When the activist already knows more than you do: lessons from Elliott at Honeywell

When Elliott Investment Management disclosed a $5 billion stake in Honeywell International in November 2024, it did not arrive with a vague complaint about "underperformance." It arrived with a fully formed structural argument: that Honeywell's conglomerate model was destroying value, that peers trading on focused multiples were leaving Honeywell shareholders roughly 50 to 75 percent of upside on the table, and that a breakup into aerospace and automation businesses was the logical correction. The thesis was detailed, cross-referenced with comparable transactions, and designed to be difficult to rebut without months of internal analysis that Elliott had already done. By early 2025, Honeywell had announced it would spin off its advanced materials division. By February 2026, it had committed to the full separation of its aerospace business.

The speed of that capitulation, and the quality of Elliott's opening position, reflects something the CNBC reporting on AI-powered hedge funds makes plain: firms are building investment theses using AI agentsAI agentsAgentic AI refers to AI systems that pursue goals autonomously by planning, taking actions through tools, and adapting based on results, with minimal step-by-step human direction.View full definition → that ingest financial filings, management commentary, competitor data, and sell-side research simultaneously, without the cognitive fatigue or time constraints of a human analyst. One fund described by CNBC has structured its entire operation around AI agents handling research, monitoring, and portfolio construction. When a firm with that architecture targets your company, the information asymmetry that CFOs once relied on to buy time has largely collapsed.

What Honeywell's management actually did

The initial response from Honeywell's then-CEO Vimal Kapur and the board was not to fight publicly. That choice mattered more than it might appear. The company acknowledged Elliott's concerns about portfolio complexity, requested time to work through the analysis with their advisors, and entered a structured dialogue rather than issuing a defensive press release about long-term strategy.

On the CFO side, the mechanics were specific. Honeywell's finance team accelerated an ongoing strategic review that had been proceeding at a normal planning cadence and reframed it as a capital allocation question rather than an organizational one. That reframing allowed the CFO to present to the board a set of scenarios with explicit value bridge analysis: what each business unit contributed to the current trading multiple, what it would trade at as a standalone, and what the transaction costs and stranded cost exposure would look like across separation scenarios.

Honeywell also moved to preempt further shareholder coordination. Elliott's 5 billion dollar position is large, but it is not a majority. What makes activists effective is their ability to convince index funds, pension managers, and other institutional holders to vote alongside them or pressure management separately. Honeywell's IR team ran a concentrated engagement program with its top 20 institutional holders in the weeks after Elliott's disclosure, presenting the same scenario analysis before Elliott could brief those holders with its own numbers. The goal was not to defend the status quo but to ensure that the board's response, when it came, was not a surprise that looked reactive.

The advanced materials spinoff announcement in February 2025 was structured as a CFO-led value creation narrative, not a concession. Management pointed to tax-efficient separation mechanics, a clean balance sheet profile for the RemainCo, and a timeline that preserved operating continuity. The aerospace separation announcement followed once management had worked through carve-out accounting complexity and debt allocation.

The results

Honeywell's stock had traded in a range roughly between $190 and $220 for much of 2023 and early 2024. After Elliott's disclosure and the subsequent strategic announcements, shares moved above $230 and held that level through early 2026. The specific attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → between Elliott's involvement, broader market conditions, and management execution is genuinely difficult to isolate, so those numbers should be read as directional rather than causal.

What is clearer is the governance outcome. Honeywell did not face a proxy fight. Elliott did not nominate directors. The separation timeline is proceeding without the disruption, legal cost, and management distraction that contested board elections typically produce. For context, the proxy fight at Disney in 2024, where Nelson Peltz ultimately withdrew after spending an estimated $25 million on the campaign, illustrates what the alternative looks like in time and resource terms even when the activist loses.

Anthropic and OpenAI are both reportedly pushing for investment-grade credit ratings ahead of potential IPOs, according to the Financial Times. That detail is worth noting in this context because it signals that AI-infrastructure companies are moving toward the capital markets toolkit that makes sustained activist-style research even more accessible: cheaper financing, longer time horizons, and deeper balance sheets to support positions.

What transfers, and where the context differs

The Honeywell case produces four concrete lessons for CFOs managing an activist situation or trying to prevent one.

First, the information gap is gone. Assume any serious activist targeting your company has already modeled your segment-level EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.View full definition →, your peer trading multiples, and your capital allocation history in more detail than your own quarterly board materials show. Prepare your value bridge analysis before the 13D lands, not after.

Second, institutional holder relationships are not an IR soft skill. They are a structural defense. If you cannot get a senior call scheduled with your top 20 holders within 72 hours of a hostile disclosure, that relationship is not at the level it needs to be. Elliott's ability to move other shareholders depends directly on whether those shareholders hear your narrative first or second.

Third, CFOs should separate the question of whether an activist is right from the question of how to respond. Honeywell's management appears to have concluded, privately, that Elliott's structural argument had merit. Acting on that conclusion early, on their own terms and timeline, cost them less than resistance would have. That is a financial judgment, not a governance one.

Fourth, AI-driven research changes the timing assumption. Historic playbooks assumed activists needed months to build a credible thesis. AI agents compress that. CFOs running strategic reviews on a three-year cycle need to shorten that cadence or accept that an outside firm may complete the analysis first and present it to your board in a 13D filing.

Where context differs: Honeywell had the balance sheet, the operating scale, and the institutional shareholder base to absorb an activist engagement with professional resources on all sides. Smaller companies face the same AI-driven research capacity from activists but with far fewer IR and advisory resources to counter it. In that context, the priority is even more basic: keep your own financial story clean, your segment disclosure granular, and your institutional relationships current. A CFO who can answer Elliott-style questions before they are asked is far harder to embarrass publicly than one who cannot.

The Honeywell outcome was not inevitable. It reflects a finance team that chose to engage on substance rather than posture, and moved fast enough to shape the process. That is the repeatable variable.