Mirion's Order Book Is Booming. The Market Is Paying for the Margin It Just Gave Up.

A 62% order surge and $1.25 billion backlog collide with a withdrawn 30% EBITDA target, front-loaded AI spend and a stock 44% below its peak.

MIR · Earnings Call · 2026-10-28

The order book is not the problem

Mirion's third quarter was, on the demand side, the best print the company has produced as a public company. Orders rose 62% year over year excluding M&A to $340 million; nuclear power orders more than doubled. Total orders including Paragon and Certrec were $386 million, up 85%. Backlog reached $1.25 billion, versus $816 million a year ago, and management laid out $735 million of backlog to recognize in Q4 2026 and full-year 2027 — more than $200 million above last year's visibility. The backlog build is not theoretical: current-year conversion is guided to $292 million versus $212 million, and 2027 conversion to $443 million versus $285 million. This is where the large order pipeline matters. Q3 included $140 million of large opportunity wins, or $122 million net of the Chinese debooking discussed last quarter. Management also flagged two France newbuild orders with EDF totaling more than $80 million that were awarded in September but are not in backlog pending EU funding. The nuclear demand is real and it is riding a global theme: the AI data center buildout has made reliable baseload power scarce, and utilities are responding with life extensions and uprates. Tomo Logan cited roughly 1.59 gigawatts of Constellation upgrades, 433 MW at Vistra and 190 MW at Calvert Cliffs, plus an NRC proposal to allow 40-year license renewals. But he was careful to separate Mirion from the data-center hype: "“these 3 waves are not contingent upon future events. Most specifically data center growth.” — Thomas D. Logan, Founder, Chairman, and CEO · 2026-10-28" About 80% of nuclear power revenue comes from the operating fleet, not newbuild timelines. The catch is conversion. Q3 organic revenue grew only 6.5%, slightly below the high-single-digit guide, even as orders surged. Q4 is guided to mid-to-high-teens organic growth, and management says the backlog gives 95% revenue coverage at the midpoint. Investors have heard backlog promises before; the next quarter has to show the revenue inflection rather than another step-up in awards.

The margin target that wasn't

The real change on this call was not the order book. It was the quiet surrender of the 2028 30% adjusted EBITDA margin target. Logan called it "too aggressive" and tied the decision to three things: a multi-year flat spot in medical, slower-than-expected margin expansion at Paragon, and front-loaded spending on AI and cybersecurity. The explanation deserves to be quoted at length because it reframes the equity story:

We have a long history of driving margin growth. However, we are in the midst of an opportunity. It would be shortsighted on our part to prioritize a near term margin target to the detriment of an opportunity that we believe will drive long term shareholder value.
Thomas D. Logan, Founder, Chairman, and CEO · 2026-10-28

That is a defensible strategic choice. It is also a direct reversal of what management told investors in prior quarters. In April, Logan said of the 30% target: "“We continue to stand by it.” — Tom Logan, Founder, Chairman and CEO · 2026-04-29" In July, he was still saying "“we are still holding to a 30 target... it is an audacious target.” — Thomas D. Logan, Founder, Chairman, and CEO · 2026-07-29" In February, he called it "“not a gimme putt” — Thomas Logan, Founder, Chairman and CEO · 2026-02-11." The target had become a lodestar for the bull case. Removing it is not a small guidance tweak; it changes the terminal margin math. The quarterly numbers explain why. Adjusted EBITDA margin was 23.3%, down from 23.5%, with nuclear and safety margin contracting to 25.7% from 28.1%. Medical expanded 200 basis points to 37.8%, but that was below internal expectations. Brian Schopfer was explicit: "“We are holding the low end of our prior adjusted EBITDA dollar range and lowering the high end.” — Brian Schopfer, CFO and Medical Group President · 2026-10-28" The margin target is gone, replaced by a pledge of 1-2 points of annual expansion. Logan later put a finer point on it: "“1 to 2 points of margin expansion annually” — Thomas D. Logan, Founder, Chairman, and CEO · 2026-10-28." The spending behind that reset is the AI investment and AI spend line. Mirion has spent almost $4 million of opex year-to-date on AI, expects $5 million-plus for the full year, and says it has already actioned investments above the initial budget. Logan described three buckets — data infrastructure, internal productivity, and customer-facing product development — and admitted the P&L benefit is back-end loaded. "“we are spending far more on AI and are likely to continue to do that than we anticipated 3 years ago” — Thomas D. Logan, Founder, Chairman, and CEO · 2026-10-28," he said. That is the trade: slower near-term margin for a claim on AI-enabled instrumentation and software.

What the tape is saying

The stock has already voted. MIR is down 43.7% from its October 2025 peak of $29.75 and down 10.8% over the last 90 trading days, closing at $18.78 on June 23 before drifting lower into the print. The global AI data-center trade has also cooled: the AI data center basket is still up big over 360 days but was a 90-day decliner, down roughly 5.4%. Mirion's orders are company-specific, but its multiple is not immune to the theme's wobble. The balance sheet is the offset. Q3 free cash flow was $47 million, bringing year-to-date FCF to $107 million, or 58% of adjusted EBITDA. Guidance for free cash flow was raised to $170-$185 million. Net leverage ended at 3.1x, or 2.8x excluding opportunistic buybacks, and management targets about 2.6x by year-end. Schopfer was blunt about the priority: "“that 3 number... has to have a 2 in front of it” — Brian Schopfer, CFO and Medical Group President · 2026-10-28" when the year closes. The Net leverage path supports the 40/40/20 capital allocation framework — roughly 20% capex/organic, 40% bolt-on M&A, 40% back to shareholders via buybacks — but it also limits how much stock can be retired while the margin reset is digested.

The bar into 2027

The bull case is no longer "30% EBITDA by 2028." It is "order conversion plus 1-2 points of annual margin expansion," with AI and Paragon as the swing factors. The fundamentals show a business that has already improved: gross margin was 49.9% in the latest filed quarter, up nearly 4 points year over year, and free cash flow reached $41 million in that period. But the latest reported revenue of $267 million still has to prove the Q4 step-up. What would make this a genuine re-rating rather than a dead-cat bounce: Q4 organic growth landing at the high end of mid-to-high teens, Paragon margins stabilizing as legacy contracts roll off, AI spend showing a visible ROI path at next year's investor day, and the EDF France awards converting from "awarded but unfunded" into backlog. What would confirm the bear case: another quarter of 6-7% organic growth against 60%+ orders, further Paragon charges, or another increase in AI opex without a matching revenue line. The Department of Energy pipeline and the data centers demand signal are real, but they are not yet the P&L. Mirion has the orders; now it has to earn back the margin multiple it just gave away.