In late November, Roche's (RHHBY 0.77%) subsidiary Genentech and the chipmaker Nvidia (NVDA 6.18%) agreed to collaborate on developing new medicines using artificial intelligence (AI) over the next few years. While the pair didn't announce the financial terms of their partnership, the move is just one of a slew of similar deals between big pharmas and businesses making AI platforms to improve the drug discovery and development process.

Now, investors are left grappling with the question of what to make of the announcement, given that few concrete details are available. Let's clear up some of the confusion by putting the new collaboration into context and discussing some of its likely goals and impacts in the near future.

This revolution is just getting started

In theory, AI has a lot to offer drug developers like Roche. The drug development process can take upwards of a decade, not to mention billions of dollars in research and development (R&D) expenses. Even under the best conditions, candidates for new medicines can fail to perform as anticipated when tested in clinical trials, and failures are the rule rather than the exception.

If using AI in some capacity could slash costs, reduce the number of stumbles, or speed up the process, the cost savings could be tremendous. It's plausible that it'd lead to safer and more effective drugs, too.

But few big pharmas have the in-house technical know-how necessary to develop or implement such an AI system independently. Hence, Roche is teaming up with Nvidia, a leading AI company building a suite of services explicitly to help biopharma businesses operationalize AI.

With Nvidia's BioNeMo AI platform for drug development, customers can, among many other functions, predict how a given molecule or therapy candidate will interact with its intended physiological targets, paving the way for picking higher-quality leads to start working with in the laboratory. They can also train AI models to address their specific needs across most stages of the development process.

Roche isn't putting all its eggs in one basket when it comes to AI. It's also collaborating with Recursion Pharmaceuticals, a biotech focusing on using AI to empower drug development efforts, partially owned by Nvidia. In terms of how important the collaboration is, consider that Roche paid $150 million up front. It also committed to paying as much as $300 million in milestones and royalties for up to 40 different programs, making for a total potential consideration of more than $12 billion over the next 10 years or so.

Given that it currently has around $8 billion in cash, equivalents, and short-term investments on hand, it's reasonable to assume it's taking a bit of a shotgun approach to the programs it's working on with Recursion, with the expectation being that, ultimately, only a few of its attempts will be worth the full payment.

Recursion isn't Roche's only bet on scrappy young biotechs using AI, though. Since late 2020, it's been partnering with private biotech Genesis Therapeutics on unspecified programs, as well as others like Reverie Labs and Dyno Therapeutics. It also bought another company, Prescient Design, in 2021. With so many different attempts at generating some value from AI, it's quite likely the business will find a measure of success with at least one.

Wait for a firmer plan or actual results

In broad terms, it's a plus that Roche is making strides to optimize its drug development operations with AI. Securing collaborations with capstone players in AI, like Nvidia, is unambiguously positive and could indeed make Genentech a more efficient drugmaker in the short term. If, over the next couple of years, the company can advance more of its pre-clinical programs into clinical trials than before, it'll be evidence that AI is leading to higher productivity in a financially impactful way.

Management's alternative course of action -- waiting to dabble until the AI technology for drug development is proven useful -- would allow competitors uncontested access to this generation's most promising new digital toolset. Obviously, that doesn't sound like a good plan at all.

But as you may have noticed, there aren't too many clues as to which disease areas or biotechnologies will be the focus of its investments and partnerships in AI. So, there's little to reassure shareholders that resources are being allocated well.

In the near future, at least one big pharma player will very likely find a way to turn their use of AI into a competitive advantage. Presently, investors are mostly left with generalities and statements of broad ambitions from Roche's management rather than concrete roadmaps for how it will turn its use of AI into a durable driver of shareholder value. Until that changes, the best course of action is to hold off on buying (or selling) shares.