This past 2 years, venture funding has been increasingly moving toward companies that are building AI solutions for different industries. As a matter of fact, by some accounts, AI companies now account for more than half of the venture capital funding and majority of the mega deals (funding rounds of over $100M) are going to AI-related companies. These include companies building infrastructure for a future with AI (e.g., data centers, compute capacity, etc) and those that are building horizontal and vertical AI applications. Horizontal AI refers to applications that are sector agnostic and are focused on foundational layers such as the large language model builders (e.g., OpenAI, Gemini, etc) and those that focus on functions such as customer service (e.g., call centers,) general agentic AI, and voice AI. Vertical applications refer to those products that focus on specific industries such as finance and healthcare and build solutions that address specific workflows in those industries. Vertical applications are especially relevant in an industry like healthcare since the data is very unique and the workflows are very specialized.
For AI to fulfill its lofty promise and provide a massive lift to our productivity and economic growth, it will need to perform up to a certain level in the real world. That means it will need to be accurate and reliable in its output and either help people complete activities faster and better or completely automate some activities. If it falls short of that level of performance, its impact will be limited as it will not be relied upon enough for it to make a difference. In a recent moderated debate about whether AI will replace doctors, one of the topics we discussed was whether the markets, private and public, were over-valuing AI companies, meaning that the valuations that these companies are commanding are too high. My response was that if the markets are lavishing premium valuations on these companies because of their short-term promise, then the companies are over-valued. It will take time for us to figure out how to incorporate this technology into the different industries, especially complicated ones like healthcare. Also, the performance of AI models is a work-in-progress. As such, if the models don’t immediately perform up to the hype or if the adoption rates are lower due to incomplete workflow integration or lack of training, this does not mean AI won’t be a game-changer long-term. So, if those valuations are based on the long-term promise of AI, those companies may actually be under-valued.
If the horizontal and vertical applications of AI are coming, who will then make the most money from building and launching these products? Is it the incumbents who have tons of money, technology DNA, and relationships with businesses and consumers? Or, will it be the nimble startups that can focus on specific areas and build great products that the incumbents are too slow and big to act on? How do you make money betting on this question? If you think Microsoft and Google will use their enormous resources to build AI products and release them to their existing customers, then buying their stock now makes the most sense. If you think the likes of OpenAI (hard to think of them as a startup now but they actually are!) will be the next titans of tech, then putting your money into their private funding rounds is the way to go. This topic is of great importance and there is no shortage of opinions on it. I’ve been studying what has transpired in the space for the last few years and looking for trends and patterns that can answer this question. It is clear to me that there is not a binary answer to this question. What I see is that given the potential of AI and the potential economic gains, nobody will sit it out. It is true that the OpenAI and Claude’s of the world are new tech companies that have built foundation models and have already carved out a prominent place for themselves in the tech landscape, but Google, Amazon, Microsoft, and other big tech are also making massive bets in AI models, infrastructure, and development tools.
In healthcare, the first wave of innovation in AI has seen companies that are creating applications for many administrative and clinical workflows and raise significant dollars. Applications as varied as note taking for doctors to drug discovery have fetched hundreds of million dollars of investments per round to deliver on this promise. These are new companies such as Abridge, Ambiance, Nym, Cohere, etc that are developing point solutions (at least in their initial versions) for clinical workflows (e.g., note taking) or administrative workflows (e.g., prior authorization.) That means startups are creating value by taking on focused opportunities to improve healthcare and building new solutions for them. Score one for the startups vs. the incumbents in healthcare! Abridge has raised over $500M to go after note-taking. Abridge and Ambiance are delivering their note-taking solutions in the Epic EHR and have existing integrations with the behemoth of health IT. That means Epic customers can enjoy AI solutions within their existing system of record and not have to switch to a new system to gain benefits from these new technologies. Looks like the customers and the AI startups are the real winners here because the customers can continue to use their EHR of choice and benefit from AI functionalities within it and AI startups can innovate and deliver their solutions in the customers’ environment of choice.
Companies like Recursion and Altos Labs, not big tech or pharma, have raised billions of dollars to do better drug discovery using AI. As I write in my book, AI Doctor, biopharma has been trying to get into this game through partnerships with these startups. The promise of the mountains of data that already exists in these large biopharma companies combined with the AI technologies of these drug discovery startups seems like a marriage made in heaven!! Looks like the innovators will reap significant economic benefits and the incumbents will have new drugs to develop and commercialize. And, people will get speedier access to life-saving treatments. No winners or losers here!
The reality, unfortunately, is more complicated! As we will discuss in part II of this series, while there has been significant innovation, funding, and some initial success by the AI startups, the incumbents are not sitting idle and making their environment and data available to these startups so they can be the big winners of the AI era. We will delve into recent developments that can signal a fierce battle ahead between the incumbents and the startups and examine the implications for the customers, founders, and investors.
