Investment in AI-driven drug discovery continues to grow, reflecting how strongly the industry believes faster in silico discovery can reshape the front end of R&D. The first half of 2026 has brought a steady run of new AI platform deals, some centered on small molecule discovery and others pointing to a broader buildout of AI capabilities across pharmaceutical research. Greater discovery speed also raises the cost of getting development wrong once a candidate enters formulation and manufacturing.
The bottleneck in AI drug discovery
Faster discovery and more computational power do not make molecules easier to develop. As AI expands chemical space and surfaces more complex small-molecule candidates, sponsors and early-phase development partners still face the same developability questions that determine whether an asset can move forward.
Poor solubility, limited bioavailability, stability concerns, and manufacturability constraints still shape what happens next. For sponsors working with poorly soluble APIs, that remains the real bottleneck. AI may improve how quickly molecules are found and prioritized, but development strategy still helps shape which candidates can become clinic-ready products.
When discovery moves faster than development
AI can help sponsors identify promising candidates faster, but faster discovery does not make those candidates easier to formulate or manufacture. Greater discovery output often means more programs encounter familiar developability problems later. As discovery pushes further into complex chemical space, more candidates arrive with poor aqueous solubility, limited bioavailability, high melting points, poor solvent solubility, or processing constraints that narrow the available formulation path. Generative AI platforms are typically optimized to maximize target potency and selectivity rather than developability, which biases their output toward higher molecular weight, higher logP, and reduced aqueous solubility; territory where physicochemical liabilities tend to concentrate.
A molecule may look strong in early screening but still become difficult once formulation scientists and development partners begin building a stable dosage form, generating meaningful exposure, or designing a process that can hold up beyond the bench. Those issues are not secondary. They shape whether a program can move efficiently toward the clinic and whether the product can be manufactured in a way that is scalable and repeatable.
Legacy screening assumptions matter more now because many of them were built around a narrower range of compounds. Applied too rigidly, those assumptions can push sponsors to deprioritize molecules that may still have a viable path forward with the right formulation strategy.
In AI-driven development, that is a costly risk to take, because some of the most promising candidates may also be the least likely to fit older ideas of what a developable molecule should look like. For early-phase teams, the job is not simply to screen for what looks easiest to handle. It is to understand whether a strong molecule can be supported through the right combination of formulation design, biopharm assessment, and manufacturing planning.
What early development needs to answer sooner
Greater discovery speed puts more weight on early development decisions. If a project is going to move quickly without creating larger problems later, sponsors need clear answers to a few questions much earlier. Those include:
- Can the molecule reach therapeutic exposure at a practical dose?
- Can it remain stable through processing and storage?
- Is there a manufacturing path that can hold up as the program advances?
These questions often determine where delays emerge, where material gets wasted, and where teams end up revisiting work that should have been settled sooner. In that environment, early development has less room to operate as a supporting function and needs to play a more direct role in showing whether a promising candidate can move forward in a way that is both scientifically sound and operationally realistic.
Why development still depends on proof
An integrated decision loop must anchor development, not prediction alone. AI can help narrow options and improve prioritization, but it cannot show whether a candidate will hold up once formulation work begins. That still requires a sequence of in vitro screening, in silico PBPK modeling, and in vivo confirmation.
Early in development, formulation scientists and development partners can use that work to define practical limits around dissolution, absorption, and stability, then assess whether a formulation approach is improving performance in a way that can support the program. Done well, it leads to stronger decisions, more efficient use of limited API, and fewer late-stage surprises.
Where formulation flexibility matters most
As more AI-surfaced candidates move into development, formulation strategy becomes more important because many of these compounds fall outside the comfort zone of standard approaches. The pattern is especially clear when a candidate brings poor solvent solubility, high melting points, limited thermal tolerance, or narrow processing windows. In those cases, sponsors need more than a default path. They need enough formulation latitude to work around the properties of the molecule without giving up what made it valuable.
Expanded formulation latitude often requires attention to:
- Bioavailability-enabling approaches: For some poorly soluble compounds, advanced amorphous solid dispersions may offer a practical path to improved absorption.
- Processing flexibility: When heat, solvents, or material properties limit conventional routes, solvent-free fusion processing, or other nonstandard approaches may open viable options.
- Broader formulation design space: Greater excipient flexibility and more tailored multi-component systems can give formulators room to balance performance, stability, and manufacturability.
Development still decides what advances
AI will continue changing how small molecules are discovered, but the candidates that generate the most interest in silico still have to succeed under real development conditions. They need formulation strategies that can support performance and manufacturing paths that can hold up as programs advance.
As discovery becomes faster, the pressure on development only grows. Pharmaceutical organizations that manage that transition well will be better positioned to carry strong molecules into the clinic without compromising what made them worth pursuing.
Photo: Yuuji, Getty Images
Elizabeth Hickman serves as Chief Executive Officer at AustinPx, a leading contract development and manufacturing organization (CDMO) specializing in bioavailability enhancement of orally delivered small molecule drug candidates. Elizabeth brings over two decades of expertise in the biotech and pharmaceutical sectors, formerly serving in leadership positions for West Pharmaceutical Services, Catalent, and Pii. Elizabeth holds a BA in Microbiology from The University of Texas at Austin and an MBA in Marketing from San Diego State University.
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.
