미리온(MIR) 2026년 3분기 실적 발표회: 수주 급증, 매출 가이던스 상향

핵심 요약

• 미리온 테크놀로지스(Mirion Technologies, NYSE: MIR)는 2026년 3분기 매출이 전년 동기 대비 24.5% 증가한 2억 7,780만 달러를 기록했다고 발표했습니다. 자체 매출(Organic revenue)은 6.5% 증가했습니다.
• M&A를 제외한 핵심 수주는 62% 증가한 3억 4,000만 달러를 기록했습니다. 파라곤(Paragon) 및 서트렉(Certrec)의 4,600만 달러를 포함한 총수주는 85% 증가한 3억 8,600만 달러를 기록했습니다.
• 수주잔고는 전년 동기의 8억 1,600만 달러에서 12억 5,000만 달러로 증가했습니다. 약 7억 3,500만 달러는 2026년 4분기 및 2027년 전체 기간에 걸쳐 매출로 인식될 예정입니다.
• 경영진은 연간 자체 매출 성장률 가이던스를 7%~8%로, 총매출 성장률 가이던스를 24%~25%로 상향 조정했습니다. 4분기 자체 매출 성장률은 10%대 중후반에 달할 것으로 예상됩니다.
• 조정 EBITDA는 6,480만 달러로 23.7% 증가했으나, 마진은 파라곤의 레거시 프로젝트 및 기타 원자력 및 안전(Nuclear & Safety) 부문의 역풍으로 인해 20bp 하락한 23.3%를 기록했습니다.
• 미리온은 AI 및 사이버 보안 투자 확대, 파라곤의 더딘 마진 개선, 의료(Medical) 부문의 성장 둔화 등을 이유로 단기 조정 EBITDA 마진 목표치인 30%를 철회했습니다. 경영진은 향후 계획에 따라 연간 1~2%포인트의 마진 확대 가능성은 여전히 존재한다고 보고 있습니다.

주요 재무 데이터

사업 및 영업 성과

원자력 및 안전 부문 매출은 34% 증가한 1억 9,400만 달러를 기록했으며, 자체 성장률은 6.3%였습니다. 기존 원전 기설용량 및 소형모듈원자로(SMR) 수요가 수주잔고에서 매출로 전환되면서 원자력 발전 매출은 자체 기준 14% 성장했습니다.

M&A를 제외한 원자력 발전 수주는 105% 증가했으며, 이전에 언급된 중국 계약 취소(de-booking) 영향을 제외하면 134% 증가했습니다. 미리온은 이번 분기 동안 1,600만 달러의 SMR 수주를 기록하여 11개 설계업체에 걸쳐 연초 대비 누적 SMR 수주액을 5,400만 달러로 늘렸습니다.

파라곤 매출은 3분기에 12%, 연초 대비 누적으로 22% 증가했습니다. 경영진은 원자력 발전 및 미국 에너지부(DOE) 관련 활동에 힘입어 연간 최소 25% 성장할 것이라는 전망을 유지했습니다. 그러나 레거시 프로젝트 조정과 개발 지출 증가가 부문 수익성에 부담을 주었습니다.

원자력 및 안전 부문의 조정 EBITDA는 23% 증가한 4,980만 달러를 기록했으나, 마진은 28.1%에서 25.7%로 축소되었습니다. 경영진은 파라곤 및 전년도 성과급 조정을 제외하면 부문 마진이 28.3%였을 것이라고 설명했습니다.

의료 부문 매출과 자체 매출은 모두 6.7% 증가한 8,400만 달러를 기록했습니다. 방사선 치료 품질 보증(RTQA)을 비롯해 3개 최종 시장이 모두 기여했습니다. 핵의학 및 선량측정 서비스는 2분기 감소 이후 자체 성장으로 전환했습니다.

의료 부문 조정 EBITDA는 12.4% 증가한 3,170만 달러를 기록했습니다. 핵의학 제품 품질 비용이 개선폭을 제한했으나, RTQA 성과와 영업 레버리지 효과로 마진은 200bp 확대된 37.8%를 나타냈습니다.

수주잔고의 가시성이 크게 강화되었습니다. 미리온은 지난해 같은 시점의 2억 1,200만 달러 대비 2026년 4분기 수주잔고 중 2억 9,200만 달러가 매출로 전환될 것으로 예상하며, 2027년 수주잔고는 전년 동기 2억 8,500만 달러 대비 4억 4,300만 달러를 기록 중입니다. 경영진은 기존 매출과 4분기 수주잔고가 연간 매출 가이던스 중간값의 95%를 충족한다고 밝혔습니다.

경영진 가이던스

경영진은 4분기 연결 자체 매출 성장률이 10%대 중후반을 기록할 것으로 예상합니다. 의료 부문은 5% 안팎의 성장률을 기록할 것으로 예상되는 반면, 파라곤의 12월 실적을 포함한 원자력 및 안전 부문은 10%대 후반 이상의 높은 성장률을 보일 것으로 기대됩니다.

4분기 자사주 매입이나 M&A가 없다고 가정할 때, 가이던스 중간값 기준 연말 순차입금 비율은 약 2.6배로 감소할 것으로 예상됩니다. 경영진은 연말 순차입금 비율이 3배 미만을 유지해야 한다고 강조했습니다.

리스크 및 주시 영역

미리온은 기존 2028년 마진 목표가 더 이상 적절하지 않다고 판단하고 있습니다. 회사는 선제적인 AI 및 사이버 보안 지출, 파라곤 일부 사업의 마진 개선 지연, 의료 부문의 2년에 걸친 성장 둔화 등을 원인으로 꼽았습니다.

첫 9개월 동안의 AI 관련 영업비용은 총 400만 달러에 육박했으며, 경영진은 2026년 전체로는 추가 설비투자를 포함해 최소 약 500만 달러에 달할 것으로 예상합니다. 2027년에는 투자가 더욱 확대될 전망이나, 미리온은 손익분기점 달성 시점에 대한 일정을 제시하지 않았습니다.

기타 마진 압박 요인으로는 인플레이션, 파라곤의 레거시 계약, 센싱(Sensing) 및 북미 지역 실적, 의료 제품 품질 비용, 전년도 성과급 기저효과 등이 있습니다. 신규 대형 계약에는 지수 연동 인상 조항이 포함되어 있지만, 원자력 및 안전 부문은 의료 부문에 비해 가격 인상분이 손익계산서에 반영되기까지 더 오랜 시간이 소요됩니다.

회사는 또한 중국과 일본에서 의료 부문의 회복 속도에 대해 신중한 태도를 유지하고 있습니다. 프랑스 신규 원전 수주 건 중 8,000만 달러 이상이 EU 자금 지원 승인을 기다리고 있어 분기 말 수주 및 수주잔고에는 포함되지 않았습니다.

애널리스트 Q&A 하이라이트

• 마진 궤적: 경영진은 사업이 상당한 영업 레버리지와 자구책 마련 기회를 유지하고 있다고 밝혔습니다. 토마스 로건(Thomas Logan) CEO는 연간 1~2%포인트의 마진 확대가 달성 가능할 수 있음을 시사했으나, 공식적인 장기 가이던스는 추후 발표될 예정입니다.
• AI 투자 수익: 지출 항목은 데이터 인프라, 임직원 생산성, 고객 대면 제품을 포함합니다. 경영진은 의료 제품 출시가 이미 가속화되고 있으나, AI 투자의 손익분기점 도달 시점에 대해서는 가이던스를 제시하지 않았습니다.
• 자본 배분: 미리온은 대략 20-40-40 프레임워크를 설명했습니다. 현금흐름의 약 20%는 자본 지출 및 자체 이니셔티브에, 40%는 볼트온 M&A에, 40%는 주로 자사주 매입을 통한 주주 환원에 사용됩니다.
• 대형 수주 기회: 회사는 4분기에 약 1억 2,500만 달러 규모의 추가 대형 수주를 목표로 하고 있습니다. 여기에는 이미 수주했으나 EU 자금 지원 승인을 기다리는 프랑스 신규 원전 프로젝트의 8,000만 달러 이상과 기타 약 4,000만 달러 규모의 수주 기회가 포함됩니다.
• SMR 동향: 경영진은 특히 자본력과 성숙한 기술을 보유한 개발사를 중심으로 SMR 분야 전반의 성장이 가속화되고 있다고 밝혔습니다. 미리온은 프리미엄 가격을 책정하는 공급업체로서의 입지를 계속 강화하고 있습니다.
• 기설용량 수요: 수명 연장 작업은 이미 미국과 유럽에서 수주잔고에 기여하고 있습니다. 경영진은 출력 증강(power uprates)과 관련된 기회는 더 긴 시차를 두고 가시화될 가능성이 높다고 언급했습니다.

실적발표 콘퍼런스 콜 전문

전체 실적 발표 컨퍼런스 콜 녹취록

경영진 발표

Operator

Good day, everyone. My name is Megan, and I will be your conference operator today. At this time, I would like to welcome you to the Mirion Third Quarter 2026 Earnings Call. [Operator Instructions] At this time, I would like to turn the call over to Eric Linn, Head of Investor Relations and Treasurer.

Eric Linn

Thank you, and good morning. Welcome to Mirion's Third Quarter 2026 Earnings Conference Call. Joining me this morning are Mirion's Founder, Chairman and CEO, Tom Logan; and Mirion's CFO and Medical Group President, Brian Schopfer.

Before we begin today's prepared remarks, allow me to remind you that comments made during this call will include forward-looking statements, and actual results may differ materially from those projected in the forward-looking statements. The factors that could cause actual results to differ are discussed in our annual reports on Form 10-K, quarterly reports on Form 10-Q and in Mirion's other SEC filings under the caption Risk Factors. Quarterly references within today's discussion are related to the third quarter ended September 30, 2026, unless otherwise noted.

The comments made during this call will also include certain financial measures that were not prepared in accordance with generally accepted accounting principles. Reconciliation of those non-GAAP financial measures to the most directly comparable GAAP financial measures can be found in the appendix of the presentation accompanying today's call. All earnings materials can be found in the Investor Relations section of our website at www.mirion.com. With that, let me now turn the call over to Tom, who will begin on Panel 3.

Thomas Logan

Eric, thank you, and good morning, everyone. Yesterday, after the market closed, we issued our third quarter results, and I'm pleased to report a quarter that reflects the continued positive momentum in our markets. Strong orders, including $140 million of large opportunity wins, revenue growth that demonstrates backlog conversion and full year guidance that has us expecting mid- to high teens organic revenue growth in Q4.

Let's start with orders, which remain the key bellwether for our business. Here, we delivered yet another strong quarter, up 62% year-over-year, excluding M&A, to $340 million. The nuclear power end market led the way with an impressive 105% increase also excluding M&A. On the revenue performance, you'll recall that in July, I told you we were gearing up for a strong second half acceleration. Third quarter organic revenue grew 6.5%, and we are tracking towards mid- to high teens organic revenue growth in the fourth quarter. Our expanding backlog is doing what is expected, converting into accelerated revenue. That momentum, together with the visibility we now have through 9 months informs our revised full year 2026 guidance, which Brian will walk through.

I would also note that we expect net leverage to improve to approximately 2.6x by year-end, which reflects both the earnings growth and the cash generation that is expected in our fourth quarter. Let me emphasize that our debt structure is very low cost with an average coupon of below 3% or less than half of what it would be if it were in the form of straight debt. Thus, our effective leverage exposure is arguably substantially lower than many people believe when looking at screens.

On the subject of guidance and before we go to Panel 4, I'd like to address our previously announced 2028 margin target, which we now believe is too aggressive. Our medical organic growth has been tempered over the last few years due to external factors, and we believe that margin expansion at Paragon will be behind initial expectations. Additionally, I'm prioritizing significant investments in AI and cybersecurity to capture the full potential from accelerating demand in our end markets.

We have a long history of driving margin growth. However, we are in the midst of a generational opportunity. It would be shortsighted on our part to prioritize near-term margin target to the detriment of an opportunity that we believe will drive long-term shareholder value. Instead, we are committed to annual margin expansion driven by operating leverage, software, procurement and self-help initiatives.

Now let's turn to Panel 4, which illustrates our Q3 order book details versus Q3 of last year. Before M&A, core orders grew 62%. Total orders, including a $46 million contribution from Paragon and Certrec grew 85% to $386 million despite the Chinese de-booking previously discussed on our Q2 call. Nuclear and Safety orders, excluding the Chinese de-booking and M&A, increased approximately $130 million, driven by a doubling of orders in the nuclear power end market. Labs and Research orders nearly doubled, helped by two large Department of Energy orders. Defense and diversified orders grew double digits and medical orders increased approximately $19 million versus last year due to strong intake in the RTQA end market. Overall, the Medical segment demonstrated order growth across all three end markets.

Panel 5 puts the broader order trajectory in context. This chart shows trailing 12-month orders, excluding M&A, and I think the picture speaks for itself. You can see the step-up beginning in the fourth quarter of last year, which further accelerates into Q3. On a trailing 12-month basis, orders have grown 37%, excluding M&A. This growth primarily reflects nuclear power end market growth. And importantly, the installed base was the largest year-over-year contributor. Recall that we have a presence in more than 98% of the world's nuclear reactors and approximately 80% of our nuclear power revenue comes from today's operating fleet. This is not growth built on unpredictable new build time lines. It is growth built on reactors that are running today.

It's also reflective of what I described last quarter, shown on Panel 6, as the waves of installed base customer demand. The catch-up capital spending after decades of profound capital rationing, the life extensions and power uprates that represent near-term capacity increases and the digital transformation that is beginning to take root. What I'd highlight today is that all 3 waves are now beginning to materialize.

The first wave is catch-up capital spending. For the August NEI report, U.S. fleet capital spending dollars per megawatt hour reached 9.7 in 2025. That's up from 7.8 in 2024 and 6.4 in 2021. The second wave is life extensions and power updates. In recent weeks, we saw operate announcements totaling approximately 1.5 gigawatts, 890 megawatts across Constellation's units in Illinois, Pennsylvania and New Jersey, 433 megawatts across Vistra's fleet in Pennsylvania and Ohio and 190 megawatts at Calvert Cliffs in Maryland. Alongside those upgrades, 96% of U.S. reactors have now been approved for 20-year life extensions and the required downstream investment work is already underway. And there may be even more runway.

In July, the NRC proposed a rule that would allow license renewals of up to 40 years, double today's 20-year limit. If finalized, operators could plan for decades of additional operation and every added year of operation means continued demand for the instrumentation, software and services that we provide. The final wave is digital transformation. We were previously awarded an order to digitally upgrade part of the reactor protection system at the first operating unit of a large installed nuclear power plant. That equipment has now shipped with installation beginning in November, and we expect to see additional opportunity with follow-on units. This is a clear example of digital transformation moving from concept to commitment across our installed base, and we are still very early in this wave.

Importantly, these 3 waves are not contingent upon future events, most specifically data center growth. Fleet operators have already made long-term commitments based upon a profound change in market dynamics, which are almost certainly likely to continue given the significant deficit in global electrical generating capacity. What will likely follow is a multiyear, possibly multi-decade capital growth cycle that will affect every category of instrumentation, software and services that we provide to nuclear power plant operators.

Let me pivot now to our backlog on Panel 7, which illustrates the positive trend in revenue recognition. Total backlog now stands at $1.25 billion, up from $816 million a year ago. That growth includes Paragon and Certrec, but even excluding M&A, we are seeing a sizable increase in expected backlog conversion in each of these subsequent years. Current year backlog recognition is projected to be $292 million versus $212 million last year. 2027 backlog is $443 million versus $285 million. Put those together, and we have approximately $735 million of backlog to be recognized in Q4 '26 and full year '27, more than $200 million greater than this time last year. This is the heart of the story, total backlog growth driving an increase in near-term revenue recognition.

I'll conclude my prepared remarks on Panel 8, which shows the improving revenue visibility that backlog is providing. At the midpoint of our revised guidance, we now have 95% revenue coverage for the full year. That compares to 94% at this point in each of the last 3 years. Here's the math. Year-to-date revenue through September is $802 million. We have $292 million of backlog expected to convert in the fourth quarter. And together, that is approximately $1.09 billion against the midpoint of our full year revenue guidance of $1.15 billion. The remaining go-get is on the order of $60 million, and most of that reflects intra-quarter medical business, which is the kind of revenue we have historically delivered and Paragon's peaks business that does not appear in backlog. The combination of year-to-date performance in the fourth quarter backlog meaningfully derisks the full year.

To summarize, we are delivering impressive orders performance, converting expanding backlog into revenue and remain on track for mid- to high teens organic revenue growth in Q4. We believe the generational tailwinds in nuclear power are only beginning to show up in our results and will set us up for strength in 2027. With that, let me turn it over to Brian to walk through the financials and the details of our revised guide. Brian?

Brian Schopfer

Thank you, Tom, and good morning, everyone. I'll continue the prepared remarks on Slide 9, outlining our financial performance. Third quarter total revenue was $278 million, an increase of 24.5% versus last year's third quarter. Organic revenue growth was 6.5%, with both segments inflecting higher from what we saw in the first half of the year. Third quarter adjusted EBITDA was $65 million or 23.7% higher than last year. Margins were 23.3% compared to 23.5% last year and below the modest year-over-year expansion we guided in July.

A contraction in Nuclear and Safety margins was driven primarily by onetime adjustments to legacy projects within Paragon and headwinds in our Sensing and North America businesses. This was more than offset by a 200 basis point expansion in Medical margins. Margins would have expanded approximately 150 basis points if you exclude a 50 basis point headwind from M&A, 50 basis points from certain onetime items and almost 70 basis points from AI-related initiatives in the quarter.

On that last point, we've spent almost $4 million in year-to-date operating expenses on AI initiatives that we expect to drive significant value in the coming years. Note, through 3 quarters, we've already actioned AI investments in excess of our initial full year budget. We opportunistically repurchased $56 million worth of shares in the third quarter, which equated to 3.6 million shares. We completed our initial $100 million share repurchase program announced in December of 2024. And in August, our Board approved a new $250 million program, of which $235 million remains. We generated $47 million of adjusted free cash flow, bringing our year-to-date total to $107 million or double the prior year. This equates to a 58% conversion of adjusted EBITDA through 9 months. Lastly, as Tom outlined, we experienced excellent Q3 order growth across both segments.

Slide 10 details the continued progress against our large opportunity pipeline. In the third quarter, we were awarded $140 million of large opportunities or $122 million net of the Chinese newbuild de-booking we discussed in July. Previously highlighted awards in July included the European installed base order and the U.S. Department of Energy order. In terms of new awards, we received a Paragon order with the U.S. Department of Energy and an Egyptian large-scale new build order. This brings our year-to-date awards to $244 million, which is $234 million higher than the same period last year.

Importantly, we have not lost any of the large opportunities in our pipeline this year. We are targeting approximately $125 million of additional awards in the fourth quarter. Included in this $125 million are two orders totaling more than $80 million that were actually awarded in September that are not currently in our backlog as the project is awaiting EU funding approval. These are related to new builds in France and are part of our strategic alliance with EDF. Before digging into the quarterly financial results, first, an update on nuclear power -- on the nuclear power end market on Slide 11.

Nuclear power orders, excluding M&A, grew 105% in the third quarter or 108% excluding FX. Excluding the Chinese de-booking we discussed last quarter, orders were actually up 134%. We booked $16 million of SMR orders in the quarter, roughly in line with the $17 million in last year's third quarter, bringing year-to-date SMR orders to $54 million, which came from 11 different SMR designers. Nuclear power organic revenue grew 14%, driven by growth in the installed base and SMRs. This is a meaningful inflection from the flat organic growth we reported in the first half as the backlog we have been building has begun converting into revenue. We remain confident that nuclear power will see double-digit organic revenue growth for the full year, which implies further acceleration in the fourth quarter.

Let's get into the quarterly financials beginning on Slide 12. Consolidated third quarter revenue grew 24.5% to $277.8 million. Approximately 18.6% of growth came from acquisitions, mainly Paragon, partially offset by a 60 basis point headwind from FX. Organic revenue growth was 6.5%, slightly below the high single-digit guide we gave in July. As a reminder, we closed the Certrec acquisition on July 31, 2025. Third quarter adjusted EBITDA was $64.8 million or 23.7% better than last year's third quarter. Adjusted EBITDA margins were 23.3% compared to 23.5% last year as margin expansion in Medical was more than offset by expected contraction in Nuclear and Safety.

Excluding Paragon, margins would have been 23.9% or up for the quarter. Adjusted EPS totaled $0.12 per share in the quarter, flat versus the $0.12 we reported last year and up from $0.11 on a comparable basis that includes stock-based compensation. As a reminder, beginning in 2026, we are including stock-based compensation in our adjusted EPS calculation. We have an adjusted EPS reconciliation slide in the appendix that has the details for your modeling.

Turning to the Nuclear and Safety segment on Slide 13. Third quarter revenue was $194 million, up 34%. Organic revenue was 6.3%, in line with our July guidance. As previously mentioned, nuclear power end market revenue grew 14% organically, driven by the installed base at SMRs. Labs and Research declined, which was partially offset by an increase in Defense and Diversified. We continue to see strong performance out of Paragon with revenue growth of 12% for the quarter and 22% year-to-date. We continue to be on track to achieve at least 25% year-over-year growth for Paragon. Adjusted EBITDA grew 23% to $49.8 million. Margins contracted to 25.7% from 28.1% last year.

As we indicated in July, this reflects dilution from Paragon, inclusive of margin pressure from those legacy contracts, mix and the lapping of a reduction in incentive compensation in 2025. Excluding Paragon and the impact of incentive compensation adjustments in the prior year, margins would have been 28.3% or up for the quarter.

On to the Medical segment on Slide 14. Third quarter revenue was $84 million, up 6.7%. Organic revenue also grew 6.7%, modestly below the high single-digit growth we had expected in July. While we did have a very robust order growth in the quarter, order timing heavily favored the back end of September. All three end markets contributed to organic growth. RTQA was the largest driver, reflecting continued strength in the OEM sector. Encouragingly, nuclear medicine and dosimetry services, both of which declined in the second quarter, grew organically in the third quarter.

Medical segment Q3 adjusted EBITDA was $31.7 million or 12.4% better than last year. Margins expanded 200 basis points to 37.8%, driven by RTQA performance and operating leverage from higher revenue growth. Although medical margins expanded, the size of the increase was below our internal expectations, partially driven by specific product onetime quality cost impact within our nuclear medicine end market.

Turning to adjusted free cash flow on Slide 15. We generated $47 million in the third quarter to end the first 9 months of 2026 with $107 million or 58% conversion of adjusted EBITDA versus $53 million and 35%, respectively, last year. Net working capital was a source of cash in the quarter, a trend that we expect to continue into Q4. We also continue to benefit from lower cash taxes, which are expected to be better than our initial guidance and from our capital structure with a blended cost of debt expected to be approximately 2.8% for the full year. We ended the quarter with net leverage of 3.1x or 2.8x, excluding our opportunistic share repurchases we did in the quarter. We expect net leverage to improve to approximately 2.6x by year-end at the midpoint, assuming no fourth quarter share repurchases or M&A.

Turning to Slide 16. We are revising our full year 2026 guidance, which reflects better visibility with only a quarter to go. We are raising our total revenue growth guidance to 24% to 25%, up from 22% to 24%, which includes acquisition-related tailwinds and a slight full year tailwind on FX, driven by the euro. We're also raising our organic revenue growth guidance to 7% to 8%, up from 5% to 7%, reflecting the inflection with the second half performance and the conversion of our growing backlog.

On our fourth quarter organic revenue growth, consolidated organic revenue growth is expected to be in the mid- to high teens. Medical is expected to grow mid-single digits plus, while Nuclear and Safety, including Paragon's December results, is expected to grow high teens plus, driven by nuclear Power. We now expect adjusted EBITDA of $285 million to $292 million and margins of approximately 24.7% to 25.2%.

We're holding the low end of our prior adjusted EBITDA dollar range and lowering the high end, reflecting our third quarter margin headwinds in Nuclear and Safety and lower-than-expected medical margin expansion. Even so, this range implies adjusted EBITDA margin expansion in the fourth quarter versus last year for the total company. We expect to see slight margin pressure in Medical due to prior year comps, mainly around incentive comp and other onetime items. And in Nuclear and Safety, we expect the best margin expansion quarter of the year.

Moving to adjusted free cash flow. We are raising our guidance to $170 million to $185 million from $155 million to $175 million, which represents 60% to 63% conversion of adjusted EBITDA, up from 54% to 58%. We are tightening our adjusted EPS range to between $0.50 to $0.54 per share versus $0.48 to $0.55 per share previously.

Lastly, we've also updated our guidance to reflect the fourth quarter euro to dollar rate of 1.12. With that, we're happy to take your questions.

Operator

[Operator Instructions] Our first question will come from Tomo Sano with JPMorgan.

질의응답

Tomohiko Sano

Maybe just on nuclear margins. I understand some of the pressure reflects the legacy Paragon mix as well as investments you're making in areas like AI. If we look through those headwinds, how should we think about the underlying margin trajectory of the nuclear business? And then as we move into 2027, should we expect those headwinds to begin to moderate and see more of the underlying operating leverage come through?

Thomas Logan

Thomas, why don't I start? And my view here continues to be very, very constructive that we see enormous potential to drive margin expansion overall. And as I indicated, there are really three broad themes that are causing us to kind of step away from this 30-point near-term EBITDA margin target. Clearly, one of them is AI and cyber, where if we look at where we are today, what our outlook is on the promise, the potential, but also on the cyber side, the risk, we are spending far more on AI and are likely to continue to do that than we anticipated three years ago when we set those targets overall.

Secondly, it is on Paragon. And on Paragon, just to flesh out the picture here a little bit. If you look at the margins on their core business, they're very strong. They are net accretive to overall Mirion margins. But where we are -- I wouldn't say upside down, but where we are gaining better insight really relates to a large-scale product development initiative that is fleshing out our positioning in instrumentation and control and nuclear power plants. And here, the spend is a little higher than anticipated. And I think the overall time line to driving margin accretion in this category is going to be a bit longer, all of which is to say we're thrilled with Paragon. We are doing very well on the integration campaign overall. But as we've gained more insights about how this component of their business is likely to evolve over time, we're just simply taking a more conservative viewpoint on it.

The final piece of the puzzle is on medical. And as noted, really due to the macro. We have seen a two-year flat spot in medical sector growth overall. And to be clear, this is a sector-wide issue. We have outperformed our competitors in this space. We are seeing a recovery. But again, in the context of where we were three years ago, we have just kind of re-rated our expectations overall. But everything that we've talked about historically in terms of the continued potential to drive improved absorption dynamics, improved procurement initiatives to harness the internal productivity of AI to drive greater product volume through AI-enabled services and all the other self-help measures continue to be there.

Again, we have a long history of margin expansion and our targets internally are very aggressive. We simply removed that 30-point near-term target because we felt it was becoming a distraction and creating a bit of overhang.

Tomohiko Sano

And maybe one on capital allocation. With free cash flow improving and leverage coming down, how are you thinking about priorities among M&A, further deleveraging and share repurchase as we move into 2027?

Brian Schopfer

Yes. I think what we've been talking about, Tomo, externally is a bit of a kind of a 20-40-40 model. About 20%-ish of our cash flow going towards CapEx and organic growth initiatives, about 40%. We've been thinking about through the M&A lens, and that's all going to be kind of bolt-on type deals. And then the last 40% deploying really back to the shareholder, right? And we don't have a dividend today. So that's pretty much entirely in the form of share repurchases. And I think we'll continue to evaluate that.

But I would note, we are exceedingly happy with the cash flow performance of the company. We expect to continue to have momentum in that space. And I think we're very well positioned for that to continue.

Operator

Your next question will come from Quinn Fredrickson with Baird.

Quinn Fredrickson

I wanted to start off just on the AI investment spending. Could you maybe discuss why now is the right time to lean further? And where would you expect your AI investments to translate into returns in the P&L? Would it be primarily better growth or margin expansion?

Thomas Logan

Yes. So Quinn, there are really many dimensions to this overall as we think about our AI spend. And I'm going to lump cyber in there because as we've seen with -- most recently with this PHP exploitation, cyber risk is absolutely exploding, really driven by the enhanced capabilities derived from Anthropic and the other frontier models overall.

As we think about AI spending right now, there are really three broad components to it. One is the data infrastructure that is really kind of creating the frictionless flow of data with the proper ontology within the enterprise overall. A second area is in internal productivity, equipping from executive suite down to shop floor, all of our colleagues with AI capabilities, training them on them and going through the process, really optimizing token spend as people come up that curve, where we are being very careful about managing that dynamic, putting people on the appropriate model for their needs and again, continuing to monitor that as people develop a more mature relationship with this technology.

The third element is on customer-facing product development. And here, we have many dozens of identified initiatives where our view is that AI can meaningfully enhance the utility of some existing solution, whether it's instrumentation, software or services that we're providing to the customer. And we've taken a very disciplined approach to creating a hierarchy of priorities and making sure that our spend is aligned with those initiatives overall.

But to be clear, here as everywhere, there is front-end loading on AI expenses, again, as we develop the infrastructure, as we train our team members and as we do the development work on customer-facing solutions that are to be launched commercially. We've not yet guided in terms of when we expect we will achieve a breakeven on this, and we'll probably provide some more insight around that as we approach our Investor Day next year. But to be clear, we're very committed to this. We want to be a leader in AI. We believe we are based upon the feedback of Anthropic and others in comparison, broadly speaking, with the industrial technology peer set.

And we think ultimately, this is going to create enormous value for the enterprise. But again, recognizing that front-end loading and recognizing the call on both OpEx and CapEx, certainly, this has been a factor in just kind of pulling back a little bit on that margin target.

Quinn Fredrickson

Okay. Second question would just be on Paragon. You discussed the margins there. Is there any change in the revenue expectations? I think you're talking about 25% pro forma growth. And just given where you're at year-to-date, should we be thinking about a pretty big 4Q for Paragon? And just if you could discuss what drives that?

Brian Schopfer

Yes. I think I said at least 25%. I think we continue to be super happy with where we sit in that cycle from that. The third quarter, Quinn, was their hardest comp, right? It's the quarter that they transacted in. And I would say the fourth quarter is set up nicely for success here with very good coverage and very good momentum in the space.

I think the other thing that's interesting under the covers maybe of Paragon is there's definitely a mix of nuclear power revenue coming through, but also increased opportunity set and revenue within the DOE space that we continue to be very excited about. So we're -- we feel good about where we sit there. There's a -- we have a great team, and we're super happy with where we're at from an execution standpoint, and we're very confident in the fourth quarter.

Operator

Your next question will come from Joe Ritchie with Goldman Sachs.

Joseph Ritchie

This is [indiscernible] on for Joe. I just wanted to spend a minute on the order growth that you're seeing and how that relates to the organic growth guide as well. So in the third quarter, organic order growth was up 62%, substantially ahead of your revenue growth in the quarter as well as your full year guide. Can you talk to the mix that you're seeing within the orders today? How much of it reflects more multiyear projects versus shorter cycle orders and just expected conversion profile over the near term as well as medium term?

Brian Schopfer

First off, just a broad comment. I think we're seeing a broad mix actually across the order base, right? We talked about a new build order in the third quarter. For what it's worth, I also mentioned some orders we actually got from EDF that are not in our backlog. So those don't run through our order rates. They're not sitting in the backlog today. We're waiting on those projects more broadly to get funded via the EU level. But those would be incremental new builds on top of what we saw even in the third quarter.

The SMR number actually wasn't the biggest dynamic in this quarter from a growth perspective. Actually, we had negative technically order growth, right? -- flat is probably a better way. So the installed base, the DOE momentum from an order perspective, all were very, very encouraging. And then as you think about it because we've been getting a lot of questions about, hey, is your backlog just growing and it's going to convert kind of way later down the line.

We put in this page, I think, Page 7 of the presentation that kind of begin that outlines just how this backlog kind of rolls off. And we segregated that page between kind of the legacy business so you can compare backwards and Paragon, so we didn't distort things. And I think what you see here is just within the current year, the legacy Mirion business, taking Paragon aside for a second, we have 18% more dollars in backlog to convert this year. Next year, that number is up 26% after the third quarter, right? And then it goes on from there. I think two years out, be 28 is up 46%, much smaller numbers, obviously.

The point is we're seeing -- we're not just seeing backlog build that converts later. We're seeing backlog build and come in that converts nearer term, both in year this year, but also even into '27. And I think that is encouraging and is something we continue to believe will be something that we continue to see kind of into the fourth quarter. So I think we're super encouraged about the order momentum and the backlog growth, but also how that will and is translating into revenue.

Joseph Ritchie

Got it. That's helpful. And just for my second question, I wanted to touch upon the Medical segment. It's not come up much yet, but some puts and takes around the guide for this year. I recognize you have an easy comp as well in the fourth quarter. So if you can talk to the implied mid-single-digit growth and what you're seeing on the ground within RTQA, that would be helpful.

Brian Schopfer

Yes. Look, I think we were very pleased with the order dynamic in medical in Q3. I think the challenge we had in Q3 specifically is it came very late in the quarter, which isn't surprising. Hospital year-end activity is actually a big September thing. And I think in the medical business, that is not an abnormal dynamic. I think that also helps set us up for Q4 a bit. I think on RTQA, it's really interesting. We continue to see very good software growth on the order side. Product, I would say, has been okay this year, better in the U.S. than in Europe. Again, we continue to see some signs of life in China, which we're continue to be cautious about. I think Japan continues to be a bit temper from a bounce back standpoint.

But I would say the team is doing a nice job in what's been a difficult medical environment. I think if you went and looked more broadly, you'd see that. But we're -- we got some great new products. We've added some leadership to that team, which we're excited about. And I think we're -- we continue to be cautious but optimistic on medical going forward. So I think that's how we're thinking about and approaching the fourth quarter.

I think we're -- it was good, we had talked about kind of the second quarter being slightly down, but that was really about China. And if you took out China in the prior year, you saw mid-single-digit kind of plus growth there. You see it again this quarter, a little bit less than our high single-digit number. I recognize that. We're kind of guiding to mid-single-digit plus. I think that's a good rhythm we're in, and we'll see how we exit the year and then how that implies to our guide on '27 when we get there.

Operator

Your next question will come from Andy Kaplowitz with Citigroup.

Andrew Kaplowitz

Tom or Brian, it's nice to see the backlog boost in Q3. I know you talked about $125 million of new large bookings for Q4. You mentioned $80 million you already booked in Q4. But how would you compare the large project award environment coming into '27 versus the large project award environment for '26? And there obviously, as you know, are some big opportunities in Europe outside of France and in the U.S. potentially in '27. So can you talk about Mirion's positioning to win them?

Thomas Logan

So from a macro picture, Andy, what we see, again, is if we begin with nuclear and then we can talk about medical a little bit, but most of this is about nuclear. On the utility scale new build opportunities, we continue to see an improving pipeline, some of which is still not yet kind of in the zone of what we would put in that high engagement pipeline. But definitely, as we look about -- as we look at the core dynamics impacting each of the major players, which would include Westinghouse, EDF Framatome, Rosatom, KEPCO and others, we see very encouraging signs there that I think will support continued growth. And we have not reflected that in the close-in large order pipeline. But to be clear, that dynamic continues to improve.

The second element would be the small modular reactor players. I know this is a sector that people are nervous about or have some degree of caution as they think about it overall. But our point of view is simpler. While we do share the view that over some period of time, maybe it's 5 years, maybe it's longer, maybe it's shorter, there will be a consolidation in this sector from the arguably over 100 discrete projects worldwide down to a handful of players that really make this market and develop it into something powerful and sustainable overall.

But our point of view is far simpler. And we take the view that there is an existential race here that the developers of the small modular technology all have the dream of being SpaceX. And they recognize that they are competing in a very dynamic field, and there's an incredible urgency to move ahead quickly to do things in a really bold fashion, and that was best demonstrated in the four new reactors that went critical in and around July 4, which was just an extraordinary achievement overall. But we look at the grouping of the people who actually have capital, the people who actually have strong balance sheets, and we take the view that they are going to spend that money because they need to. They need to do that to be first past the post overall.

And what I would say here is that we're incredibly proud of our positioning here, where if you look at the categories of instrumentation and services and software that we are selling into the sector overall and you handicap it by kind of life phase in terms of what has traded, what is yet to trade, we have developed an incredibly dominant position in this sector, again, with respect to our overall capability set. And we expect to see continued growth here. And we see it as perhaps a bit more tangible than the casual market observer may be, again, because we're taking a capital approach to it rather than a high-level macro approach.

Final piece of the puzzle relates to major system upgrade work. and the operating fleet overall. And we continue to see, again, swelling activity there triggered by the life extension decisions and the upgrade decisions that ultimately drive instrumentation package upgrades. So as it relates to the general dynamic there, we're very bullish about how the current dynamic will translate into 2027 and beyond. But it's not just nuclear power. We also have, I think, most notably, the Department of Energy where when we look at the kind of the in- queue project opportunities relating to environmental remediation, modernizing our nuclear arsenal and continuing to support the development of advanced reactor technology. Again, we see a very favorable dynamic there. So I'm approaching your question at a higher level, but I just want to make clear that as we look at the dynamics that feed that large order pipeline, which ultimately feeds backlog, the dynamic is great.

Andrew Kaplowitz

That's helpful color, Tom. Maybe just to level set us. I mean, I understand sort of the 30% margin target off the table. But if we think about what you can do in margin then over the next couple of years. I think the old dynamic was that Mirion could do 50% plus incremental's. Is that what we should level set toward? Is there any sort of puts and takes that you talk about over the next couple of years that we need to think about in terms of incremental margin targets?

Thomas Logan

Yes. I'm going to speak first and Brian is going to cringe, he can correct me after I speak. But again, I'm incredibly bullish on the margin expansion story here. Our decision to basically withdraw that 30-point near-term target is really driven by the fact that this has become a bit of a distraction and dynamics have changed, particularly as it relates to AI and cyber spending and the requisite front-end loading associated with that overall. But my view, given the powerful operating leverage of the business and again, all of these self-help opportunities and the integration pathway of Paragon, et cetera, is that as we look ahead, there's a very strong opportunity for us to drive 1 to 2 points of margin expansion annually. But I think we'll provide mature -- more mature guidance in and around that as we talk about our updated long-range plan. Brian, how do you want to correct that?

Brian Schopfer

No, I think everything you said is spot on. And we're not pulling back from expanding margins. I think we just want to -- we want to not have people worry about we're going to stretch to hit a target. And we want to make sure that we continue to operate effectively but be responsible.

Operator

Your next question will come from Nick Amicucci with Evercore.

Nicholas Amicucci

I did want to just kind of drill in a little bit on the capital allocation question that was asked prior, Brian, if we can. So just kind of thinking through 4Q and kind of the opportunity set there. I understand the 2040, 40 kind of structure. But I mean, I guess, how should we view kind of any potential acceleration of kind of share repurchases as we kind of think through just given where the stock is trading and kind of that opportunity set there?

Brian Schopfer

I don't think we want to play our hand on what we'll do there. I think the thing we do have to be mindful of in the fourth quarter is we need to get the leverage printed down for sure under three. And I think that's an absolute and has to happen in the fourth quarter. So I think that is no matter what we do, either share repurchase or on the inorganic side, like that 3 number has to have a 2 in front of it when we print at year-end, and we're very cognizant of that.

Nicholas Amicucci

Got it. And then you touched upon it a couple of times here, Brian, but I just wanted to clarify to you. Of the $125 million of incremental opportunity of the large opportunity pipeline still available in 2026, $80 million is associated with the -- with those European projects with EDF that you had referenced. So like we're talking about $45 million incremental on top of that.

Brian Schopfer

Yes, that's correct. And again, like we actually have an order for those, but we didn't put it in, we can't put it in our backlog until there's formal funding from the EU for those six reactors. Those are all new builds for what it's worth. And this was all part of the strategic deal that we cut with EDF back 2 years ago. So we're super proud of the team for getting this done. But that will come into backlog and through orders once the funding happens. But yes, there's about another $40 million-ish of other larger opportunities, which, look, that would be quite a hall for the year.

I'd also remind you, just while we're on the orders topic one more time. The fourth quarter last year was our largest ever orders quarter. FX was also -- the euro was well above $1.15, but also we won a ton of orders. So that's just a dynamic to keep in mind here on the fourth quarter, even with that potential to win $125 million or $140 million of orders.

Operator

Your next question will come from Jeff Grampp with Northland Capital Markets.

Jeffrey Grampp

On the AI spend front, Tom, I was curious to get maybe a little more clarification on. It sounded like you guys are ahead of budget, obviously, no surprise there. But can you quantify that a bit more in terms of like where does AI spending ultimately shake out this year? And should our expectations be that, that grows into '27? Or how are you guys thinking about scaling those investments?

Brian Schopfer

Yes. Maybe I'll answer the question. Look, I think I said we've almost done $4 million year-to-date, so through three quarters. That will obviously grow in the fourth quarter. My guess is it will be 5 or above total for the end of the year. I do think it will grow some next year for sure. And most of that today is sitting at corporate. So the corporate number is where a lot of that sits for what it's worth. But the benefits begin -- are sitting in the business, right? Tom talked about a lot of those.

I'll tell you, on the medical side, we're definitely seeing faster releases of products because of the spend. So the time lines are getting faster. We had some new releases at one of the most recent shows, et cetera. So there's quite a bit of good dynamic happening underneath, but that's about where we're going to land from a P&L perspective. And then on the CapEx side, you can add a couple of million dollars to that in this year. It will be less than what's through what's been expensed through. And that probably grows a bit next year as well as we continue to build out the infrastructure.

Jeffrey Grampp

Got it. Okay. Appreciate those details. And with respect to kind of the larger opportunities, obviously, you guys highlight some of the Eight life extension announcements we've recently seen. Is any of that in backlog at the end of Q3? Is any of that in large order pipeline? Or how material could some of these announcements be for Mirion?

Thomas Logan

As we look at the life extension-related activity, yes, that does inform some of the backlog, not only in the U.S. but also in Europe. As it relates to uprate activity, there's a bit more of a lag. So that is something that's yet to come in terms of how that may translate into opportunities for us.

Operator

Your next question comes from Chris Moore with CJS Securities.[Technical difficulty] Your next question will come from Brandon Knutson with Morgan Stanley.

Brandon Knutson

To start, backlog growth has been very strong. And some of the backlog here to be deliverable over multiple years. As we're seeing broad input cost inflation, what are you all seeing in terms of inflation? And then how does pricing work in the backlog? Are you priced for that going forward?

Brian Schopfer

Yes. I mean, look, we're definitely seeing inflation just like everybody else. I think our procurement programs that we put in place a couple of years ago have actually very much helped us be out in front of this. I think price cost was on a dollars basis this quarter was flattish. So on a rate basis, it hurts us a little bit. And we're -- our ability to get price in the market is something we continue to drive the teams on. It's a little easier near term on the medical side because of the book-and-bill dynamic, a little bit harder on the nuclear safety kind, not necessarily to get price, but for it to come through the P&L.

On the larger contracts, everything new that we've signed has escalation clauses in it that are tied to some appropriate index depending on where that contract is. So our ability to kind of keep up with inflation from a large contract standpoint, we feel like we have a very good handle on. And I think as we continue to see kind of the supply-demand dynamics working, that probably works in our favor from a pricing perspective. But price is a hard one to talk about publicly just because we don't want to overplay our hand too much. But we're definitely seeing cost inputs come up. I think our margin profile also helps us, right, where even on the -- where we're seeing it, our ability to pass it on is been good. So that's how we're kind of thinking about it at this point.

Brandon Knutson

Great. And then a quick one on SMRs. I think you said that the orders this quarter represent 11 different SMR providers, which shows good relationships across the many providers that are out there. I'd also imagine this gives you good visibility into how the projects are moving along. So have you seen any slowdown in terms of maybe any just a slowdown in terms of them being able to deliver on time or any delays there?

Thomas Logan

No, Brendan, I think we're seeing the opposite. I think we're seeing an acceleration, a strengthening of the sector overall. It's not universally true, but it is broadly true. And I think it's a robust dynamic that time lines continue to move to the left. And again, it's all about who has capital and who has a mature technological readiness level. But as a sector overall, we're seeing things move a bit faster and a bit more tangibly.

Brian Schopfer

There's a real competition in that space for who wins, right, or the couple that wins. So you're seeing that definitely play out for sure.

Operator

Your next question will come from James West with Melius Research.

James West

Tom or Brian, you guys a couple of years ago gave a dollar per megawatt assumption for new builds. And I think at the time, you were trying to describe how much larger it was with SMRs versus 1000. But have you updated that number since the Paragon transaction? And I'm assuming it's considerably higher, if you have a magnitude or an official estimate, I'd love to hear it.

Thomas Logan

We have not, James. We'll do that again prior to the investor event next year. But directionally, you're correct. I mean it's a combination both of enhanced organic capabilities as well as the inorganic principally through Paragon and Certrec that augment our wallet share. But also other factors that will go into this update are the fact that when we gave that guidance a few years ago, we are still using a default operating lifespan of 60 years for a nuclear reactor, whereas now, I think arguably, that number would be 100. And it will drive a bit of a different dynamic. Our mix will be different. Our footprint will be broader.

And so that number will go up for utility scale. And as it relates to the SMR dynamic, we obviously have gained an enormous amount of knowledge about the different players, the different technologies. And ultimately, where that center of gravity is evolving toward. And so it's premature really to toss out a number there, but I would just continue to hold to the directional guidance we've given before that on a per megawatt basis, the opportunity set is considerably higher than it is for utility scale.

James West

Okay. Got it. That's fair enough. And then maybe just a follow-up on the SMRs and the competitive environment there, and Brian alluded to it, I think, in the answer to one of the last questions. The urgency effect from these SMR companies, how is that having an impact on your ability to price into these projects?

Thomas Logan

I think it's the old saying good, fast, cheap, choose to. And we are really good. We can be really fast, but we are a premium priced player in the marketplace. And obviously, if you look at our track record of performance to date in the SMR sector, again, we're very proud of the team for what they've done and really forged just amazing partnerships with the majority of these players where they're trying to do great things. And we want to walk hand-in-hand with them as they do these things. Again, I'd cite the July 4 criticality events, which just, I mean, amazing and how quickly that came together and what it took to make it happen. And we're proud to be a part of that. So we are really focused on being a great partner to the sector overall. And with that, I think we'll be fairly paid for it. But the dynamic so far is quite good.

Operator

This concludes the question-and-answer session. I'd now like to turn the call over to Tom Logan, Chairman and CEO of Mirion.

Thomas Logan

Well, folks, we really appreciate the time and attention today again. Very proud of the performance of the enterprise and a big thanks to the -- all of my Mirion colleagues worldwide for delivering a strong quarter, but I think more importantly, setting us up for success in Q4 and into 2027. So thank you for your time again. We'll look forward to our next call a few months down the road.