The surge in AI-enabled discovery has moved from promise to infrastructure. In late 2025, NVIDIA announced a new partnership to build what it described as the most powerful supercomputer owned and operated in the pharmaceutical sector.1
This “AI factory” is intended to accelerate work in drug discovery and, potentially, downstream functions as well, signalling how aggressively the industry is investing to compress early R&D cycles and widen what can be explored.
Computational advances have dramatically increased the speed of candidate identification and prioritisation.
But once a molecule transitions from predictive modelling into experimental development, its true property profile begins to dictate the path forward.
Compounds that are promising in silico often present solubility limitations, high melt temperatures or stability constraints that complicate formulation and manufacturing, reports Dave Miller (pictured), Chief Scientific Officer at AustinPx.
Not yet a Subscriber?
This is a small extract of the full article which is available ONLY to premium content subscribers. Click below to get premium content on Manufacturing Chemist.
Subscribe now Already a subscriber? Sign in here.