CEO leaning into the promise of AI for biotech? Not so fast, it can hurt firm value
By Xingrui (George) Mou
Readers of this article are also reading:
Big Value in Big Data? Here's how to signal the market about it
Executive Summary
While AI/ML innovation has swept into the communications of CEOs across many industries of late, an EY report has shown that talk about the promise of AI/ML to transform research in health sciences earnings/analyst calls has not been as pronounced despite corresponding regulatory approval submission filings increasing significantly.
Research published in the Strategic Entrepreneurship Journal suggests that health science firm CEOs that maintain high innovation rhetoric in shareholder letters get a 23.47% higher market value than average, but if that language is sudden or inconsistent over time, it negatively affects market value by 9.913%. However “follow-through” (ie regulatory FDA approval submissions that are 7x competitors and echo that innovation language) can mitigate the negative effects.
Health science CEOs in particular should be cautious about drastically ramping back up their AI innovation rhetoric beyond that of their competitors. Such increases should always be accompanied by significantly more product development than their competitors as well, which is becoming increasingly harder to do given FDA submission trends.
Impact: the promise of (re)entering the AI/ML drug race slower to develop now
Arguably the biggest splash at JP Morgan’s Healthcare Conference in January 2024 was NVIDIA’s announcement that they were partnering with Recursion and Amgen to beta test an AI/ML platform for novel drug discovery – and the talk was unsurprisingly thick with innovation rhetoric: “the future”, “groundbreaking work”, “incredible.” With the promise of assisting to identify therapeutic targets that are challenging to be discovered by traditional drug screening approaches, thus shortening drug development timeline and cutting research costs, it is indeed hard to imagine divorcing the AI/ML solutions from that language.
The global market value for AI/ML-generated drugs was $0.6 billion in 2022 and is estimated to grow continuously from 2023 to 2027 at a CAGR of 45.7%, reaching $4.0 billion in 2027 (“AI In Drug Discovery Market Size and Companies, Healthcare Industry News, LinkedIn, 2023”). For instance, AI/ML is gaining pronounced attention due to its superior capability to identify common disease signals from a complicated cell biology network, while predicting how various drug molecules can interact with the common disease markers. Such processes can significantly boost the drug discovery timeline compared to traditional drug screening techniques. It is very natural for a health science CEO to feel pressure to address this range of innovation, particularly since the launch of AI chatbots like ChatGPT in 2022 Q4 has rejuvenated the AI conversation.
AI/ML has gained a significant increase in attention within CEO communications to shareholders spanning all range of sectors - financial services, software engineering, manufacturing, as well as the drug discovery and medical device market. The October 2023 CEO Outlook Report published by EY found that the health science industry had the 2nd highest count of AI or GenAI mentions in 2022 Q1 earnings/analyst calls, making up 15% of all six industries analyzed in the report (“Is the AI buzz creating too much noise for CEO’s to cut through?” EY CEO Outlook Pulse, October 2023).
Surprisingly, mentions of AI/ML in health science experienced a sharp drop between 2022 Q1 and 2022 Q4 of around 40%, and though rebounding by Q3 2023 had only reached back to previous levels while all the other industries had doubled or tripled previous mentions. Health science CEO’s were clearly more hesitant to publicly embrace the language of this innovation this time around – and to stay around the average for any one firm (rather than drifting above one standard deviation from it) means similarly staying moderated in that language.
Yet since AI/ML models have been recently approved by US FDA as alternative models to animal models for pre-clinical testing (FDA Modernization Act 2.0, FDA, 2022), the FDA has reported a drastically increased number of new drug applications (170 submissions) based on AI/ML in 2022 alone (“FDA sees rapid uptick in drug and biologic submissions with AI/ML components”, Regulatory Focus, 2023). So for a firm to meet their AI/ML rhetoric with new drug/device submissions at a clip many multiples above their competitors’ average will only get harder - thus the major mitigation against shareholder blowback less available.
Findings: health science CEO’s “cheap-talk” on innovation has real consequences
The healthcare industry can sometimes be less desirable to investors due to a combination of stringent regulatory policies, long product development timelines, and significant sunk costs. Therefore, to maintain high investor confidence and interest in a company over that time, leaders in the pharma and device spaces must sometimes adopt aggressive innovation language about their “next big thing.” But how much of that talk is too much? A recent paper published in Strategic Entrepreneurship Journal suggests that maintaining a high level of this rhetoric (“entrepreneurship orientation” (EO) rhetoric) – through which top managers in public healthcare companies demonstrate innovation, proactiveness, and risk-taking to their shareholders – can help them to outperform their competitors in certain indicators of market value, but with great risk. If this language is introduced suddenly or inconsistently over time, the market responds by significantly punishing the firm’s market value. This market punishment is mitigated by follow-through on the rhetoric with actual novel drug or device submissions to the US FDA (“entry commitment” (EC)).
In this study, the authors analyzed all health science firms identified in the Compustat North America annual fundamental database according to SIC classification codes 283 for drugs and 384 for medical devices, along with any companies involved in clinical studies according to the U.S. National Library of Medicine, from years 2004 to 2012. Tobin’s Q (the ratio of market value to asset’s replacement cost) was calculated by the firm’s average stock price in each year; Tobin’s Q is similar to market-to-book value, which is far more commonly tracked, but is a multiple against current asset replacement costs and not past costs. The authors calculated EO rhetoric by counting all the instances of keywords in the CEO shareholder letters of that year that match known innovation language (e.g., “creativity”, “novelty”, “visionary”, “exploratory”, “bold”, “brash”, “competitive”, “aggressive”, “ambitious”, “opportunity”, “leading”, “forward-looking”, etc.). Additionally, to understand how hard-information such as actual follow-through can mitigate the effect of innovation language on Tobin’s Q, the authors analyzed EC through quantifying the number of new drug submissions or new medical device premarket approval applications in a 3-year period.
After controlling for the firm’s financial characteristics (total assets, return on assets, debt ratio, etc.), the research found that the health science firms outperform their competitors with a 23.47% higher Tobin’s Q (herein used interchangeably with “market value” in the rest of the article though not the same) if they demonstrate one standard deviation above the mean of all companies’ innovation rhetoric. However, a sudden increase in innovation rhetoric leads to depressed market value for the company – the presumption being that investors think the increase in rhetoric is being used to distract them from negative business information - with one standard deviation increase in innovation rhetoric leading to 9.913% reduction in Tobin’s Q year-to-year.
Additional analysis performed by the authors demonstrated that this negative effect of a sudden increase in innovation rhetoric can be mitigated by a concurrent increase in EC. Specifically, health science companies needed to submit 7x more than the industry average of new drugs or medical devices to ensure no negative effects on market value when innovation rhetoric is greatly increased. This study thus provides insights into strategies for health science firm CEO’s to avoid giving “cheap talk”, but rather implement an approach to introduce concurrent increase in innovation rhetoric that is commensurate with their actual “follow-through” commitment.
Recommendations: slow and steady wins the valuation race (until it doesn’t)
While AI/ML-assisted drug/device design is becoming a global trend on which health science CEO’s should act now, blindly ramping up their rhetoric on AI/ML innovation with shareholders without accompanying demonstrable innovations is likely unwise. Repeated mentions of innovation that are not backed by increased regulatory filings could hurt market value. Health science leadership should consider these factors in their innovation efforts and communications:
Day One: Monitoring shareholder letter communication references to AI/ML and other innovation promises RELATIVE to their competitors – being average is ok at first, but going too-far / too-fast will face blowback.
After: Aiming to increase innovation rhetoric slowly and keeping it consistent going forward, as sustaining extraordinary innovation rhetoric over time does often does provide a bump in market value if can be reached.
Finally: Ensuring the firm is truly equipped to submit new drug/device FDA applications that mention AI/ML at a rate that can far exceed competitors if they intend to increase the language above the industry standard.
____________________________
Xingrui (George) Mou is a PhD candidate in Biomedical Engineering at Duke University and a member of the Duke Advanced Degree Consulting Club. The research applications proposed in this article are solely the views of the author and do not necessarily reflect the views of the original academic journal article authors nor any individual member of our Editorial Board.