Writing about aerospace and electronic systems, particularly with defense applications. Areas of interest include radar, sonar, space, satellites, unmanned plaforms, hypersonic platforms, and artificial intelligence.
Millimeter-wave radar can identify pollinators from
their wingbeat signatures, but the model has to be trained on the
species mix a particular grower actually has in the field. That problem
is fundamentally regional—and the institutional machinery built to solve
regional agricultural problems already exists. It is called Cooperative
Extension.
A Policy Brief | May 2026
Bottom Line Up Front
The new mmWave radar pollinator-identification technique reported by Antony et al. in PNAS Nexus works because a machine-learning classifier was trained on radar returns from known species.1
A grower in California's Central Valley, in Maine's wild blueberry
barrens, and in North Dakota's sunflower country needs three different
classifiers, because the local pollinator community is different in each
place. Building those regional classifiers is exactly the kind of
applied, place-based agricultural research that the U.S. land-grant
university system and state Departments of Agriculture have been doing
since the Hatch Act of 1887 and the Smith-Lever Act of 1914. The funding
rails already exist: EPA-funded Managed Pollinator Protection Plans
(MP3s) currently route money through state agriculture departments to
land-grant entomology programs in dozens of states; USDA-NIFA's Crop
Protection and Pest Management Program and Specialty Crop Research
Initiative are natural funding lines for the radar work; the National
Managed Pollinator Protection Plans Working Group already coordinates
extension specialists across institutions including MSU, Cornell, Penn
State, Oregon State, the University of Nebraska-Lincoln, and the
Institute of American Indian Arts.23
The proposal here is straightforward: state agriculture bureaus and
their land-grant partners should fund regional radar-classifier training
as a public good, distribute pre-trained models to growers as the
Cooperative Extension Service has always distributed varietal trial
results, soil maps, and IPM thresholds, and federate the resulting data
into a national reference library housed at USDA-ARS bee research labs.
Why the Training Problem Will Not Solve Itself in the Market
The Antony et al. demonstration used three European species—Apis mellifera, Bombus terrestris, and Vespula vulgaris—deliberately
chosen because they are common, behaviorally distinct, and amenable to
controlled lab data collection. Even within that small set, the
classifier required several thousand wingbeat samples per species to
converge.1 A working U.S. system needs reference signatures for, at a minimum:
The Western honeybee (Apis mellifera) and any commercially relevant managed alternatives—blue orchard mason bee (Osmia lignaria), alfalfa leafcutter (Megachile rotundata), bumblebee species being commercially propagated for greenhouse and tunnel crops.
The native bumblebees—roughly 46 Bombus species in North America, with regional dominance shifting from B. impatiens in the East to B. vosnesenskii in California to B. occidentalis in the Pacific Northwest.
The solitary bees that do disproportionate work on specific crops—Peponapis on cucurbits, Andrena on early tree fruit, Xylocopa on passion fruit and a range of legumes.
The non-bee pollinators that increasingly matter under
managed-honeybee scarcity—syrphid hoverflies, beetles, certain moths and
butterflies.
The non-target insects that produce confusable wingbeat
signatures—yellowjackets, paper wasps, bee-mimic flies, and certain
beetles.
That is, conservatively, a couple of hundred species before the
classifier is operationally useful, and the species mix that matters in
any given orchard might be only a dozen of them. Each species needs
controlled data collection with positive identification by a trained
taxonomist, which is the binding labor constraint.
The reason a private firm will not build this library on its own is
that the economics do not close. A single grower cannot justify funding
species training that benefits competitors; a single sensor company has
no incentive to release reference signatures into the public domain
because they are differentiating intellectual property; and the
reference data is jointly produced—it requires field entomologists,
taxonomic specialists, and radar engineers working together. This is the
canonical structure of an applied-research public good, and the
canonical institutional answer in American agriculture is the land-grant
university extension system.
Why the Land-Grant Model Fits This Problem Exactly
Three features of the U.S. agricultural research system make it the
right institutional vehicle for regional radar-classifier training, and
all three are direct consequences of the 19th- and early-20th-century
statutes that created it.
The Hatch Act of 1887 established State Agricultural Experiment
Stations at every land-grant institution, with a specific mandate for
applied research keyed to local crops, climates, and pests. The
Smith-Lever Act of 1914 created the Cooperative Extension Service, which
tied federal, state, and county funding together to push research
findings out to working farmers through county extension agents. The
Morrill Acts created the land-grant universities themselves; every U.S.
state has at least one (some have two or three, including 1890
institutions and tribal colleges in the system).
The pollinator-monitoring problem maps onto this structure with
almost no friction. The State Agricultural Experiment Station produces
the reference data through controlled field collection and lab radar
measurements. The Cooperative Extension specialist translates the
resulting classifier into a tool the working grower can actually use.
The county extension agent, who already visits farms in the district,
distributes and supports the deployment. None of this is
hypothetical—exactly this institutional pipeline already operates for
soil testing, varietal trials, integrated pest management thresholds,
and pesticide applicator certification.
It also already operates for pollinators specifically. EPA's Managed
Pollinator Protection Plan framework has, since 2017, routed federal
funding to state Departments of Agriculture for pollinator stewardship;
in Michigan, for example, the Department of Agriculture and Rural
Development has secured annual EPA funding flowing to MSU Extension to
implement the state's MP3, with MSU producing crop-specific stewardship
guides for blueberries and vegetables and providing pesticide applicator
training that includes pollinator-protection certification.2
USDA-NIFA's Crop Protection and Pest Management Program supports this
work through grants such as 2021-70006-35450 to MSU and the North
Central IPM Center.2
The National Managed Pollinator Protection Plans Working Group hosted
by MSU Extension coordinates extension entomologists across at least a
dozen states.3
The radar classifier training program would slot into this existing
structure as one more funded activity under the same MP3 / NIFA
umbrella, not as a new program requiring its own statutory authority.
Regional Spotlight
California's Central Valley and UC Davis
If the federated land-grant model is going to work anywhere first,
the Central Valley is the obvious starting point. California produces
approximately 80 percent of the world's almonds on roughly 1.4 million
bearing acres, and that single crop pulls in approximately 70 percent of
all U.S. commercial honeybee colonies for a six-week pollination window
each February—the largest annual pollination event on the planet. The
2024–2025 commercial colony losses hit those operations particularly
hard, and the 2026 pollination season opened with the tightest bee
supply California growers had seen in years.7
The institutional anchor is already in place. The Harry H. Laidlaw
Jr. Honey Bee Research Facility at UC Davis—described by the university
itself as "the largest and most comprehensive state-supported apiculture
facility in North America and the only one in California"—has the
entomology faculty, the field infrastructure, and the federal
partnership that a regional radar-classifier program would need.8
Dr. Elina Lastro Niño, UC Cooperative Extension's only apiculture
specialist in the state, has run the Laidlaw facility's extension
program since 2014 and founded the California Master Beekeeper Program
in 2016; her lab's stated research portfolio explicitly includes
"precision beekeeping... [investigating] the use of cutting edge
technologies to make beekeeping more efficient and sustainable."9
Drs. Neal Williams (pollination ecology, native bees), Brian Johnson
(honey bee behavior and genetics), and Rachel Vannette (chemical ecology
of pollination) round out the bee biology faculty. A USDA-ARS bee
research lab opened on the UC Davis campus, providing the federal-state
co-location that the proposed federated reference architecture needs.10
The species library that matters in the Central Valley
The classifier UC Davis would build for California growers needs to discriminate, at minimum:
Apis mellifera—the managed honeybee, the workhorse of almond pollination.
Bombus vosnesenskii—the yellow-faced bumble bee, the Central Valley's most common native bumblebee.
Bombus crotchii—Crotch's bumble bee, a CESA candidate
species since September 2022 (reinstated after litigation), which has
declined approximately 98 percent in relative abundance and now persists
in only about 20 percent of its historic range. Surveys for this
species are required under California Department of Fish and Wildlife
protocols on any project with suitable habitat in its range, which
historically included the Central Valley.11
Xylocopa varipuncta and X. tabaniformis—valley carpenter bees, ecologically important and acoustically distinctive.
Peponapis pruinosa—the squash bee, dominant pollinator of cucurbit crops.
Multiple Andrena, Osmia, and Megachile species supplementing managed pollination on stone fruit, berries, and seed crops.
Common syrphid hoverflies and the major confusable wasps (Vespula, Polistes).
The regulatory and funding hooks
California offers an unusually rich set of state-level levers. The
California Department of Food and Agriculture's Pollinator Habitat
Program, funded with $15 million by the Budget Act of 2021 (SB 170,
Skinner), already routes habitat-establishment grants through Resource
Conservation Districts and the University of California system.12
The California Department of Pesticide Regulation's pollinator-zone
restrictions took effect in January 2025, creating a regulatory
environment in which species-resolved monitoring data is directly useful
for compliance documentation. AB 1319, signed by Governor Newsom on 11
October 2025, established the "provisional candidate species" category
under CESA, expanding the universe of pollinator species that may need
monitoring data on California agricultural lands.13
What UC Davis would do with funding from a CDFA / CDPR / EPA MP3
partnership is not speculative. The Laidlaw facility is already
collecting honeybee data across most of the operationally relevant
variables; adding a 60–94 GHz radar bench, a controlled-acoustic flight
chamber, and a UC Davis College of Engineering RF/ML partnership to
produce labeled wingbeat training data is a marginal extension of
capabilities the facility already has. The output—an annually updated
California Pollinator Classifier, distributed at no cost through UC
Cooperative Extension county offices and the California Master Beekeeper
Program—would be the working model that other land-grant programs could
replicate, modified for their own regional species mix.
A Concrete Proposal: Federated Regional Classifiers
The right architecture is federated. Each land-grant entomology
program, in partnership with its state Department of Agriculture, builds
a local classifier covering the species mix that matters in its state's
principal pollinator-dependent crops. Each contributes its labeled
training data to a national reference library. The classifier becomes a
public artifact distributed at no cost to growers in the region, with
annual model updates as new species are added.
Radar data collection on identified specimens, classifier training
Land-grant electrical engineering / agricultural engineering programs in partnership with entomology
UAV imaging programs at most ag schools; precision agriculture labs
National reference library, model federation, cross-state validation
USDA-ARS regional bee labs (Beltsville MD, Logan UT, Tucson AZ, Baton Rouge LA, Stoneville MS, Davis CA)
USDA-ARS National Plant Germplasm System
Distribution to growers, deployment support, training
Cooperative Extension Service, county extension agents
NRCS Web Soil Survey delivery; IPM threshold tables
Funding
EPA MP3 grants; USDA-NIFA SCRI and CPPM; state Specialty Crop Block Grants
MSU Extension MP3 funding model (NIFA grant 2021-70006-35450)
The output the grower sees is a downloadable, regularly updated
classifier model that runs on whatever sensor platform the grower has
bought (or rented). The classifier itself—the list of species it can
recognize, and how well—is a public artifact in the same sense that USDA
Plant Hardiness Zones, the NRCS Web Soil Survey, USDA-NIFA crop variety
trials, and Cornell's eBird are public artifacts. Sensor manufacturers compete on hardware and dashboard; they do not compete on whether Bombus impatiens versus Bombus griseocollis is correctly classified, any more than they compete on whether the soil under the orchard is a Hanford sandy loam.
Why a Public-Goods Approach Beats the Commercial-Only Path
A purely commercial path produces three failures the federated public-goods architecture avoids.
First, it fragments the data. Each sensor vendor builds a proprietary
species library; libraries do not interoperate; cross-vendor data
comparison becomes impossible; researchers studying pollinator decline
cannot pool data across deployments. This is the
IoT-platform-fragmentation problem that has plagued precision
agriculture more broadly, and it slows everyone down.
Second, it under-serves the long tail of species. A vendor optimizes
its classifier for the species that drive sales—honeybees in almond
country, bumblebees in greenhouse tomato—and ignores wild and rare
species because there is no margin in identifying them. But the wild
species are precisely the ones that ecologists, regulators, and the
growers themselves need data on, both because of their ecosystem-service
contribution and because they are the indicators of habitat health.
Third, it locks growers into single-vendor stacks. A grower who buys a
sensor from Vendor A loses access to its classifier when switching to
Vendor B; switching costs are high; competition is reduced; long-run
prices are higher. This is the same dynamic that has driven the
open-data and right-to-repair conversations across the rest of
agriculture.
The public-goods classifier path fixes all three: data is federated,
the long tail of species gets covered because the marginal cost of
adding a species to a public library is paid once and amortized across
all users, and growers can take their classifier with them across
hardware vendors. None of this prevents the sensor companies from
competing vigorously on hardware and software around the classifier; it
simply takes the species library off the table as a moat.
Forward Action Items
Three things should happen this funding cycle if the institutional pipeline is to be ready when the radar hardware is fielded.
State Departments of Agriculture should add radar-classifier training as an explicit deliverable in their next MP3 cycle.
The MP3 framework already accommodates research and monitoring
activities; New York's Pollinator Protection Plan, for example, lists
"research and monitoring of pollinators to better understand, prevent
and recover from pollinator losses" as a core pillar.4 Adding a radar-classifier line item is a marginal change.
Land-grant entomology programs should partner with their
on-campus electrical and agricultural engineering departments to stand
up reference-data collection capability. The hardware—a 60–94
GHz mmWave radar sensor, a clean-room cylinder enclosure for individual
specimens, and the analysis pipeline—is well within the capability of
any university with an active radar or precision-agriculture lab. The
Trinity College team's published methodology and dataset provide a
working starting point.1
USDA-ARS should designate one of its existing bee research labs as the national reference repository.
Beltsville, Maryland is the natural choice given its central role in
the 2025 colony-loss forensics; Logan, Utah is the alternative given its
specialization in non-Apis pollinators and its existing relationship
with Western land-grant institutions.5
None of this requires new statutory authority, new federal programs,
or new funding lines. It requires reading the existing program language
broadly enough to recognize that training a radar classifier is the same
kind of activity as running a variety trial: place-based, applied,
publicly funded, publicly distributed, and indispensable to a working
agriculture sector that needs to know more about its pollinators than it
currently does.
Verified Sources
Antony, L., White, C., Marchetti, N., Donohue, I., Stout,
J. C., and Narbudowicz, A., "Harnessing mmWave signals and machine
learning for noninvasive taxonomic classification of insects," PNAS Nexus, vol. 5, no. 4, pgag096, April 2026. DOI: 10.1093/pnasnexus/pgag096. https://academic.oup.com/pnasnexus/article/5/4/pgag096/8662959
Michigan State University Extension, "Michigan Pollinator
Protection Plan Resources" (MSU implements Michigan MP3 with EPA funding
via Michigan Department of Agriculture and Rural Development; supported
by USDA-NIFA Crop Protection and Pest Management Program grant
2021-70006-35450 and the North Central IPM Center, NIFA grants
2018-70006-28883 and 2022-70006-38001). https://www.canr.msu.edu/resources/michigan-managed-pollinator-protection-plan
National Managed Pollinator Protection Plans Working
Group, hosted by Michigan State University Extension. Membership
includes extension entomologists at MSU, Cornell, Penn State, Oregon
State, University of Nebraska-Lincoln, University of Minnesota, and the
Institute of American Indian Arts Land-Grant Programs. https://www.canr.msu.edu/resources/national-mp3-working-group
"Bees Are Gold: Almond Pollination in 2026," AgNet West, February 2026 (BeeHero interview at World Ag Expo on tight bee supplies after the 2024–25 commercial colony collapse). https://agnetwest.com/almond-pollination-2026-bee-hero/
Humboldt County Beekeepers event listing describing the
Harry H. Laidlaw Jr. Honey Bee Research Facility at UC Davis as "the
largest and most comprehensive state-supported apiculture facility in
North America and the only one in California." UC Davis Department of
Entomology and Nematology, Bee Biology Program, "People." https://beebiology.ucdavis.edu/people/
E. L. Niño Bee Lab, UC Davis, "Dr. Elina L. Niño" (UCCE
Apiculture Specialist since 2014; founded California Master Beekeeper
Program in 2016; lab portfolio includes precision beekeeping research). https://elninobeelab.ucdavis.edu/people/elina-nino
California Department of Fish and Wildlife, "State and Federally Listed Endangered and Threatened Animals of California" (Bombus crotchii
reinstated as CESA candidate species 30 September 2022 following Third
District Court of Appeal reversal). California survey protocols: CDFW,
"Survey Considerations for California Endangered Species Act (CESA)
Candidate Bumble Bee Species," June 2023. https://nrm.dfg.ca.gov/FileHandler.ashx?DocumentID=109405
California Grants Portal, "Pollinator Habitat Program"
(CDFA, $15 million allocation under Budget Act of 2021, SB 170 Skinner;
eligible recipients include Resource Conservation Districts and the
University of California system). https://www.grants.ca.gov/grants/pollinator-habitat-program/
The Case for a Manned-Unmanned Heavy Strike Architecture
The right answer to "after the B-52" may not be another bomber. It may be a manned mission controller flying with a flight of large unmanned cruise-missile trucks — a heavy-class CCA family of systems that solves the surety, industrial-base, and affordability problems simultaneously, and that General Atomics is uniquely positioned to deliver on the Indo-Pacific timeline.
BLUF: The Heavy Bomber Analysis of Alternatives (AoA) the U.S. Air Force has penciled into its FY2027 budget will, on current institutional trajectory, almost certainly recommend another crewed subsonic stand-off bomber that grows in cost and shrinks in quantity over the course of a fifteen-year acquisition cycle. The analytically defensible answer is different: a heavy-class Collaborative Combat Aircraft (CCA) family of systems consisting of a manned mission-commander node — possibly a B-21 derivative, a JetZero blended-wing-body airframe, or a militarized large commercial platform — operating with a tethered flight of large unmanned cruise-missile trucks at standoff distance. The architecture preserves nuclear surety doctrine by keeping release authority with a qualified human crew, preserves the bomber career field by elevating rather than eliminating it, diversifies the heavy-aircraft industrial base by pushing truck production outside the Plant 42 single-point-of-failure, and shifts the cost-exchange ratio against A2/AD adversaries by treating long-range strike as a magazine-depth problem rather than a platform problem. The combat-relevant question coming out of Operation Epic Fury was never how many bomber sorties could be generated. It was how many cruise missiles those sorties could carry, and how quickly the magazines could be regenerated. A heavy CCA architecture answers that question. A clean-sheet crewed bomber, however well executed, does not. The deepest argument for the architecture — the one rarely spoken aloud in Air Force forums but owed to the men who paid the cost of doing it the old way — is that the technology of 2026 finally gives the United States a way to do the long-range conventional strike mission without populating another Hanoi Hilton. And the Indo-Pacific timeline argument, which dominates everything else, points to a single vendor whose institutional clock matches the strategic clock: General Atomics Aeronautical Systems, whose Avenger-derived heavy-truck variant is the only credible path to a fielded unmanned heavy-strike capability inside the 2027–2032 strategic window.
The misframed question
The trade press has covered the New Heavy Bomber AoA almost exclusively as a B-52 replacement story, which is the frame the budget submission itself invites by placing the line item inside the B-52 System Improvements Program Element. The frame is wrong. The B-52, in its B-52J configuration, is not a bomber in any operationally meaningful sense after about 2035. It is an aerial weapons truck — a subsonic launch platform whose entire combat utility, by design and by Air Force CONOPS, consists of carrying air-launched cruise missiles to a standoff release box outside the contested airspace and pickling them at coordinates supplied by off-board command-and-control.
That is not a bomber mission. It is a logistics mission with a finger on the pickle.
Once that observation is admitted into the analytical frame, the AoA's central question becomes much sharper. It is not "what platform replaces the B-52?" It is "what is the most cost-effective architecture for delivering large numbers of stand-off cruise missiles to release boxes in contested theaters, over the next sixty years, at a magazine-regeneration rate the threat dictates?" Asked that way, the answer almost certainly is not another crewed airplane.
The architecture
The proposed family of systems consists of three elements.
The manned mission controller is a crewed aircraft optimized for command-and-control rather than weapons carriage, orbiting in a relatively benign launch box at the leading edge of the threat ring. It carries the mission commander, the weapons systems officer, the nuclear surety crew, and the communications and battle-management suite necessary to coordinate strikes across a multi-platform formation. Its weapons load is modest — perhaps four to eight LRSOs internally as a hedge — because its primary contribution is human judgment and release authority, not magazine depth. Candidate airframes include a B-21 variant with reconfigured mission systems, a JetZero Z-class blended-wing-body, a Boeing or Lockheed clean-sheet command-and-control platform, or — in the most aggressive cost-cutting case — a derivative of an existing large platform such as a P-8 successor or an E-7 follow-on.
The unmanned heavy truck is a clean-sheet, long-endurance, subsonic, modestly-low-observable airframe in the size class between an MQ-9 and a 737, optimized for internal cruise-missile carriage and trans-oceanic range. Each truck carries between sixteen and twenty-four LRSOs or JASSM-ERs internally, plus growth provisions for whatever follows them. It has no pressurized crew compartment, no environmental control system sized for humans, no ejection seats, no flight deck, no crew rest. Its endurance is bounded by fuel and oil and engine MTBF rather than crew duty day, which means twenty-hour-plus sorties without aerial refueling on the ingress leg become routine. Its autonomy is deliberately limited: formation flight, commanded weapons release, autonomous recovery to a designated alternate if the manned-node link is severed. The hard autonomy problems — beyond-line-of-sight tactical decision-making, EMS-denied operations independent of the manned node — are deferred to future spirals.
The link architecture between the two is the critical enabling technology and, fortunately, the most mature. The manned-to-truck datalink is short-range, directional, line-of-sight, hardened against jamming, and operates inside a formation envelope the manned platform can see and influence directly. This is a vastly easier electromagnetic-spectrum problem than the truck-to-CONUS-control-station architecture a fully autonomous unmanned bomber would require, and it borrows directly from the CCA fighter-class work already in progress under the Air Force's Increment 1 program with General Atomics and Anduril.
A representative force structure: thirty manned nodes, two hundred unmanned trucks, distributed across Barksdale, Minot, and a forward-rotational footprint at Andersen, Diego Garcia, RAF Fairford, and the Australian bare bases. Each manned node deploys with a flight of four to six trucks, generating between eighty and one hundred forty LRSO-class shots per package. The numbers are arbitrary at this stage of the analysis; the order of magnitude is not.
Why this beats a clean-sheet bomber on every axis the AoA actually scores
Cost per delivered missile. A clean-sheet crewed bomber, at the procurement quantities the Air Force has actually historically achieved on bomber programs (the B-2's twenty-one-aircraft buy is the cautionary case, the B-21's hundred-aircraft buy the optimistic one), delivers stand-off cruise missiles at a marginal cost dominated by airframe amortization. A truck program at two-hundred-aircraft scale, built to commercial-grade tooling on a dedicated production line, delivers them at a marginal cost approaching the missile itself. The cost-exchange ratio against an A2/AD adversary improves by something like a factor of three.
Magazine regeneration. The Epic Fury campaign, in its operational pause moments between strike packages, was constrained not by aircraft availability but by missile reload. Trucks decouple the regeneration constraint from the crew duty cycle: a truck that has expended its magazine can recover to any base with a missile-handling capability and be returned to the line in hours, not days. The strike-tempo math changes substantially.
Industrial base. This is the strongest argument and the one that should resonate hardest in OSD CAPE and on the Hill. There is exactly one company building heavy bombers in the United States, on exactly one production line, in a special-access program that no second source has ever seen the inside of. A heavy-truck program of two hundred airframes is large enough to support a credible second prime — General Atomics, Kratos, Anduril at scale, or a Boeing or Lockheed clean-sheet team — on a dedicated commercial-style production line outside Plant 42. The Air Force has been saying for a decade that it wants exactly this kind of diversification. A heavy CCA truck program is the first credible vehicle for actually delivering it.
Loss tolerance. The political and operational cost of an unmanned loss is roughly two orders of magnitude lower than a downed manned crew. That permits operating profiles — closer to the threat ring, in worse weather, against denser SAM coverage — that are unacceptable for a manned platform. The second-order consequence is that the standoff-range requirement on the missile shrinks, which means the missile gets cheaper, which means magazine depth per dollar grows again. This is the kind of cascading affordability improvement the bomber community has not seen in a procurement program in fifty years.
Force-structure flexibility. A manned controller with six trucks has more mid-mission flexibility than a single manned bomber, not less. Trucks can be peeled off individually for pop-up targets. Strike packages can be scaled in real time. Combatant commanders gain the kind of graduated, scalable response options the current bomber force structurally cannot deliver without committing whole airframes to half-loaded missions.
How the institutional objections get resolved
The architecture's deepest virtue is that it accepts the institutional constraints rather than fighting them.
Nuclear surety, the showstopper for a fully autonomous truck, becomes tractable here because the human in the loop is preserved exactly where doctrine, statute, and treaty obligation require. Release authority resides with a qualified two-person crew aboard the manned node, with positive command-and-control over the unmanned shooter via a hardened directional link. The trucks are dumb pipes from a surety perspective; they cannot release weapons absent the manned node's authorization. STRATCOM, the Nuclear Weapons Surety Group, and the New START or follow-on treaty-monitoring apparatus can sign off on this architecture because it preserves the existing surety paradigm with the airframe-to-weapon physical separation extended by a tactical air gap. Compared to "fully autonomous nuclear release" — which is and will remain a non-starter — this is a small doctrinal lift. It is, in fact, the only architecture I can construct that gets unmanned heavy strike past the surety community.
The pilot-lobby problem, which would kill a fully unmanned bomber outright, reverses sign under this architecture. The manned mission commander is now responsible for a magazine of two hundred-plus missiles instead of twenty, executing target prioritization and weapon assignment decisions across a flight. That is a more demanding cognitive task than the current B-52 stand-off mission, not a less demanding one. The bomber career field keeps its rated billets, gains an argument that those billets are more critical to the kill chain rather than less, and gets to point at the manned node as the irreplaceable human-judgment element of long-range strike. AFGSC can recruit to it, train to it, and promote out of it. The career field survives — possibly in a stronger institutional position than it occupies today.
Optionality, the argument senior commanders deploy whenever they want a manned platform retained, also reverses. The flock-and-shepherd architecture is more flexible than a single manned airframe, not less, for exactly the reasons combatant commanders care about: scalable response, distributed magazine, graduated escalation control.
Datalink dependency, the argument deployed against unmanned operations in contested EMS environments, gets compartmentalized rather than solved. The hard autonomy problem — beyond-tactical-horizon operations independent of the manned node — is deferred to a future spiral. Day-one autonomy reduces to "follow the formation, execute commanded launches, autonomously navigate to a designated recovery field if the link drops," which is mature MQ-9-class capability, not a research program.
The Rapid Dragon proof
Skeptics will object that the Air Force has already chosen a different solution to the magazine-depth problem: Rapid Dragon, the palletized JASSM launch system that fires cruise missiles out the back of a C-17 or C-130 cargo aircraft. Air & Space Forces Magazine reported in May 2026 that the program is going operational in 2027.
Rapid Dragon is, in fact, the conceptual proof that the architecture argued here is correct. If a C-17 with pallets is an acceptable cruise-missile shooter — and the Air Force has now formally decided that it is — then the launch platform doesn't need to be a bomber, doesn't need a bomb bay, doesn't need a crew flying the weapon to release, and doesn't need the institutional baggage of the bomber career field. Rapid Dragon is a logistics workaround built on top of the existing mobility fleet because the bomber fleet's magazine depth is inadequate to the threat and the bomber procurement pipeline is too slow to fix it. A purpose-built heavy unmanned truck is what Rapid Dragon would have been if the Air Force had been willing to design clean-sheet for the mission instead of bolting it onto a transport.
The trajectory from Rapid Dragon to a heavy CCA truck is the same trajectory that took the Predator from a CIA reconnaissance asset to the MQ-9 Reaper. It is the trajectory the Air Force has already proven it can execute, on a class of platform it has already proven it can field. The only thing missing is the requirements document.
The argument the bomber community has earned
There is one further argument for this architecture that does not usually appear in trade-press coverage of bomber procurement, and that the AoA's analytical apparatus is structurally unequipped to make. It is the argument from history, and it is owed plainly to the men who paid the cost of doing the long-range conventional strike mission with the technology of the previous era. Sometimes the unsaid thing has to be said.
Linebacker II, December 18-29, 1972, cost the U.S. Air Force fifteen B-52s shot down over eleven nights of operations against Hanoi and Haiphong. Thirty-three crewmen were killed or listed missing in action. Thirty-three more were captured and added to the population of the Hỏa Lò Prison — the Hanoi Hilton — that already held the Rolling Thunder shootdown survivors of 1965-1968: Stockdale, McCain, Denton, Risner, Alvarez, Day, and the men who had by then been in captivity for as much as seven and a half years. The B-52D and B-52G losses came overwhelmingly to SA-2 Guideline batteries that the North Vietnamese had reloaded between strike waves because the SAM-suppression cycle was not keeping up with the bomber stream's predictable routing and timing.
The B-52 crews over Hanoi were not flying into the most heavily defended airspace on earth at that moment because doing so was intrinsically valuable. They flew into it because the conventional standoff weapons of the era did not exist. The AGM-28 Hound Dog was being retired. The AGM-69 SRAM was nuclear-only. The crews flew into the threat ring because the weapons of 1972 made them. That is the operational fact at the center of the Linebacker II loss tally, and it is the fact from which every subsequent generation of long-range strike thinking has flowed.
The same fact, in a different form, accounts for most of the rest of the Hanoi Hilton population. The F-105 Thud drivers of Rolling Thunder flew into the same SAM rings on the same predictable Route Pack VI ingress profiles out of Takhli and Korat day after day because the targeting authorities in Washington required it and the standoff weapons to do otherwise did not exist. The Navy A-4 and F-4 crews off Yankee Station ran the same flight profiles for the same reasons. Those losses, like the B-52 losses of Linebacker II, were imposed by a weapons-range constraint that no longer exists. The whole subsequent half-century arc of standoff weapons development — from the AGM-86B ALCM through SLAM-ER, JASSM, and JASSM-ER, to the LRSO now in development — was the engineering response to the lesson Hanoi taught the Air Force in 1972. The truck architecture is the next logical step in that arc, which is to recognize that once standoff range exceeds the threat ring by a comfortable margin, the human in the cockpit is providing courage and judgment to a mission that needs only the judgment.
A heavy CCA truck flight carrying LRSOs releases two hundred-plus nautical miles outside the SA-2 successor envelopes — outside the S-400 envelope, outside the HQ-9B envelope, outside whatever comes next — and never sees the threat ring at all. The crews who would have flown those B-52s into the equivalent of Linebacker II in 2030 or 2040 are, in this architecture, sitting in a manned controller orbiting safely outside the engagement zone, doing the target-prioritization and release-authority work that actually requires their judgment, and going home to base at the end of the duty cycle. The mission gets done. The crews come home. The cost-exchange ratio against the threat improves by a factor that makes the whole enterprise of long-range strike sustainable in a way it has not been since 1965.
There is a darker version of this argument worth stating plainly, because the AoA's institutional framing will obscure it otherwise. The reason the Air Force has been institutionally slow to embrace unmanned heavy strike is not really about surety doctrine or career-field preservation, important as those factors are. It is that the rated bomber community draws much of its institutional legitimacy from a lineage that includes Linebacker II, the Doolittle Raid, the Schweinfurt-Regensburg missions, and the Berlin Airlift — missions where the courage of the aircrew was inseparable from the operational outcome. An architecture that says, in effect, "we don't need to send crews into the threat ring anymore" is read inside the community as a threat to the lineage itself, even when the operational logic is unimpeachable. That reading is wrong, and the Vietnam comparison is the cleanest available way to demonstrate why.
The B-52 crews who flew Linebacker II did not fly into the SA-2 belts because riding into SAM rings was the point of being a bomber pilot. They flew because the technology of 1972 left no alternative and the targets had to be hit. The crews who flew the F-105 Wild Weasel missions out of Korat did not ride into the SAM rings because riding into SAM rings was the point of being a fighter pilot. They rode in because the SEAD problem of 1967 had not been solved and somebody had to solve it with their bodies. In both cases, the lineage being honored is the lineage of getting the mission done with whatever the technology of the moment makes possible, not the lineage of putting human bodies into harm's way as an end in itself. A heavy CCA truck architecture is, read correctly, the most respectful possible inheritance of that lineage. It says: the next generation of bomber crews will get the mission done with substantially fewer of them ending up at the equivalent of the Hanoi Hilton, because the technology of 2040 makes that possible in a way the technology of 1972 did not. The crews are still flying. They are still making the hard decisions about when and where and what to release. They are still operating in a contested theater under the same statutory and ROE constraints. They are simply not riding the missile to the launch box anymore, because the missile no longer needs them to.
The names on the Linebacker II memorial at Andersen Air Force Base, and the names of the Hanoi Hilton returnees who came home in February and March 1973 with the rest of Operation Homecoming, are the strongest available argument for the architecture, not the strongest available argument against it. The men who flew those missions were, to a man, technical professionals who would have been the first to use whatever tool the technology of their moment had given them to reduce friendly losses while accomplishing the mission. James Stockdale's writings on operational ethics, John McCain's later work on military reform, Robinson Risner's reflections on his captivity — none of them romanticize the cost. They paid it because it was the cost the technology of 1965-1972 imposed, and they paid it without flinching. The cost the technology of 2026 imposes is different, and the AoA's job, properly understood, is to make sure the Air Force pays that lower cost when it can rather than reflexively paying the higher cost out of institutional habit.
That argument should be made, in those terms, to the bomber community itself — because the bomber community, more than any other constituency in this debate, has earned the right to have it made to them honestly. The men in the Hanoi Hilton would have understood the argument immediately. The institutional successors to the men who sent them there sometimes need it spelled out. Spelling it out is not a slight to the lineage. It is the honest extension of it.
The vendor question: GA-ASI and the Indo-Pacific clock
The argument made so far has treated the heavy-truck variant as a generic procurement problem. It is not. There is a specific vendor whose institutional clock matches the strategic clock the Indo-Pacific window imposes, and identifying that vendor changes the AoA conversation from "should the Air Force eventually field unmanned heavy strike?" to "the Air Force can field unmanned heavy strike in this decade if it chooses to, from a vendor whose institutional record makes the choice defensible." Those are very different conversations. Only the second produces a fielded capability inside the strategic window that actually matters.
The vendor is General Atomics Aeronautical Systems. The reasons are specific.
Avenger is the right starting airframe. The MQ-20 Avenger is the only existing U.S. unmanned platform with most of the right DNA for a heavy-strike-truck mission. It is jet-powered, which means the cruise speed and altitude band match a heavy strike profile rather than the loiter profile of the propeller-driven Predator/Reaper line. It has an internal weapons bay, which is the structural feature that distinguishes a real combat aircraft from an externally-loaded ISR drone with bombs hung off it. It was designed from inception with modest signature management — the S-duct inlet, the internal carriage, the planform shaping — which means GA-ASI has already lived through the configuration-control discipline of building a low-observable airframe and knows where the trade-space lives. GA-ASI has been flying Avenger variants continuously since 2009, including the Predator-C demonstrator, the extended-wing Avenger ER, and the company-funded MQ-20 evolution that fed directly into the YFQ-42A CCA proposal that won Increment 1.
That last point is the one that actually matters for schedule. GA-ASI is not starting from a paper airplane. They are starting from an airframe family that has accumulated thousands of flight hours, has a mature flight control system, has been integrated with multiple weapons configurations on government and company-funded test programs, and has — through the YFQ-42A — already passed through the autonomy-stack and CCA-avionics maturation cycle that the Air Force requires for a credible combat unmanned platform. Every other U.S. prime would be starting from preliminary design. GA-ASI is starting from Block 1.
The scaling problem is tractable, not trivial. The Avenger is roughly the size of a large fighter — about 44 feet long, 66-foot wingspan in the ER variant, internal payload around 3,500 pounds, gross takeoff weight in the 18,000-pound class. That is too small for the heavy-truck mission. To carry sixteen to twenty-four LRSO-class missiles internally, the airframe needs to be in the 80,000- to 150,000-pound MTOW class with an internal bay length on the order of fifty feet — roughly the size class of a regional jet. Scaling an existing airframe by a factor of five to eight in MTOW is not a trivial engineering exercise, but it is much more tractable than the alternatives, and the reasons are specific. The Avenger family's flight control laws are already validated for jet-powered, swept-wing, internally-loaded UAS operations across a broad airspeed and altitude envelope. The autonomy stack — from the software that won the YFQ-42A down-select through the CCA Increment 1 work — is platform-agnostic at the architectural level, which is precisely the design principle the Air Force has been pushing on the CCA program from the beginning. The mission systems suite, the satcom-and-LOS link architecture, the ground control station architecture, and the deployment-and-recovery procedures are all already mature and already certified for U.S. military operations.
What scales hard is structures, propulsion, landing gear, and the integration of those subsystems into an airframe whose dynamic response is qualitatively different from the parent. GA-ASI has structures and propulsion organic capability. Landing gear and large-aircraft integration are the gaps, and the natural answer is the same answer the JetZero program took: partner with a heavy-aircraft prime — Boeing, Lockheed-Skunk-Works, or Northrop-Scaled-Composites — for the airframe-scale-up work while GA-ASI retains mission systems, autonomy, and program lead. This is the structure that has worked for the X-47B (Northrop with Navy), the JetZero BWB demonstrator (JetZero–Scaled Composites), and the original Predator (GA-ASI organic). Precedent exists.
The more aggressive path, and the one that fits the Indo-Pacific timeline: GA-ASI proposes a clean-sheet design that inherits the Avenger family's flight control architecture, autonomy stack, mission systems, and operational concept of employment, but scales the airframe to the heavy-truck mission directly rather than evolving Avenger upward in stages. This is the approach Boeing took with the 757 — same basic configuration as the 737, scaled for a different mission, sharing crew training and maintenance commonality but designed clean. It avoids the structural-margin compromises that come from scaling a fighter-sized airframe to a transport-sized mission.
The Indo-Pacific timeline is the binding constraint. Public statements from INDOPACOM, from CSBA, and from the Mitchell Institute have converged on a window of strategic concern centered on roughly 2027–2032 for a potential Taiwan contingency. The B-21 force in that window is small — perhaps twenty to forty aircraft, depending on production-rate execution — and the B-52J fleet in that window is mid-modernization, with re-engining incomplete and the AESA radar still in low-rate production. The magazine-depth problem the truck architecture solves is most acute precisely during this window, and the AoA's current schedule produces a fielded truck no earlier than the mid-2040s. That is a strategic-timing mismatch of roughly fifteen years.
A GA-ASI-led approach is the only credible path for closing that gap. GA-ASI has historically been able to execute UAS programs on commercial-style timelines — the original Predator went from contract to combat employment in roughly four years, the MQ-9 Reaper in five, the MQ-20 Avenger demonstrator in three from program initiation to first flight. Those timelines are achievable because GA-ASI operates outside the major-program acquisition apparatus on most of its work, with company-funded development carrying programs to a maturity level where DoD can buy a substantially de-risked product rather than fund the development itself. The YFQ-42A is the most recent example: General Atomics flew the airframe on company funds before the CCA Increment 1 down-select, which is part of why they won.
A heavy-truck program executed on GA-ASI's normal program tempo — clean-sheet design start in 2026, first flight in 2029-2030, IOC in 2032-2033 — is achievable. It is not achievable for any other U.S. prime starting from where the industry sits today, because no other prime has the autonomy stack, the mission systems, the unmanned-systems CONOPS experience, and the company-funded development culture all in the same organization.
GA-ASI's nuclear-surety credibility is the underrated argument. This bears directly on whether the truck variant survives Milestone B, which is the determinant of whether the program survives at all. GA-ASI's MQ-9 Reaper has, since the early 2010s, been the platform on which the Air Force has worked out the operational doctrine for armed unmanned combat aircraft under positive command-and-control. Every lesson the surety community has been institutionally unwilling to apply to nuclear-capable platforms — about link reliability, about beyond-line-of-sight command authority, about pilot-in-the-loop architectures, about the ROE-to-software translation problem — has been worked out on GA-ASI's airframes. The MQ-9 fleet has accumulated more than ten million flight hours. Its safety record is quantifiable. Its command-authority architecture is auditable. Its concept of employment is mature.
When the AoA gets to the surety question — and it will, because nothing else matters until it does — the question that determines whether the truck variant survives is whether STRATCOM and the Nuclear Weapons Surety Group can be persuaded that the manned-node-plus-unmanned-truck architecture meets the standards of positive control. That persuasion is much easier if the truck variant comes from a vendor whose airframes have been operating at the relevant standard, in the relevant CONOPS, for fifteen-plus years. It is much harder from a vendor that has to demonstrate the standard for the first time on the program of record. Northrop Grumman would face that surety-demonstration burden. Boeing and Lockheed both would. GA-ASI does not, because the demonstration record exists.
The Lynx parallel. There is a useful structural parallel from GA-ASI's own program history. The Lynx SAR/GMTI program was, at inception, an unusual proposition — a high-performance synthetic aperture radar developed inside an unmanned-systems prime rather than at one of the traditional radar houses (Raytheon, Northrop, Lockheed). The institutional skepticism about whether GA-ASI could field a competitive SAR was real in the early 2000s, and was wrong. Lynx ended up being the radar that defined the operational utility of medium-altitude long-endurance ISR for two decades, and the CCD-mode improvements developed in subsequent blocks extended that utility into mission spaces the original program had not contemplated.
The institutional skepticism about whether GA-ASI can field a heavy-class CCA truck is the same skepticism, in the same form, with the same answer. The advantage of being a vertically-integrated unmanned-systems prime is precisely that you can take a class of mission that the major airframe primes consider too small to be interesting and the major sensor primes consider too far outside their core competence, and execute it as an integrated program. That is what GA-ASI did with the Predator/Reaper/Lynx combination, and it is what they would do with a heavy-truck-plus-mission-systems-plus-autonomy-stack offering. The pattern is the company's institutional core competence.
What it would take
Three institutional moves are required, in sequence.
First, AFGSC writes the manned-node-plus-truck architecture into the AoA requirements document at inception, with the truck variant as a parallel procurement to the manned node rather than a follow-on spiral. Parallel procurement matters because it puts the truck on the same milestone schedule as the manned node, and that schedule discipline is the only mechanism that prevents the historical pattern in which unmanned spirals are repeatedly deferred at each subsequent milestone gate and quietly disappear before EMD.
Second, the AoA splits the truck procurement onto an accelerated track that uses the CCA Increment 1 acquisition pathway: company-funded prototype demonstrators, rapid down-select, fly-before-buy procurement, and fielded force-design experimentation alongside the existing B-52 and B-21 fleets. GA-ASI is invited as a full participant in the AoA, with their company-funded program-of-record proposal evaluated against the same criteria as the traditional primes. That move is somewhat unusual in heavy-aircraft AoAs — the traditional primes have historically been the only firms with the credibility to propose against bomber requirements — but the precedent exists from CCA Increment 1, where General Atomics and Anduril both displaced the traditional fighter primes in the down-select. The heavy-truck mission is, in this framing, just the next size class up the same procurement pathway.
Third, Congress writes an average-procurement-unit-cost (APUC) cap on the truck variant into the FY28 or FY29 authorization language at AoA inception. APUC caps have worked, in the modern era, exactly once on a bomber program — the original LRS-B cost cap of $550 million in 2010 dollars, which held primarily because Northrop Grumman and the Air Force agreed it was binding. Whether the same discipline could be imposed on a heavy CCA truck depends on whether anyone on the Hill is willing to spend political capital on bomber procurement again so soon after the B-21 fight. The combat-relevant lessons of Epic Fury — and the much harder lessons that a Pacific contingency would generate — argue that the political capital is available, if anyone chooses to spend it.
A representative timeline emerges from those three moves: AoA initiation in FY27, GA-ASI and one or two other vendors selected for company-funded prototype work in FY28, first flight of a representative truck airframe in FY30, fly-off in FY31, Milestone C and LRIP in FY32, IOC with a fielded truck-and-manned-node force structure by FY34. That fits inside the Indo-Pacific window. The fifteen-year nominal AoA cycle does not.
The fact that this timeline is achievable on the truck side and is not achievable on the manned-node side is itself an argument for treating the two as separate procurements with different schedule constraints. It is also, incidentally, an argument for using the existing B-52J fleet as the manned-node bridge platform in the early IOC period — a B-52J equipped with the necessary battle-management and link suite is operationally adequate for the manned-node mission while the clean-sheet manned platform works through its own Milestone B and EMD cycle. That bridge-platform role gives the B-52J fleet a coherent operational purpose during the 2030s and 2040s that justifies the investment now being made in CERP and RMP, and resolves the strategic-coherence problem the AoA's existence currently creates for the B-52J program.
In other words: the B-52J is not the platform being replaced. The B-52J is the manned node in the early-IOC architecture, controlling the GA-ASI-derived truck flock until a clean-sheet manned controller is ready to take over the role in the late 2040s. That framing inverts the entire AoA conversation. It says the B-52J modernization is not a sunk cost about to be obsoleted — it is the bridging investment that makes the heavy-CCA architecture fieldable on the strategic timeline. The eleven billion dollars in CERP-and-RMP exposure becomes, under this framing, a precisely-targeted investment in the air leg of the long-range strike force structure for the period 2030 through 2050.
The honest analytical case
The architecture sketched here is not a sure thing. The truck-side autonomy work is real, the link-architecture engineering is non-trivial, the surety doctrine work is delicate, and the inevitable creep of penetrating-strike requirements onto the manned-node program would kill affordability in exactly the way it has killed every other bomber-adjacent program of the last forty years.
But the alternative — another crewed subsonic missile truck procured at the same unit costs, in the same numbers, with the same institutional incentives that produced the B-1B cost trajectory and the B-2's twenty-one-aircraft buy — is a known failure mode. The Air Force has been there twice. There is no analytical reason to expect a third attempt to perform better than the first two.
The deepest virtue of a heavy CCA architecture is that it lets the Air Force stop pretending its problem is a B-52 replacement and start solving the problem it actually has, which is that long-range conventional strike has become a magazine-depth problem in a contested-EMS world. Reframing from platform-centric to magazine-centric long-range strike is the analytically honest output the AoA could produce, if the institutional incentives align to let it. And the historical-ethical argument for doing so is the one the bomber community itself, going back to the men in the Hỏa Lò cells, has the standing to make most powerfully — if it chooses to.
The Indo-Pacific clock is the constraint the AoA needs to be designed against. GA-ASI is the vendor whose institutional clock matches it. Those two facts, taken together, are the strongest argument for the architecture: not that it is the best long-term answer, though it probably is, but that it is the only answer that produces a fielded capability inside the strategic window that actually matters. The choice is between fielding a heavy-CCA truck capability by 2034 from a vendor with the institutional record to make the surety argument credible, or fielding nothing and hoping the B-52J magazine and the B-21 force structure are enough to handle whatever the 2027–2032 window brings.
They probably will not be. But the next twelve months of AoA framing will tell us whether the Air Force is having a real conversation about long-range strike architecture, or whether it is once again budgeting for another crewed airplane it cannot afford in the numbers the threat requires.
Verified sources
U.S. Air Force, FY2027 President's Budget, Research, Development, Test & Evaluation, Volume IV — B-52 System Improvements Program Element (Heavy Bomber Analysis of Alternatives, Advanced Concept Demonstration line), April 21, 2026. https://www.af.mil/LinkClick.aspx?fileticket=JMxt76kxvY0%3d&portalid=1
Trimble, S. "New Heavy Bomber Study Appears In U.S. Air Force Spending Plans." Aviation Week Network / Aerospace Daily & Defense Report, May 2026. https://aviationweek.com/defense/aircraft-propulsion/new-heavy-bomber-study-appears-us-air-force-spending-plans
Losey, S. "After the B-52? Air Force to Study More Heavy Bomber Options." Air & Space Forces Magazine, May 8, 2026. https://www.airandspaceforces.com/after-the-b-52-air-force-to-study-more-heavy-bomber-options/
Hadley, G. "Air Force to Field Cruise Missiles on Cargo Plane Pallets in 2027." Air & Space Forces Magazine, May 8, 2026. https://www.airandspaceforces.com/air-force-fielding-cruise-missiles-cargo-planes-pallets-2027/
Trevithick, J. "USAF Is Going To Explore What Will Finally Replace The B-52." The War Zone, May 7, 2026. https://www.twz.com/air/usaf-is-going-to-explore-what-will-finally-replace-the-b-52
Hudson, A. "JetZero Pitches Blended Wing Body Tanker as 'Game-Changer'." Air & Space Forces Magazine, June 27, 2025. https://www.airandspaceforces.com/jetzero-air-force-blended-wing-tanker/
Trevithick, J. "100 B-21 Stealth Bomber Fleet Size Target Unchanged For Now Despite Production Acceleration." The War Zone, February 24, 2026. https://www.twz.com/air/100-b-21-stealth-bomber-fleet-size-target-unchanged-for-now-despite-production-acceleration
Mitchell Institute for Aerospace Studies. Mark Gunzinger, various publications on stand-off strike, arsenal aircraft concepts, and bomber force structure. https://www.mitchellaerospacepower.org/
Congressional Research Service. "Report to Congress on U.S. Strategic Bombers." Reposted by USNI News, June 19, 2025. https://news.usni.org/2025/06/19/report-to-congress-on-u-s-strategic-bombers
U.S. Central Command. "Operation Epic Fury Fact Sheet, March 18, 2026." https://media.defense.gov/2026/Mar/18/2003900300/-1/-1/1/OPERATION-EPIC-FURY-FACT-SHEET-MARCH-18.PDF
Government Accountability Office. Weapon Systems Annual Assessment (B-52 RMP and CERP entries), GAO-25-107569, June 2025.
Air Force Research Laboratory and DARPA, public release materials on the Rapid Dragon palletized munitions program, 2021–2026.
Air Force Collaborative Combat Aircraft (CCA) Increment 1 program documentation: General Atomics YFQ-42A and Anduril YFQ-44A, public release materials, 2024–2026.
General Atomics Aeronautical Systems, Inc. MQ-20 Avenger / Predator-C public release materials and program history, 2009–2026. https://www.ga-asi.com/
Michel, Marshall L. III. The Eleven Days of Christmas: America's Last Vietnam Battle. Encounter Books, 2002. The standard operational history of Linebacker II, including B-52 loss tabulations and SA-2 employment patterns.
McCarthy, James R., Brig. Gen. USAF (ret.) and George B. Allison, Lt. Col. USAF (ret.). Linebacker II: A View from the Rock. USAF Southeast Asia Monograph Series, Volume VI, Monograph 8. Office of Air Force History, 1979.
Hobson, Chris. Vietnam Air Losses: USAF, USN, USMC Fixed-Wing Aircraft Losses in Southeast Asia 1961-1973. Midland Publishing, 2001. Comprehensive shootdown record including B-52, F-105, and F-4 losses cited.
Stockdale, James B., VADM USN. Thoughts of a Philosophical Fighter Pilot. Hoover Institution Press, 1995. Operational-ethics writing referenced.
Risner, Robinson, Brig. Gen. USAF. The Passing of the Night: My Seven Years as a Prisoner of the North Vietnamese. Random House, 1973. Captivity-period reflections referenced.
Operation Homecoming records, Defense POW/MIA Accounting Agency, regarding the February-March 1973 release of U.S. prisoners of war from North Vietnamese custody, including the Hanoi Hilton population.
Doerry, A. W., et al. Various technical publications on Lynx SAR/GMTI system architecture, coherent change detection (CCD) modes, and operational employment, Sandia National Laboratories and General Atomics Aeronautical Systems, 2002–2018.
A short editorial note for the file: the analytical case made in the GA-ASI section is the one most likely to produce a fielded capability inside the strategic window. It is also the case least likely to be made by anyone inside the AoA process, because the AoA's institutional sponsors — the bomber-prime contractors, the bomber-rated officer corps, the heavy-aircraft acquisition apparatus at AFLCMC — do not include General Atomics, do not include the unmanned-systems community, and do not include the CCA Increment 1 program office. The argument has to come from outside that apparatus, from the Hill or from the Office of the Secretary of Defense, and it has to be made before the AoA's analytical scope is locked in FY27. After that, it becomes another deferred spiral, and the strategic window closes without an answer. The window is open now. It will not be open in 2028.
A Heavy Bomber Analysis of Alternatives surfaces in the FY2027 budget request, alongside a re-engining program already 50% over its initial estimate and a radar effort that has triggered a Nunn-McCurdy notification.
BLUF: The U.S. Air Force has quietly inserted a "New Heavy Bomber Analysis of Alternatives" into its Fiscal Year 2027 Research, Development, Test & Evaluation (RDT&E) Volume IV budget submission, requesting $1 million to begin defining requirements for a possible B-52 successor — even as the service simultaneously commits roughly $11–12 billion to convert 76 B-52H Stratofortresses into the upgraded B-52J configuration. The AoA is buried within the "B-52 System Improvements" Program Element under an "Advanced Concept Demonstration" line that previously funded a $3.872 million classified proof-of-concept demonstration on a B-52 in FY2025. The Aviation Week Network was first to identify the line item. The placement is notable: AoAs typically precede a new aircraft's initial operational capability by 10 to 15 years, which would land any New Heavy Bomber fielding decision in the late 2030s — precisely when the Commercial Engine Replacement Program (CERP) is now scheduled to complete fleet re-engining and the Radar Modernization Program (RMP) AN/APQ-188 fielding will still be in low-rate production. The analytical contradictions — relevance, affordability, and industrial base — are not subtle.
What the budget actually says
The Air Force's FY2027 budget justification places the New Heavy Bomber AoA inside the same Program Element that funds the AWWP heavy pylon, LRASM/JASSM integration, and Advanced Weapon Integration. The study would follow a roughly $3 million proof-of-concept effort that concluded in fiscal 2025, which involved a "demonstration on the B-52," and Aviation Week reported that it was not immediately clear how the proof-of-concept aligned with the AoA. According to the budget narrative quoted by The Aviationist and confirmed by The War Zone, the FY2027 work scope will develop key performance parameters, key system attributes, and additional performance attributes for a follow-on heavy bomber, and address programmatic, requirements, capabilities, and vendor options. Aviation Week Network
The dollar figure is small — $1 million is, in Pentagon terms, a rounding error — but the line item is the first time since the Long-Range Strike Bomber (LRS-B) competition that produced the B-21 Raider in 2015 that the Air Force has formally written a new heavy bomber requirement into its books. As Aviation Week's Steve Trimble noted, the placement inside a B-52 program element strongly suggests the eventual target is the Stratofortress itself, not the B-21 or B-1B. The B-1B Lancer and B-2A Spirit are programmed for retirement during the 2030s as the B-21 enters service; the B-52J is the only legacy heavy bomber the service intends to retain past 2050.
A modernization in trouble
The AoA emerges against a backdrop of B-52J modernization difficulties that have eroded confidence in the original 2030 IOC plan.
The Radar Modernization Program (RMP), which replaces the mechanically scanned AN/APQ-166 with Raytheon's AN/APQ-188 — an AESA derived primarily from the AN/APG-79 used on F/A-18E/F and EA-18G aircraft, with elements from the F-15E/EX's AN/APG-82 — has slipped repeatedly. The Air & Space Forces Magazine reported in 2025 that a deviation report was filed in April, with formal notification to Congress expected imminently, and that the cost increase raised the total price tag from $2.3 billion to an estimated $2.6 billion. The Air Force characterized the breach as "non-critical" under Nunn-McCurdy (a baseline cost or schedule growth of up to 15 percent). The first AN/APQ-188-equipped B-52H ferried from Boeing San Antonio to Edwards AFB on December 8, 2025, with ground and flight testing planned through 2026. According to DOT&E, the Air Force plans a low-rate initial production decision in the fourth quarter of FY2026, followed by initial operational test and evaluation and a final production decision in fiscal 2028. To contain costs, the service is now pursuing a "minimum viable product" radar configuration that defers some originally specified capabilities for later insertion — exactly the kind of capability descope a senior radar engineer would expect when AESA aperture, antenna, and signal-processing requirements collide with a 1950s nose section and obsolete radomes. The Asia LiveAir & Space Forces Magazine
The Commercial Engine Replacement Program (CERP) has fared worse. The Air Force originally hoped to see F130-engined B-52s reach IOC in 2030, but that subsequently slipped to 2033, and it could be another three years after that before the entire fleet is re-engined — 15 years since the original contract for the new engines was awarded. The Air Force has said the program will see a "non-critical" Nunn-McCurdy breach. The 2024 House Armed Services testimony in which Air Force acquisition executive Andrew Hunter acknowledged a roughly 50% cost growth from the middle-tier acquisition prototype baseline placed total program estimates in the $9 billion range, up from prior figures around $8 billion, with cost growth driven primarily by Boeing's airframe-integration scope rather than the Rolls-Royce F130 engines themselves. Inlet redesign issues and Boeing proposal delays have been cited by GAO as principal schedule drivers. CERP passed Critical Design Review in 2025 — three years late — and Boeing received a roughly $2 billion task order in December 2025 to modify and test the first two B-52s with the new eight-engine, F130-powered configuration. The War ZoneAir & Space Forces Magazine
The aggregate exposure on the two programs alone is now in the neighborhood of $11.6 billion, before counting avionics, glass cockpit, landing gear, electrical-power, and SATCOM upgrades that round out the B-52J spec.
The B-21 baseline — and the gap it leaves
Northrop Grumman's B-21 Raider is the only U.S. heavy bomber currently in production. The U.S. Air Force and Northrop Grumman finalized an agreement to boost annual B-21 production capacity by 25 percent, with the main focus on accelerating fielding now set to begin in 2027 at Ellsworth AFB. While specific output figures remain classified, industry reporting from Aviation Week suggests production may reach up to eight aircraft per year, and Air Force officials also indicated that the final fleet size figure could remain classified. The acquisition program of record remains "at least 100" aircraft, leveraging $4.5 billion in supplemental funding from the FY2026 reconciliation bill. The War ZoneAviation A2Z
Senior commanders are pushing past that floor. INDOPACOM Commander Adm. Samuel Paparo told lawmakers in April 2026 he favors a 200-aircraft B-21 fleet; Air Force Global Strike Command has previously signaled a total bomber-force requirement around 225 aircraft. U.S. Strategic Command has reportedly advocated for 145.
The strategic logic of the AoA can be read in this gap. With B-1Bs and B-2s heading for retirement and the B-21 capped — for now — at 100, even an upgraded 76-aircraft B-52J fleet leaves the United States with a long-range strike force of roughly 176 heavy bombers covering global commitments. Mark Gunzinger, the former DASD and B-52 pilot who now writes for the Mitchell Institute, told Air & Space Forces Magazine the AoA is best understood not narrowly as a B-52 replacement question, but as a re-examination of the stand-off strike mission area as a whole. Options on the table, in his view, include a clean-sheet stand-off bomber, additional B-52 modifications, a dedicated cruise-missile carrier ("arsenal aircraft"), and additional B-21 buys.
Operation Epic Fury underscored relevance
The B-52's continuing utility was demonstrated emphatically during Operation Epic Fury, the U.S. Central Command campaign against Iran that ran from February 28 through early April 2026. U.S. forces hit 1,700 targets in Iran — including 300 new sites — while adding B-1 Lancer and B-52 Stratofortress bombers to the air attack, and the B-52s primarily used "stand-off" weapons that can be launched from a distance beyond anti-aircraft fire. CENTCOM publicly stated that B-52s struck Iranian ballistic-missile and command-and-control posts during the operation's first 100 hours, and B-52 night-strike sorties were flown from RAF Fairford. Once SEAD/DEAD operations achieved local air superiority, B-52s were flown directly over Iranian territory — the first such use of the type in that airspace. Stars and Stripes
The campaign reinforced the CONOPS now underwriting the B-52J: a high-volume stand-off missile truck operating outside contested airspace, transitioning to direct overflight only after penetrating assets (B-2 today, B-21 in future) have rolled back integrated air defenses. It is precisely this stand-off role — and the question of whether a 1962-vintage airframe is the right vehicle to perform it through 2055 — that the AoA appears designed to interrogate.
What a successor might look like
The Air Force budget submission specifies no particular configuration, and the $1 million FY27 ask buys planning, not preliminary design. Two design vectors are nonetheless worth flagging.
The first is the blended wing body (BWB). The service has invested in JetZero's Z4 BWB demonstrator (built by Northrop Grumman's Scaled Composites under a roughly $235 million cost-share contract), with first flight targeted for 2027. The Air Force has been careful to describe the BWB program as decoupled from any specific program of record, but the planform's combination of high internal volume, 30–50% range improvement over tube-and-wing designs, and modest reduction in radar cross-section — short of true low-observable treatment — is well matched to a stand-off bomber CONOPS. A common BWB platform across NGAS (Next Generation Air-refueling System), NGAL (next-generation airlift), and a New Heavy Bomber would be an obvious industrial-base efficiency, though one freighted with all the technical and program risks Northrop Grumman lived through on the B-2.
The second is additional B-21 procurement, possibly with payload-optimized derivatives. Northrop Grumman's modular open-systems architecture on the Raider is the program feature most often cited for its growth potential. A "B-21 Plus" focused on stand-off carriage rather than penetrating strike would inherit the production line, supply chain, and trained workforce now operating at Plant 42 in Palmdale, but would forfeit the BWB's payload-volume advantages.
The Air Force's documented willingness to also study "future B-52 requirements and costs" within the same AoA suggests a third path: another round of B-52 upgrades, potentially including a second-generation engine option, expanded SATCOM, and the AESA radar growth path Lt. Gen. Andrew Gebara has alluded to publicly.
The hard question: relevance versus affordability
Stephen Trimble's observation in Aviation Week — that an AoA precedes IOC by 10–15 years — is the single most important framing fact for any reader. If the AoA produces a Milestone A decision in FY2029, an EMD start in roughly FY2032, and a first flight before 2040, the New Heavy Bomber would enter service alongside a B-52J fleet whose re-engining was completed only in 2036 and whose AN/APQ-188 radars were still rolling off the production line in the early 2030s. The financial case for completing the full B-52J program then rests almost entirely on bridging a roughly 15-to-20-year gap, after which the airframe is replaced.
That math has not gone unnoticed in the comments sections of Aviation Week and the trade press, where the most common reaction has been that the United States should finish what it has started — get the B-21 to Ellsworth, finish CERP and RMP, and absorb the operational lessons of Epic Fury — before opening another bomber program. Defenders of the AoA argue, accurately, that $1 million in pre-Milestone A planning is cheap insurance against precisely the kind of capability gap the LRS-B program was launched in 2011 to close after the cancelled Next-Generation Bomber.
Both views can be correct. What is harder to argue is that a 76-aircraft fleet — re-engined, re-radared, re-cockpited, and re-pyloned at a unit cost approaching $150 million per tail — is a sustainable solution into the second half of this century. The smallest U.S. bomber type also has the smallest mission-capable margin: in 2024, the B-52 maintained a 53.77% MC rate, and the fleet is heavily tasked between Barksdale, Minot, deployed CENTCOM rotations, and the nuclear-deterrent mission. Pulling jets out of the line for CERP and RMP induction is itself a force-availability problem — one Air Force Global Strike Command leadership has flagged repeatedly. The Asia Live
The FY2027 budget hearings before the House Armed Services Subcommittee on Seapower and Projection Forces, scheduled for May 13, 2026, will be the first formal opportunity for Congress to test the analytical coherence of running an AoA for a B-52 successor in parallel with the Service's largest legacy-bomber sustainment investment in a generation.
