May 4, 2026
Via Email and Certified Mail
Board of Directors
Innventure, Inc.
6900 Tavistock Lakes Blvd, Suite 400
Orlando, FL 32827

Dear Members of the Board:
In February, we wrote you a hard letter. We meant every word of it.

This letter does two things: Closes our February campaign on the record on which we opened it and explains why. With the structural questions of that letter now substantially addressed, we believe Innventure represents one of the most compelling risk-adjusted opportunities in the public markets today and why we are remaining holders in size.

Following the company’s response to the Schedule 13D filed by Commonwealth Asset Management on February 18, we were compelled to take the lead on direct, sustained engagement with this Board and management team — via serious conversationsregarding overhead, capital allocation, internal and public relations, and governance. Those conversations were candid. They were specific. They were, at times, uncomfortable. They produced results.

We are writing today, voluntarily and on the record, to acknowledge that this Board engaged seriously, we believe management is executing, and the Company is in a fundamentally different place than it was when we sent that February letter.

That deserves to be said as plainly as the criticism was.

The receipts.

In February, we called a ~14% G&A reduction the bare minimum dressed up as progress. G&A is now down approximately 61% year-over-year on the comparison management has disclosed, with cash G&A at the parent down approximately 56%. Management has guided 2026 to track the new lower run rate, with more to come. That is structural. That is real.

In February, we said the parent balance sheet should stop funding every operating company through dilution at the Innventure level. We believe the necessity for that dynamic is over. AeroFlexx and Refinity are now raising directly, where dilution belongs. Accelsius’ $665 million Series B, anchored by Johnson Controls and Legrand, fully funds its scale-up without another dollar needed from Innventure. The April 23 capital allocation framework formalized a $250–$350 million parent capital buffer, with surplus capital returned to shareholders. That is the policy a disciplined holding company runs. Thisis now the Company’s stated policy, time will tell.

In February, we said three executive directors on a nine-person board was not oversight. Bruce Brown is now Lead Independent Director. Daniel Hennessy has resigned. John Hewitt has been elected to fill that seat as an independent Class I director. Catriona Fallon has been nominated to stand for election as an independent Class II director at the June 17 Annual Meeting and is expected to serve on the Audit Committee. Dave Yablunosky’s term expires at the 2026 Annual Meeting and the Company has indicated he will not stand for re-election as a management director. The math has changed: the executive-director count comes down, the independent count goes up, and two seats have been filled with operators chosen for what this Company is becoming, not for what it has been.

Hewitt is not a governance hire. He is a former senior operating executive of Vertiv — the public-company incumbent in data center infrastructure — and currently CEO of Robertshaw Controls, and he has been engaged with the Accelsius operating company in the period leading into this appointment. A former senior executive of the category incumbent is now an independent director of the parent of the disruptor. That is the kind of signal markets eventually price. Fallon is the public-company CFO and audit chair this platform should have had from the day it listed: Hitachi Vantara, Silver Spring Networks, audit committee chair at two public boards. The financial oversight gap we identified in February is being closed by the right person.

In February, we singled out a bubble bath partnership as the entire stated case for AeroFlexx. AeroFlexx now has Aveda as its first prestige beauty anchor and the credibility to raise outside capital. Refinity has produced a metric ton of product from real plastic waste at pilot scale and is moving to commercial demonstration.

Eleven weeks ago, we said the path was clear and short. The Company is now visibly walking it.

How this got done.

After the February letter, this Board and management team did something many boards under public pressure do not: they engaged. Real meetings. Real exchanges of information. Real willingness to be challenged on specifics — the G&A line items, the financing architecture, the AeroFlexx funding question, the Lead Independent Director appointment, the composition of the slate announced last week, the capital allocation framework, the disclosure cadence.

We brought our case. They tested it, pushed back where they disagreed, and moved decisively where they agreed. The output is what we believe to be the company you see today: leaner cost base, cleaner capital structure, a parent-level allocation policy that returns surplus to shareholders, a Board that is molding into a genuine independent - in fact and not just in form, and a disclosure posture that engages
substantively rather than defensively.

This was not Ascent dictating to a Board. It was a Board willing to do the work, with shareholders willing to do it alongside them. That is what constructive engagement is supposed to look like.

We are grateful for the way this Board and management team engaged with us. We do not say that often. We mean it here.

We worked, alongside this Board, to fix the structure around the asset. We want to spend the rest of this letter on the asset itself, because the structure was the only thing standing between Innventure’s share price and the value of what Innventure actually owns.

Why we are doubling down on Accelsius.

We want to be unambiguous about why we remain holders, we think the market is still materially mispricing what this Company actually owns. Our conviction is not philosophical. It is built on the asset, the market, the moment and the math.

The asset. Accelsius is, in our view, the single best managed and pure-play public exposure to two-phase direct-to-chip liquid cooling that exists today. NeuCool removes more than 4,500 watts of heat per socket — a threshold that single-phase direct-to-chip systems cannot economically reach and that air cooling is multiple generations away from approaching. Two-phase is not a marginal improvement on single-phase; it is the architecture that becomes mandatory once chip-level thermal design power crosses the critical point where dielectric phase change is the only viable heat-transfer mechanism. This has been stated by multiple hyperscalers across the industry alarmingly, it is not a generic thesis. That envelope has now been crossed by current-generation accelerators and will be crossed by a wide margin by what comes next. The patent position around NeuCool, the reliability data hyperscalers and colos require before deploying at
scale, and the integrated-rack form factor are not commodity advantages — they are the reasons two industrial strategics led the Series B and the reasons hyperscale-class operators are signing production orders rather than running pilots. NeuCool is not a science project. It is shipping product, deployed under
contract, with a $1 billion-plus pipeline behind it and $50 million-plus in production-volume bookings booked in Q1 2026 alone, per the Company’s public disclosure. The April 20 launch of NeuCool IR150 — the industry’s first integrated rack for two-phase liquid cooling — debuted at NVIDIA GTC and was met with a single-day move in INV of more than 30%. That reaction was not noise. That was the market beginning to price the product Accelsius now actually sells.

The market. The thermal threshold of modern AI infrastructure has broken air cooling. Current-generation NVIDIA accelerators draw on the order of 700W per GPU; next-generation parts will draw substantially more. Per published industry forecasts, liquid cooling penetration of AI servers is expected to expand from
a small minority of deployments today to the majority within the next two years, and the broader data center liquid cooling market is forecast to grow at a 20%+ CAGR through the end of the decade. Within that, two-phase — the segment Accelsius leads — is consistently identified as the fastest-growing subsegment. This is not a story we have to project ourselves into. It is happening, and it is being underwritten today by hyperscalers, colos, and AI infrastructure operators who have to solve thermal density before they can deploy anything else.

The moment. The 300MW Ontario deployment is the largest two-phase, direct-to-chip deployment on record. It is also a reference customer of a kind that does not stay isolated. The $65 million Series B led by Johnson Controls and Legrand was strategic, not financial — two of the most operationally sophisticated industrial players in the world chose Accelsius as their two-phase platform. Accelsius is targeted to exit 2026 at cash-flow breakeven, implying a December annualized run rate of over $100 million. Its cash on hand is sufficient to reach that threshold without further parent-company capital. That trajectory — from de minimis revenue to $100 million+ run rate by year-end 2026 — produce, what we believe to be the kind of inflection that has historically produced multi-bagger outcomes when the underlying asset is a controlling stake in a category leader.

The math. Accelsius’ post-Series B valuation is approximately $665 million post-money. Innventure consolidates Accelsius and remains the controlling shareholder. Based on our analysis of disclosed terms, Innventure’s stake in Accelsius alone, marked at the Series B post-money valuation with the addition of
the implied weighted pipeline immediately following, encompasses Innventure’s entire current market capitalization. AeroFlexx — a sustainable packaging company with recent traction and multiple consecutive quarters of revenue recognition — is in there discounted if not free. Refinity — with validated
technology and active capital formation — is in there for free. The parent-level operating goodwill and capital buffer is in there for free. The Innventure platform itself — the create-and-operate engine that is now producing Accelsius — is in there for free.

That is not an opinion. That is what the numbers say at today's share price. The discount exists for one reason: the issues this Board has now begun to fix were, until recently, structural impediments to the market underwriting the asset. Those impediments are coming off. The asset has not changed. The structure around the asset has.

But let us be precise. The discount was never one thing. It was three things arriving at the same time.

Governance — a Board that answered to no one and communicated even less. Opacity — a company that asked investors to trust a process it refused to explain. And timing — the unavoidable reality that markets as large as liquid cooling for hyperscale AI do not confirm overnight. The hyperscalers move slowly, then
all at once. Hyperscalers and multinational adoption? Our view, it is a matter of wanting to be first to be second. Accelsius is built for that. They waited for it. They have now been named the lead. You can be first or second – welcome to the show.

Three discounts. All three resolving. Not theoretically. On the tape. In the wins. In the names.

Mark-to-market events we will be watching.

We are going inactive into a window in which several specific events should, in our view, materially compress the discount this stock currently carries:
— Accelsius’ exit from 2026 at cash-flow breakeven and a ~$100 million annualized run rate with multiple hyperscaler adoptions.
— The 300MW Ontario campus moving from agreement to active deployment, with the associated reference effects on the rest of the pipeline.
— AeroFlexx’s direct capital raise at the operating-company level, which would establish an independent valuation mark and validate the new financing architecture.
— The first capital return to shareholders under the $250–$350 million parent buffer policy, which would convert the framework from policy into practice.
— Hewitt and Fallon’s first full cycle on the Board, including the work product of an Audit Committee chaired by an operator with two public-company audit chair seats already under her belt with another board member exchange in sights.

Any one of these, if not a combination, are catalysts.

What would bring us back.

We are confident enough in the direction of this Company to stand down today. We are also clear-eyed enough to be specific about what would change that view. We would re-engage publicly if (i) The capital allocation framework announced April 23 is abandoned or reversed — including any return to funding operating companies at the parent level beyond Accelsius, at irresponsible prices or terms; (ii) The governance refresh is undermined — any retreat from the Board independence now in motion; (iii) c Corporate overhead reverses course and expands meaningfully from the run rate established in the back half of 2025; or (iv) The Board takes any action that dilutes Innventure's ownership of Accelsius outside the ordinary course of Accelsius' own capital formation. To be clear: we expect Innventure to raise capital again. We expect them to do it at pricing, terms, and timing that serve shareholders. That is the standard. We do not expect to invoke any of these. We are stating them so there is no ambiguity about what
continued constructive engagement requires.

Our position.

We have always preferred to engage privately. We are activists when we have to be, and only for as long as we have to be. The conditions that forced us public in February have been addressed or are being addressed.

Following last week’s board action, we intend to file to go inactive.

We are not concerned to add to our position if and when we feel it makes sense. We will continue the same direct, private engagement with the Board and management. We do so with confidence in the direction the Company is heading and in the people getting it there.

Two things in closing.

First, going inactive is not going silent. The G&A reduction is structural progress, not a finish line. The governance refresh announced last week is a commitment now executed, not a completed reconstitution of every aspect of how this Board functions. Year-end 2026 and 2028 cash-flow targets are guidance, not results. We will continue to engage. We will continue to hold the Board to its own commitments on the same standard the Company applies to the businesses it builds. The Board has earned the benefit of the doubt it had not earned in February. It has not earned a free pass, and we do not believe it is asking for
one.

Second, we believe the opportunity from here is exceptional. Accelsius is a proven business at the center of a generational infrastructure transformation. Innventure is now structured — finally — to let public shareholders fully participate in what Accelsius and the broader platform are creating. The discount this stock has carried exists precisely because of the issues this Board has now begun to fix. As execution continues, we believe that discount should narrow.

Respect was earned this quarter. We are saying so on the same record we used to demand it.

Respectfully submitted,
/s/ Jonathan Loeffler
Managing Director, Ascent Capital Partners
On behalf of Ascent Capital Partners, Ascent X Innventure
(collectively with its affiliates)

Source: https://www.sec.gov/Archives/edgar/data/2001557/000204177726000004/AscentLetterInnventurefinal.pdf