The world's largest tech companies have joined the biggest pharmaceutical companies and several cutting-edge biotechs in efforts to use AI to discover new, better drugs. But so far, hype is outpacing reality.

Why it matters: The list of conditions and diseases that new drugs could treat or cure is enormous.

• We've been hearing about AI-enabled drug discovery for years, but no product has made it onto the market and there are fewer in the development pipeline than all the talk would suggest.
• At the same time, while pointing an algorithm at Alzheimer's disease may still be fantasy, there are signs that AI can alleviate human suffering and help people live better lives.

Driving the news: There have been several recent megadeals between drug and AI companies, and boldface tech names are suddenly pitching themselves as drug discovery partners. My Pro Deals colleague Katherine Davis had a good rundown this week:

• Merck on Wednesday signed a deal worth up to $1 billion with Google Cloud to use its AI platform for R&D, manufacturing and commercialization.
• Amazon's AWS launched a new application aimed at helping scientists design and test novel drugs more quickly and precisely.
• Novo Nordisk and OpenAI also announced a drug discovery partnership.
• Nvidia and Eli Lilly announced a $1 billion drug development partnership in January, and more recently Anthropic reportedly purchased AI biotech startup Coefficient Bio for $400 million.

The challenge here is well-known: Finding and testing new drugs is time-consuming and expensive and has a huge failure rate.

• And even the experimental drugs that succeed often deliver incremental improvements on existing therapies or fall short of delivering actual cures.
• AI not only can shorten the discovery and development processes, saving drug companies money for more R&D, but also increase the likelihood that drug candidates succeed in clinical trials.

Reality check: The crowd claiming AI drug discovery is overhyped includes none other than Alex Zhavoronkov, CEO of Insilico Medicine — one of the biotechs that is leading the field.

• Zhavoronkov recently wrote a piece titled "The AI drug revolution is real but the hype around it isn't" that concluded AI won't "overwrite the laws of nature and eliminate all human disease in 10 years."
• "What AI is doing is transforming biological discovery from a slow, bespoke artisan craft into a highly scalable, compute-driven engine," he continued. "The precision medicines of tomorrow will arrive years earlier, cost a fraction of what they do today, and save millions of lives in the process."

Between the lines: Insilico's rentosertib, for a progressive lung condition called idiopathic pulmonary fibrosis, was the first instance in which AI enabled the discovery of a disease target and a compound to attack it, per a Nature Medicine article discussing the drug's phase 2 results last year.

• That essentially means AI helped both define the biological problem and find a solution.

I caught up with Zhavoronkov the other night to get his perspective on big tech companies like Amazon and OpenAI getting into drug discovery. He was pretty dismissive.

• "In my field, they are irrelevant," he told me. "If you haven't delivered a few [drug candidates] clearly and also explained how you did it in a serious paper, you are not my competitor. You may be hyping up, but where are the drugs?"

I also wanted his thoughts on where all of this is going. With all of these billions of dollars sloshing around, what's real and what isn't?

• The constraints of human trials proving an experimental drug's safety and efficacy aren't going away, he told me. Those take time and money.
• His biggest ambitions lie with the kinds of drugs being developed and how they can improve human lives — and not just by curing diseases.
• "We want to find another GLP-1 scale drug, or many drugs like that," he told me. "How can we develop a drug for Parkinson's, and then repurpose it into aging? How can we find a drug for, let's say, an OB-GYN disorder and repurpose it into productive longevity, postponing menopause, for example?"

The intrigue: Zhavoronkov told me that he's had to prioritize less-novel drugs so far because those were easier to sell to large pharmaceutical companies.

• The deals he's made on "moderate novelty drugs" have made Insilico more financially sustainable, and so now the company can aim for real breakthroughs.

The bottom line: "AI helps you to go from zero to this wonderful point called developmental candidates or preclinical candidates, where you have kind of created the gun, loaded the bullet, fired the bullet — and then after that it's kind of moving at the speed of traffic," Zhavoronkov said.

"The adoption of artificial intelligence (AI)-powered tools is accelerating rapidly across all layers of healthcare systems," Nature Medicine wrote in an editorial this week.

• "Yet evidence that AI tools create value for patients, providers or health systems remains scarce."

My thought bubble: That last line is big. The AI tools aren't just coming, they're here — and there's not much of a system for determining if and how well they work.