BWX Technologies research
BWX Technologies: The Company That Has to Win, No Matter Which Reactor Wins
- 1. The Megatrend: AI's Electricity Problem
- 2. Why Nuclear, Specifically
- 3. What BWX Technologies Actually Is
- 4. Q2 2026, By the Numbers
- 5. The Moat: Naval Nuclear Propulsion
- 6. Project Pele and BANR: Betting On the Category, Not One Reactor
- 7. The DOE Reactor Pilot Program: Supplying Fuel To (Almost) Everyone
- 8. Ontario, Wyoming, and the Rest of the Commercial Build-Out
- 9. Selling the Medical Business: A Capital Discipline Story
- 10. What You're Actually Paying
- 11. What Could Go Wrong
- 12. Muffett's Take
1 The Megatrend: AI's Electricity Problem
Every conversation about the AI buildout eventually runs into the same wall, and it isn't chips. It's power. The IEA's base case has global data center electricity consumption climbing from roughly 415 terawatt-hours in 2024 to about 945 terawatt-hours by 2030 — more than doubling in six years, with AI accelerator workloads growing at something like 30% a year driving most of that increase. That demand has to come from somewhere, and the grid, as currently built, cannot absorb it fast enough. US interconnection queues now hold more than 2,600 gigawatts of proposed projects waiting to connect, average wait times are stretching toward five years, and withdrawal rates on stalled projects have run as high as 80% in some regions. Wholesale power prices near major hyperscale data center clusters have reportedly surged by as much as 267% in places where demand has outrun supply. This is not a distant, theoretical problem. It's already showing up in electricity bills and project timelines today.
Every major hyperscaler has responded the same way, and it's the reason this megatrend is worth building an entire research thesis around: Microsoft, Google, Amazon, and Meta have between them signed more than thirteen separate nuclear power agreements, together committing to something close to 9.8 gigawatts of nuclear capacity specifically earmarked for AI infrastructure, as of counts taken in mid-2026. Microsoft alone signed a $16 billion, twenty-year deal to restart the former Three Mile Island Unit 1 reactor — now rebranded the Crane Clean Energy Center — targeting commercial operation in the second half of 2027. Google struck an agreement in January 2026 with Elementl Power for up to 1,800 megawatts of advanced nuclear capacity. These aren't pilot projects or public-relations gestures. They're the largest technology companies on earth writing multi-billion-dollar, multi-decade contracts for a specific kind of electricity, because they've concluded nothing else solves their actual problem.
2 Why Nuclear, Specifically
The reason nuclear keeps winning this argument, rather than wind or solar, comes down to one unglamorous statistic: capacity factor, meaning how often a power source is actually producing electricity at rated output. Nuclear plants run above a 92% capacity factor. Wind runs around 34.3%. Solar runs around 23.4%. Put in plain terms, nuclear delivers roughly 2.7 times the reliable output of wind and close to 3.9 times that of solar, for the same nameplate capacity. An AI data center running training runs and inference workloads around the clock doesn't care how sunny or windy it is outside — it needs power at three in the morning just as much as at three in the afternoon. Nuclear is, today, the only zero-carbon technology that can promise that kind of always-on reliability at meaningful scale, which is exactly why it's the technology hyperscalers keep signing contracts for rather than simply building more solar farms.
This is the structural setup we look for across our technology and infrastructure coverage: a real, urgent, well-funded demand curve, running directly into a supply side that is genuinely hard to expand quickly. Building a new large nuclear plant from scratch takes the better part of a decade and tens of billions of dollars. What's actually moving faster is a mix of restarting mothballed reactors, extending the life of existing plants, and — the part we think is most underappreciated — a new generation of small modular reactors and microreactors specifically designed to be built faster, cheaper, and closer to the point of demand. That's the part of the nuclear renaissance where BWX Technologies sits.
3 What BWX Technologies Actually Is
BWXT isn't a utility, and it isn't betting its future on any single reactor design becoming the industry standard. It's the industrial manufacturer that sits underneath the entire US nuclear enterprise, civilian and military alike — the company that actually forges the reactor pressure vessels, fabricates the nuclear fuel, and manufactures the precision components that make a nuclear reactor an actual, licensable, operating machine rather than a blueprint. For more than half a century, BWXT has been the sole source manufacturer of reactors for the US Navy's entire nuclear-powered submarine and aircraft carrier fleet — an extraordinary, essentially uncontested position that most public companies never get to hold in any market, in any country. On top of that decades-old government relationship, BWXT has spent the past several years building out a genuinely diversified commercial nuclear business, supplying components, fuel, and engineering services to the new wave of SMR and microreactor developers racing to meet exactly the demand curve described above.
4 Q2 2026, By the Numbers
The most recent quarter is a clean illustration of a business firing on both its old and new engines at once. Revenue came in at $901.6 million, adjusted earnings per share of $1.07 beat the Street's $1.05 estimate, and GAAP net income of $89.1 million was up 14% year over year. Backlog reached $8.4 billion, up 40% year over year, with a trailing-twelve-month book-to-bill ratio of 1.7 times — meaning BWXT is booking new orders considerably faster than it's working through existing ones, which is exactly the leading indicator you want to see in a business with a multi-year growth story still in its early stages. Free cash flow came in at $115 million for the quarter, and management raised full-year 2026 free cash flow guidance to a range of $345 million to $360 million. Full-year revenue guidance was raised to approximately $3.8 billion, adjusted EBITDA guidance to $662 million to $672 million, and non-GAAP EPS guidance to $4.70 to $4.80. Raising guidance across revenue, margin, and cash flow simultaneously, in the same quarter, is not something a business does when its underlying demand is anything less than genuinely strong.
Underneath those headline numbers, the two segments told different stories worth understanding separately. Commercial Operations — the newer, SMR-and-commercial-nuclear-facing side of the business — grew revenue 72% year over year, including 33% organic growth, with segment adjusted EBITDA more than doubling to $36 million, up 123%, at an 11.9% margin. Government Operations — the legacy Navy propulsion business — grew a more modest 2%, as growth in special materials offset a temporary dip in microreactor program volumes, while segment margin actually expanded to a strong 20.9% on operational execution. Read together: the old, stable moat is still throwing off excellent margins even in a quieter quarter, while the new growth engine is accelerating hard. That's a healthier combination than either number would suggest in isolation.
5 The Moat: Naval Nuclear Propulsion
It's worth dwelling on just how unusual BWXT's core government business actually is, because "sole source supplier to the US Navy's nuclear fleet" undersells it. Every nuclear-powered submarine and aircraft carrier the United States operates runs on a reactor whose core components trace back to BWXT manufacturing, and there is, for all practical purposes, no second qualified domestic supplier for large portions of this work. The Navy's own thirty-year shipbuilding plan, updated in May 2026, calls for sustained production of two Virginia-class submarines and one Columbia-class submarine every single year for decades to come, alongside continued Ford-class aircraft carrier construction. BWXT recently announced $2.6 billion in new contracts specifically for naval nuclear reactor components tied to these programs, with deliveries scheduled across the next six to eight years, and separately won a $1.5 billion defense fuels contract from the National Nuclear Security Administration in September 2025. This is about as close to a government-guaranteed, multi-decade revenue stream as exists anywhere in the public markets — the kind of durable, high-visibility backlog we like to anchor a position around before layering on a faster-growing, higher-optionality story on top.
6 Project Pele and BANR: Betting On the Category, Not One Reactor
BWXT's most direct exposure to the "power for remote and distributed applications" story runs through two related programs. Project Pele is a Department of Defense initiative to build the country's first transportable microreactor — BWXT holds the roughly $300 million cost-type contract to build the prototype, producing 1 to 5 megawatts of electricity, with fuel and reactor components scheduled to ship to Idaho National Laboratory in 2026 and first power potentially as early as 2028. BANR, BWXT's own civil-focused advanced reactor design, leverages that same underlying technology in a 75-megawatt-thermal high-temperature gas reactor aimed at remote power and industrial heat applications — the company has a two-phase, $20.5 million partnership with the Wyoming Energy Authority specifically targeting power-hungry industrial users like trona mining operations. Neither program alone is a needle-mover for a company BWXT's size today. Together, they represent genuine, government-and-industry-validated proof that BWXT's core manufacturing and fuel expertise translates directly into the next generation of smaller, faster-to-deploy reactors — exactly the category most likely to serve data centers that need new power in years, not decades.
7 The DOE Reactor Pilot Program: Supplying Fuel To (Almost) Everyone
This is, in our view, the single most underappreciated piece of the entire BWXT story, and it's the clearest expression of the "picks and shovels" thesis we keep coming back to in this research library. In May 2025, an executive order directed the Department of Energy to stand up a Reactor Pilot Program with a goal of getting at least three test reactors to criticality by July 4, 2026, using a streamlined DOE authorization process rather than the traditional, much slower NRC licensing path. The DOE selected eleven advanced reactor projects to participate — companies including Oklo, Radiant, Terrestrial Energy, Last Energy, Natura Resources, Antares Nuclear, and others, each pursuing a different reactor design and technology bet.
BWXT isn't betting on any one of those eleven companies winning the race. Instead, it's supplying the specialized TRISO fuel — a uniquely robust, accident-tolerant fuel form — that several of these designs actually require to operate. In practice, that already means results: BWXT's TRISO fuel powered Antares Nuclear's reactor to its first successful criticality milestone under this program, with BWXT also having processed the high-assay low-enriched uranium feedstock, sourced from NNSA scrap materials, used to manufacture that fuel. This is exactly the kind of position we look for in a fast-moving, technologically uncertain race: instead of trying to guess which specific reactor design becomes the industry standard, own the specialized manufacturing capability that most of the credible contenders need regardless of who ultimately wins.
8 Ontario, Wyoming, and the Rest of the Commercial Build-Out
BWXT's commercial nuclear footprint extends well beyond the US test-reactor programs. In Canada, BWXT signed contracts worth roughly CAD $1 billion — about US $690 million — to support Ontario Power Generation's life extension of the Pickering Nuclear Generating Station and the new small modular reactor build at Darlington, where OPG installed the first basemat module in 2026. As part of that broader relationship, BWXT is also manufacturing the reactor pressure vessel for GE Hitachi's BWRX-300 SMR design — again, supplying critical hardware into a specific reactor program without needing to own or bet the company on that program's ultimate commercial success. Separately, BWXT holds an engineering contract with TerraPower to support its Natrium reactor project under construction in Kemmerer, Wyoming. Add these together with the DOE pilot program fuel supply relationships and the Navy propulsion backlog, and the picture that emerges is a company with genuine exposure to essentially every major nuclear technology path currently being pursued in North America — sodium-cooled fast reactors, high-temperature gas reactors, light-water SMRs, and traditional large-scale naval reactors alike.
9 Selling the Medical Business: A Capital Discipline Story
On August 3, 2026 — alongside the Q2 earnings release — BWXT announced it is selling its medical isotopes business, BWXT Medical together with Kinectrics' stable isotopes operations, to Nordic Capital in a transaction valued at up to $800 million, expected to close by the end of 2026. BWXT Medical was a genuinely interesting business in its own right, including a growing position in actinium-225, an alpha-emitting isotope used in next-generation targeted cancer therapies, with supply agreements already in place with NorthStar Medical Radioisotopes and Nucleus RadioPharma. But it was also a business built around an entirely different customer base, sales motion, and regulatory pathway than BWXT's core nuclear manufacturing franchise — pharmaceutical companies and hospitals, rather than the Navy, the DOE, and utility-scale reactor developers.
Management's own stated rationale is straightforward: sell a good business for a full price, in an area outside the company's core focus, and redirect that capital and management attention toward the national security and commercial nuclear power markets described throughout this note — precisely the areas benefiting most directly from the AI-driven electricity demand story. BWXT is retaining a meaningful minority stake in the divested business rather than exiting the space entirely, which reads as a sensible way to keep some exposure to a market management still believes in, without needing to run it day to day. This is exactly the kind of disciplined, non-sentimental capital allocation decision we like to see from a management team — recognizing that being good at something isn't the same as it being the best use of the next dollar of investment.
10 What You're Actually Paying
BWXT trades today at $172.56, with a market capitalization of roughly $16.45 billion, after a 52-week total return of about 31.7% — a genuinely strong run, though notably not the kind of parabolic, speculative move that's characterized some of the smaller pure-play SMR names over the same period. Wall Street's own recent moves on this name have leaned constructive: JPMorgan initiated coverage with an Overweight rating and a $230 price target, while Deutsche Bank, despite trimming its target slightly to $253 from $255, maintained a Buy rating — both targets implying meaningful upside from today's price. We'd frame BWXT's valuation less as a bet on any single catalyst and more as a bet on duration: an $8.4 billion backlog, up 40% year over year, with a 1.7 times book-to-bill ratio, gives a level of multi-year revenue visibility that few growth stories in this market can actually document with hard contracted numbers rather than narrative.
11 What Could Go Wrong
- Government program timing risk: a meaningful share of BWXT's growth narrative depends on DOE, DOD, and Navy program schedules holding — Project Pele's path to 2028 first power, or the Reactor Pilot Program's original July 2026 criticality goal, are the kind of ambitious government timelines that have a documented history of slipping.
- Government Operations growth is currently soft: 2% segment revenue growth in Q2, even with strong margins, is not the pace investors underwriting a nuclear-renaissance growth story are ultimately looking for from BWXT's largest segment — a sustained slowdown here would matter even if Commercial Operations keeps accelerating.
- The medical business sale isn't closed yet: the transaction with Nordic Capital is subject to customary regulatory approvals and isn't expected to close until the end of 2026 — regulatory delay or a deal collapse would leave BWXT holding a non-core asset longer than planned and complicate the capital-reallocation story.
- Commercial nuclear demand could still disappoint relative to hype: hyperscaler nuclear commitments are real and substantial, but independent analysis, including from groups like the Carnegie Endowment, has raised legitimate questions about whether announced gigawatt commitments will convert into actual operating capacity on the timelines currently being promised.
- Competitive and supply chain risk in commercial nuclear: BWXT is well positioned across multiple reactor technology paths today, but a wave of new entrants and government-backed competitors in fuel fabrication and component manufacturing could pressure margins in the commercial segment over time, even as the Navy relationship remains protected.
- Valuation has already re-rated meaningfully: a 52-week return above 30% means some of this thesis is already reflected in the share price. A disappointing quarter, particularly in Commercial Operations growth, could trigger a sharper pullback than the underlying long-term story would otherwise justify.
12 Muffett's Take
We think the AI-power-demand story is one of the more genuinely durable megatrends available to investors right now, for the simple reason that it isn't really a bet on a forecast — it's a bet on physics and contracts that have already been signed. Data centers need continuous power. The grid can't deliver enough of it fast enough. Nuclear is the only proven, at-scale technology that solves that specific problem, and the largest, best-capitalized technology companies in the world have already put real money behind that conclusion, to the tune of roughly 9.8 gigawatts of committed nuclear capacity and counting.
BWX Technologies is, in our view, one of the cleanest ways to own that megatrend without needing to guess which individual reactor company wins. A decades-long, essentially uncontested position as the US Navy's nuclear reactor manufacturer anchors the business with genuinely government-guaranteed cash flow. Layered on top, a commercial nuclear segment growing revenue 72% year over year, with real, signed contracts spanning DOE test reactors, Canadian SMR builds, and next-generation reactor developers across nearly every competing technology path. And a management team that just showed real capital discipline by selling a good but non-core medical isotopes business to double down on exactly this opportunity. This isn't a speculative, story-driven stock. It's a well-run industrial business with a genuinely exceptional backlog, sitting directly underneath one of the most credible growth narratives in the market today.
