When AI Becomes the Activist’s Thesis

Going into the 2026 proxy season, many expected artificial intelligence to continue shaping shareholder activism. It did, although the focus shifted. The debate initially centered on governance—specifically, whether boards had sufficient AI expertise and whether companies were appropriately disclosing and managing AI risk. Those topics remain important to investors, and the expectations behind them continue to rise. As the 2026 proxy season unfolded, a number of activist campaigns shifted their emphasis. Rather than treating AI as primarily a governance theme, activists have increasingly asked what AI means for a target’s competitive position, revenue durability, cost structure, capital spending, and, ultimately, valuation.

Valuation Emphasis

The pivot in shareholder activism’s emphasis from governance to valuation impacts not just what activists argue for and seek from companies, but how boards can respond. A governance argument can be met with a proposed governance solution, for example, adding a director with relevant expertise, enhancing disclosure, or amending committee mandates. A valuation argument requires different types of analysis because it is a claim about what the business is worth and what the board should do to realize that value. Two campaigns from the past year illustrate the point. In May, Findell Capital Management argued that the market had wrongly written off Figma, Inc. as an AI “laggard.” Three months later, Engine Capital told the board of EPAM Systems, Inc. that the market had priced the company as “a relative loser” amid fears of AI disruption and that the board should respond by either buying back stock aggressively or running a sale process. Neither activist group was asking the board to improve its understanding of AI; each was claiming that the market had misjudged the company’s future and that the board should act on that view, whether through capital allocation, a clearer investor narrative, or a transaction.

That does not mean governance has dropped out of AI-themed activism campaigns. An activist pressing a valuation thesis will still question whether the board has the requisite expertise and the process to evaluate and oversee AI matters from a strategic and risk management perspective. A board that cannot show both invites the governance critique on top of a valuation one.

The speed of this change in emphasis is notable, but the transition itself should not come as a surprise. Over the past year, the technology sector has been defined by a sharp divergence: record spending on data centers and the power and cooling they require, alongside steep declines in the shares of companies that investors fear AI will displace, such as application software. Earlier this year, Barclays reported that a record 52 percent of the 136 campaigns launched globally in the first half of 2026 targeted industrial and technology companies. That concentration is not a coincidence. Companies exposed to AI disruption, whether as adopters of the technology or as suppliers of the infrastructure that enables it, are increasingly the targets of activist campaigns. The timing also fits a well-established pattern. In each recent market dislocation—the COVID-19 drawdown, the 2022 rate repricing, the 2023 regional banking crisis, and the 2025 tariff shock—campaign launches rose materially in the six months after volatility subsided. Sharp repricings create a gap between a company’s trading price and its perceived value, and identifying ways to close that gap is the core of the activist business model. The AI-driven repricing has followed the same pattern, with a key difference: earlier gaps between price and value were opened by shocks external to the business, whereas the current repricing reflects the market’s judgment about whether AI will strengthen or erode the business itself, and so an activist arguing that a company is undervalued is also arguing that the market has misjudged the technology’s effect on the company.

New Premise, Familiar Demands

Although the AI aspect of these activism campaigns is new, the fundamental demands are familiar. For example, once the AI vocabulary is set aside, many of the past year’s highest-profile campaign letters sought steps drawn from the standard activist toolkit. Engine Capital asked EPAM Systems for a $750 million accelerated share repurchase or, failing that, a sale process. Randian Capital asked Snap to adopt a leaner and more efficient operating model. Elliott Investment Management asked Daikin Industries to close its margin gap with peers, launch an ambitious share repurchase program, and conduct a strategic review of its non-core businesses. Lauro Asset Management asked Segro plc to spin off its data center platform.

Traditional efficiency campaigns have generally been negotiable because they are arguments about degree: the activist claims costs are too high, management claims costs are coming down, the sides disagree over how much and how fast, and the matter typically settles with a cost plan, sometimes a buyback, and, often, a board seat. These campaigns follow a predictable course, and well-prepared boards can manage them.

The premise of this year’s AI campaigns is different and frequently less readily negotiable. Rather than arguing about the degree of a company’s inefficiency, these campaigns have asked where the business will stand once the AI transition has played out and whether it will lead or lag in a transition whose outcome is uncertain. That debate may leave less room for negotiation to the extent the answer is a binary one: either the company will emerge from the transition stronger, or it will be displaced. A settlement agreement between a company and an activist that adds directors to a board cannot necessarily resolve a disagreement of that kind; at most, it may buy time for the competing forecasts to be tested. As a result, these campaigns force boards to confront the hardest version of the valuation question. If the company’s future is uncertain, are shareholders better served by a certain return of capital today or by an uncertain, but potentially much larger, payoff if the company succeeds in the transition? And if the share price is depressed, is that because the market doubts the technology or because it doubts management’s ability to capitalize on it?

Lessons From the Past

Questions like these go to the fundamentals of the business, but boards have faced them before. One of the closest modern parallels to the current AI transition is the widespread adoption of the internet in the late 1990s. At that time, the pressure on boards came not from activists but from the market itself, and in particular from the fear of being labeled a laggard. Several prominent companies responded with large and, in hindsight, costly strategic bets. In November 1999, Disney shareholders approved a separate class of stock to track the performance of its internet business, GO.com, and on January 29, 2001, Disney announced it would convert the tracking stock back into Disney common and shut down the portal. AT&T decided to become a cable and wireless company, spending more than $100 billion, and its share price fell sharply as a result. Mattel’s 1999 acquisition of The Learning Company for $3.5 billion in stock, meant to carry Mattel into high-tech toys, led to the CEO’s resignation in 2000 under pressure from shareholders.

In the years that followed, those dot-com-era decisions became the activist theses of the next cycle. For example, Carl Icahn’s 2006 campaign at Time Warner was, at its core, an argument for undoing the AOL merger. One takeaway for boards is that decisions made today under pressure to act on AI will become the evidence in the campaigns that follow the next downturn. The same history also suggests where AI activism is likely to go next.

Activism follows market repricing, and activists often target the companies the market has left behind before they turn to the leaders. In 1999, activists were largely absent from the dot-coms themselves. They focused instead on the “old economy” companies that the market had abandoned, arguing for buybacks and separations, because those were the only companies at which an undervaluation argument could credibly be made. After the bubble burst, activists turned to the technology sector itself, by which point hundreds of fallen companies were sitting on cash at compressed multiples. That second wave helped build the modern activist industry: the number of activist hedge fund campaigns in the United States more than doubled between 2001 and 2006.

The parallel may be useful for considering where AI-themed activism may be headed. The first wave has already arrived: campaigns in the past year were concentrated in repriced software companies and industrial companies whose AI exposure the market has not recognized. The second wave is yet to come. The companies building and supplying AI infrastructure may be relatively insulated for now, but only for as long as the capital spending cycle holds. If that spending slows, the same activists who today argue that AI demand is an under-leveraged tailwind may argue that the spending should stop and the cash should be returned to shareholders. Boards in both groups should be preparing now, and how they should prepare depends on which group they are in.

Preparing Now

No public company should assume it is immune from an AI-themed campaign, and the first step in preparing for one is a vulnerability assessment. That assessment may begin with the question: how would an activist categorize this company today? Each answer points to a different type of campaign. A company priced as an AI casualty invites a “prove the market wrong” campaign built around capital return. A company whose AI exposure the market has not credited invites a monetize-or-sell campaign. A company with a slow AI adoption story invites a cost campaign, and a company with an AI-adjacent asset invites a sum-of-the-parts campaign that may end in an unsolicited bid. The assessment should replicate the analysis an activist would perform, using the same public data: peer benchmarking on margins and capital allocation, sum-of-the-parts valuation, and the return the company is earning on its AI spending. With that assessment in hand, boards should focus on two further priorities:

• Owning the narrative around AI by clearly articulating an evidence-based strategy. The board should ensure that the company has developed an AI thesis grounded in evidence—what the technology means for revenue durability, pricing power, cost structure, and competitive position, and what the company is doing in response—and that this thesis, rather than a generic statement about “leveraging AI,” is what investors hear. Where the answer is uncertain, the company can say so and explain how it expects to resolve the uncertainty. Getting there is itself a governance exercise in part: the board should be clear about where oversight of AI strategy and risk sits (whether at the full board or a committee), what it expects management to report, and how often.
• Knowing shareholder sentiment on AI before an activist arrives. When speaking to investors, companies should be informed about lines of activist criticism they could face in the future that have not yet been at the forefront of commentary by analysts and other market observers. Without providing a roadmap for dissent, companies can use these conversations to explore sentiments that investors might not otherwise express and to explain the company’s views and plans on topics it expects might become issues, before they do. If an activist raises such an issue before the company does, the company may find itself on the back foot when addressing the issue with investors. Thoughtful engagement with activism in mind can help the company control the narrative, better shape investors’ perspectives on potentially sensitive issues before activists magnify them, and set the tone and the groundwork for the company’s response. Revisiting the shareholder engagement program is a worthwhile exercise even when the existing program is strong—as is the case at many companies, particularly large ones. The point is to periodically test whether the engagement program is in sync with other aspects of activism preparedness and to stay ahead of the narrative.

The boards that came through the last technology transition in the strongest position were not only those that made the right bet about the internet. They were ones that made deliberate, informed decisions on their own timetable and explained those decisions clearly to investors. AI activism asks the same of directors today, with the added challenge that the central question is now about the technology itself. No one can predict with certainty where the AI transition will lead, but the process by which a board forms its own view and acts on it is within the board’s control. How well the board and management run that process and how they explain it to investors may become decisive factors, whether in heading off an activist campaign or in responding to one.