money arrives sooner than "back-loaded" suggests, though still only on delivery. And if the MASH targets Koller expects do not arrive, the only cash AstraZeneca has committed is the $40 million upfront ^[1].