Mass Without Depth Is Temporary
On 1 October, at the National Gallery of Australia in Canberra, the Sir Richard Williams Foundation held its second seminar of 2026, The Requirement for Air Combat Mass and Depth in the Indo-Pacific. One session carried the question that sits under most air force planning at the moment. As air forces rebuild combat power and add collaborative combat aircraft (CCAs) and other autonomous systems, where are the gaps, the risks and the opportunities?
Air Marshal Geoff Brown AO (Retd), Deputy Chair of the Foundation, explored that question with Lt Gen David Deptula, USAF (Ret.), Dean of the Mitchell Institute for Aerospace Studies. Deptula’s central message was that air forces must be designed to win, with enough capability and capacity to achieve national objectives and keep fighting.
Brown was more than a moderator. As a former Chief of Air Force who served as the operational commander of Australian air assets in Operation Iraqi Freedom, he pressed Deptula from the operator’s seat, testing the American case against the Australian experience. Several of his interventions sharpened the argument, and they appear in the sections that follow.
Mass is an effect, not a count
Deptula starts by refusing to equate mass with numbers. “Tail count is not combat power,” he said. Land power creates mass by physically massing forces. Airpower creates its equivalent by concentrating effects in time and space, through speed, reach, precision, coordination and sequencing.
On this logic, a small force of highly capable aircraft can produce enormous operational effects if it can penetrate, find the right targets, communicate, employ weapons and return to do it all again. A much larger force that cannot reach the battlespace, or cannot survive in it, produces very little. His Forbes article puts it starkly: a thousand aircraft orbiting outside an adversary’s engagement envelope do not provide the combat power of a hundred able to penetrate it, execute their missions and fight again.
This leads to his working definition. He would measure combat mass by asking how many relevant effects a force can generate, at what distance, against what level of opposition, and for how long. The final clause matters most, because it introduces the second half of the seminar’s title.
Depth is what makes mass last
“Mass without depth is temporary,” Deptula wrote. Depth means enough aircraft, weapons, crews, maintenance capacity, fuel, basing, command and control, and logistics to sustain effects over time. A force that generates enormous effects for two days and then runs out of weapons or serviceable aircraft lacks depth, however impressive its opening performance.
This is the half of the equation that procurement announcements tend to underplay. Aircraft are visible and countable, and they photograph well on a ramp. Weapons stockpiles, spare engines, trained maintainers, hardened shelters and fuel infrastructure do not. Yet they determine whether a force is a campaign instrument or a one-week demonstration.
For Australia, with its long distances and thin infrastructure, the point is acute. Depth is the quiet determinant of whether a capability can be sustained at the tempo and duration a crisis demands.
The problem with “affordable failure”
The phrase in Deptula’s title is deliberately sharp. His target is the habit of treating low unit price as a virtue in itself, and his objection is that cheap is not the same as cost-effective.
Brown returned to the phrase in the discussion: there is no virtue, he said, in affordable failure. He reached for a Second World War parallel in the air commander George Kenney’s comparison of air forces to a poker hand, in which the second-best hand costs a great deal of money and gets you nothing in the end. The analogy cuts against the budgeting instinct. A force that is cheap but not good enough does not save money. It spends it on a result that does not count.
Deptula offered an illustration. A $10 million aircraft that must be sent in groups of twenty to achieve what one $100 million aircraft can achieve, and loses half the formation doing so, is not cheaper. Nor is an aircraft at one-third the price more useful if inadequate range, payload, survivability or weapons capacity forces repeated sorties at greater cost and risk.
His preferred metric is cost per desired effect, and the historical case he reaches for is stealth. In Desert Storm, F-117s flew less than 2 per cent of combat sorties yet struck more than 40 per cent of the fixed targets. The first non-stealth attack package of the campaign comprised 41 aircraft, only eight of which dropped bombs, striking three aimpoints on a single target. At the same time, twenty F-117s struck 38 aimpoints on 28 targets without additional force-protection aircraft. By that comparison, one F-117 was the equivalent of 26 non-stealth aircraft. The F-117 cost more per airframe and far less per effect.
Brown asked whether the airframe is even the right thing to cost, and Deptula’s answer was that the public and most legislators, in Washington as in Canberra, think about defence the way they think about shopping. They compare unit prices and overlook that the cheaper option may need twenty or thirty units to do what the dearer one does. Warfighting, he said, is not buying two cans of beans. He conceded that a rigorous cost-per-effect analysis is much harder than comparing sticker prices, but argued that this is where the discussion has to go.
The denominator is the crux. Compare the price of a penetrating aircraft with the price of a conventional one and stealth looks extravagant. Count the whole package required to produce the effect and the conclusion reverses. That package includes the aircraft, weapons, people, fuel, bases, maintenance, supporting assets and expected losses. Deptula applies the same test to autonomous aircraft. He has no objection to a $20 million aircraft doing the job of a $100 million one, provided it actually does the same job under realistic combat conditions.
His practical test is a question for anyone quoting a flyaway price: what else has to be operating for that aircraft to be effective? An inexpensive system may depend on tankers, airborne ISR, satellite communications, electronic warfare, command nodes and suppression of enemy air defences. “Those dependencies don’t appear in the flyaway cost,” he said, “but they absolutely appear in the war plan.”
Attritable is not the same as affordable
The distinction was Brown’s opening. Attritable aircraft, he suggested, are about operational risk, and he doubted that a force should be designed to lose large numbers of platforms. Deptula’s reply was that he was spot on.
Deptula went further, insisting on separating two ideas that the contemporary debate often fuses. In his words, attritability is about risk and affordability is about resources.
Willingness to risk an aircraft is an operational choice. Commanders have always decided where and when to accept losses, and uncrewed aircraft are especially valuable when a mission warrants risking the platform but not an aircrew. He pointed to MQ-9 Reapers in Operation Epic Fury, which provided persistent overflight and combat power despite significant losses. He added that at $30 million to $40 million each, they are hardly “attritable” in any meaningful sense.
Designing a force around the expectation of loss is a different matter. That, he said, is a force-development decision, and a potentially very bad one. Every loss reduces combat capacity, and replacement depends on factories, engines, sensors, munitions, transport and theatre logistics. “An aircraft destroyed on day three,” his Forbes article observes, “cannot be replaced on day four merely because it was inexpensive.”
His line on the matter is clear: attritability can be a useful characteristic, but it should not become a strategy. It is worth noting what he is not saying. He explicitly rejects the opposite error, an aircraft made so survivable that commanders are afraid to employ it. The aim is a force prepared to absorb attrition without making attrition its theory of victory. Brown distilled the exchange neatly: whether to put an aircraft at risk is a decision for the commander at the time, not a premise to build into force development.
Why attrition is the wrong contest with China
Here his argument takes on a strategic edge. Attrition, he said, is a condition of warfare, not an attractive strategy. Every serious force must be able to endure losses and keep operating, which requires reserves, industrial depth, weapons stocks and replacement capacity. But there is a major difference between preparing for attrition and planning to win through it.
Against China, he argued, the difference is decisive. He cited the Defense Department’s description of China as the world’s leading commercial shipbuilding nation, with the industrial capacity to produce large numbers of military vessels while also sustaining extensive missile and armaments production. He was careful to note that shipbuilding is not aircraft production and that the comparison should not be pushed too far. The strategic lesson holds regardless: why would the United States voluntarily choose a method of warfare that plays to one of China’s greatest structural advantages, scale and industrial output?
“I don’t want our theory of victory to become: we can replace inexpensive airplanes faster than China can,” he said. Against a production powerhouse, that may be precisely the contest Beijing would prefer.
His Forbes article applies the same caution to Ukraine. Small, attritable drones have been central to Ukraine’s defence, and Operation Spider Web showed what small weapons can do against high-value aircraft. But Deptula reads Ukraine’s approach as a product of constraint: a grinding war of attrition shaped in part by Ukraine’s lack of a modern air force with sufficient capability and capacity, and by Russia’s suboptimal use of airpower. American strategy, he argues, should aim to avoid that model of warfare rather than emulate it. The aim is to achieve air superiority rapidly, impose effects across far greater distances and at far greater scale, and prevent a peer conflict from devolving into a prolonged contest of attrition. Spider Web, he wrote, was ingenious and strategically valuable, much like the Doolittle Raid, but like that first raid on Tokyo it was not decisive.
Brown opened the session by asking Deptula to unpack the recent air campaigns, beginning with Ukraine, and his answer went further than the Forbes article. Ukraine lacks air superiority, he said, because it lacked capacity and the right equipment, and the war has settled into a stalemate reminiscent of the First World War. Necessity then drove invention. A quadcopter carrying a grenade is better understood, in his view, as a personally controlled precision-guided munition than as a drone, and the long-range systems striking Russian oil and electricity infrastructure are cruise missiles that happen to be propeller-driven. What they demonstrate is the value of strategic attack, done on the cheap. It is the damage to Russia’s economy, he argued, rather than the losses at the front, that is getting Moscow’s attention. The wrong lesson would be to emulate a form of warfare that suits a state fighting for survival and not one preparing to deter China.
Air superiority is what makes the alternative possible. It lets a joint force operate freely while denying the enemy the same freedom, and it opens the way to attack the systems that make an adversary effective: command and control, sensors, logistics, air defences and missile forces. The purpose is to impose disproportionate costs rather than trade platforms one for one. As he put it, use brains as well as mass.
CCAs: enablement, not substitution
None of this makes Deptula an opponent of autonomy. He called it potentially one of the most consequential developments in military aviation since the jet engine, and he said he wants the United States to lead the world in autonomous combat aviation.
His view of where CCAs fit is precise. He sees them principally as combat-air enablement, not combat-air substitution. A CCA might extend sensing hundreds of miles ahead of a crewed aircraft, carry additional air-to-air weapons, perform electronic warfare, act as a communications node, conduct reconnaissance, create false targets, or accept risk in environments where commanders would hesitate to expose a crew. Those are potentially revolutionary contributions.
The question that matters, in his framing, is not whether an F-47 can be replaced by three or five inexpensive aircraft. It is what an F-47 commanding or collaborating with several autonomous aircraft can accomplish that neither could achieve alone. That is the real opportunity, and it is a statement about formations rather than airframes.
Brown supplied the Australian example. The Ghost Bat, he said, grew out of operational analysis, and one of its motivating roles was to protect the E-7A Wedgetail. Deptula, who said the Australian programme with Boeing is heading down the correct trail, called that a great concept of operations: the CCA does not replace the E-7, it is melded with it to make the surveillance and control aircraft more survivable. That is the kind of prototype-driven concept development that shows how best to employ these systems, after which capacity decisions can follow. Brown then pressed the contrast with the United States. Deptula’s view was that the operational analysis there has so far been done largely on computers, and that the Increment 1 aircraft now flying have yet to take part in many large-scale exercises. Asked whether the danger was that the concept would be treated as a replacement for crewed aircraft, he agreed, naming the F-47 as the aircraft some would like CCAs to stand in for.
Deptula noted that the Air Force’s own description is consistent with this. CCAs are intended to extend reach, awareness and survivability while operating in tandem with the rest of the combat air force, and the Air Force describes them as “optionally attritable”, not disposable. He also drew attention to a tension in the programme. The service is moving towards roughly 1,000 CCAs, with an intention to field 500 by 2032, while simultaneously confronting the need to modernise its crewed fighters, bombers, tankers, command-and-control systems, munitions and base defences. As CCAs gain the capability and survivability to fight through contested airspace, their cost rises and the affordability-and-mass premise narrows.
Prove it, then buy it in numbers
His prescription is “experiment aggressively, test ruthlessly, prove the operational effects, and then buy them in substantial numbers.” What he objects to is the move from promise to force-structure assumption before the operational proof exists.
He welcomed what the Air Force is doing: flying CCAs, conducting weapons-integration testing, experimenting in operational exercises, developing tactics and testing portable command-and-control systems. But he set out the hard questions that still need answers:
• How well do CCAs function when communications are degraded, and how autonomous are they in practice?
• What happens when GPS is denied?
• Can they identify targets reliably?
• How do they perform against advanced electronic attack?
• What is their real range with a useful payload?
• What support footprint do they require, and how many operators are actually needed?
• How quickly can sorties be regenerated?
These, he stressed, are not arguments against the technology. They are the questions professionals should answer before betting force structure on it. His related point on design sequence is the one he returns to most. Systems should be designed from the mission backward, not from the price forward. Begin with the operational effects needed to defeat the adversary, determine the range, payload, survivability, sensing, autonomy and weapons required to produce them, then challenge industry to deliver them as cheaply as possible. Reverse the sequence, setting a price and asking what capability fits inside it, and the result may be a wonderfully affordable aircraft that cannot perform the mission the force exists for. In his words, the ambition is not an aircraft whose principal virtue is that it is cheap enough to lose.
He applies the same standard to the Air Force’s massed modular aircraft concept, whose business case, he argues, closes only if it produces the desired effects without requiring so many aircraft that total programme cost exceeds that of fewer, more capable ones.
Brown offered the strongest case for modularity. On a CCA, he noted, payloads can be changed rapidly, whereas altering anything on a complex crewed platform such as the F-35 or F-47 is slow. Deptula accepted that modularity is a good concept and went straight back to cost per effect. If the mission needs three payloads at once, it takes three aircraft, and the total can approach the price of one more capable platform that already carries all three. If it takes three to do what one can do, he said, it is a toss-up. Modularity has to be more than a slide in a PowerPoint deck. The devil, he said, is in the details.
Connected when possible, operational when disconnected
A further risk sits in the networks that tie a modern force together. Deptula’s standard is simple to state: assume from the outset that a sophisticated adversary will attack communications, satellites, GPS, data links, cyber infrastructure and the electromagnetic spectrum. The requirement is therefore to be connected when possible and operational when disconnected.
Brown brought personal experience to the point. In Operation Iraqi Freedom, he said, he would have a dollar for every time a communications or network failure disrupted collection from an uncrewed surveillance aircraft. In a high-end fight every network will be under pressure, so how much confidence can anyone place in a highly networked autonomous system? Deptula’s answer was to ask how much external input a system needs in order to succeed, which is the question of organic capability against offboard capability. There is a balance to strike, he said, and the issue reaches beyond autonomous aircraft to mission command: a system cut off from its network must act and think independently. Many operations run on accurate information arriving within seconds or minutes, generally from space. If that link goes away, there has to be an alternative.
For autonomous systems, this means sufficient onboard sensing, processing, decision-making and mission understanding to keep contributing when communications are degraded or intermittent. Otherwise, as he put it, the enemy does not need to shoot the aircraft down. He only needs to sever the connection that makes it useful. His Forbes article notes that China is developing the ability to disrupt and disconnect US weapon systems through electromagnetic-spectrum warfare, and that inexpensive aircraft dependent on continuous communications may become ineffective when data links are cut.
The point also feeds back into cost. A system that looks inexpensive but needs satellite links, relay nodes and escorts to function carries those costs and vulnerabilities whether or not they appear in the price.
Geography is unforgiving
“Distance is not a detail,” Deptula said of the Indo-Pacific. “Distance is one of the defining characteristics of the problem.” A low acquisition price does not solve a shortfall in range, payload or survivability.
Brown put the same issue to him directly. If range, payload and survivability command a premium, is unit cost a simplistic measure for the Pacific? Deptula agreed emphatically and offered a comparison for his Australian audience: the distance from Canberra to Beijing is greater than the distance from Beijing to Honolulu, a fact most Americans do not appreciate. What matters is not only range but how fast a force can get there, what it can do on arrival and whether it relies on offboard systems or organic capability. Those factors shape investment, he said, and explain why the Air Force is at last expanding its production objective for the B-21, a step he called long overdue.
The consequences run through the whole force. Short-range aircraft require more tanking. More tanking requires more tankers, bases and protection. Forward bases require air and missile defence, fuel, logistics and repair capacity, and communications must function across enormous distances. All of those supporting elements become targets in their own right. Against an adversary that has spent decades building a military architecture to keep US forces at a distance, with long-range missiles, integrated air defences, contesting fighters and attacks on the networks that connect the force, range, payload, survivability, autonomy, sensing and resilient communications are, in his phrase, “not luxuries—they are requirements.”
For Australia, the implications are direct. Deptula said Australia does not need to replicate the United States Air Force, but it does need sovereign combat power sufficient to generate meaningful effects at the distances its geography imposes, without assuming that every operation begins from a sanctuary close to the target. That means range, survivability and weapons depth. It means resilient command and control, hardened and dispersed basing, fuel, maintenance, and air and missile defence. And it means numbers, correctly understood.
He sees real promise for autonomy in this setting. Autonomous systems may let a relatively modest force extend its sensing, weapons capacity and geographic coverage. But their value depends on integration with the rest of the enterprise. The right question for Australian force design, he argued, is not how many inexpensive aircraft the country can afford. It is what combination of capabilities gives Australia the sovereign ability to generate and sustain the combat effects its strategy requires.
Too few, and so too expensive
One of the most useful parts of Deptula’s case concerns why advanced aircraft look so expensive in the first place. His answer is that we often make them so by buying too few.
Brown teed up the point. The B-2 was built in small numbers and the F-22 production run was cut short, he observed, and both carried heavy development costs that fed into high unit cost and into difficulties with maintainability.
The B-2 was conceived around a fleet of more than 100 aircraft, and only 21 were delivered. The Air Force itself has identified that tiny fleet as creating unique sustainment challenges, and the active inventory today is 20. The F-22 requirement was 381 aircraft and production ended at 187. When production is truncated, fixed costs do not disappear. Development, tooling, software, test infrastructure, training and sustainment systems must be spread across far fewer aircraft. The supplier base is disrupted, economies of scale are lost, and the owner eventually pays heavily to sustain small fleets of unique systems. In the seminar Deptula added that the F-22’s original planned inventory was 750 before it fell to 381, and that the programme was terminated by Secretary Gates when it was well down the production learning curve and building aircraft at the lowest unit cost it had reached. That, he argued, is precisely when production should be increased.
“Then we point to the high unit cost,” Deptula said, “and say, ‘See, advanced aircraft are unaffordable.’ That’s backwards.” Sometimes, he observed, the most expensive way to buy an aircraft is to buy too few of them. His Forbes article draws the practical conclusion: maximise production of the F-35, F-15EX, B-21, F-47, KC-46, E-7 and CCA to bring down unit costs and build realistic attrition reserves.
Strategy, resources and the honest conversation
Underneath the force-design questions lies a harder one about money. Deptula acknowledged that a force combining advanced crewed aircraft and autonomous systems makes the resource problem harder, and that pretending technology will make the problem disappear does not help. His diagnosis is that the underlying difficulty is a mismatch between national strategy and the resources allocated to execute it.
Brown put the funding question bluntly. If the US Air Force has been underfunded for years, and the plan is to field both crewed aircraft and a large number of CCAs, where does the money come from? Deptula’s answer was that the case has to be made to Congress and the public that the strategy needs funding at that level, or else the strategy must be throttled back. The Air Force is already stretched by operations against Iran, he said, before any contest with China is considered.
He pointed to the breadth of what the US FY27 proposal must fund at once: F-35s, F-15EXs, CCAs, tankers, munitions, readiness, sustainment, infrastructure and next-generation capabilities. There are efficiencies to pursue: reduced overhead, acquisition reform, more competition, open architectures and higher production rates. But eventually there is a strategic choice. Either resource the force the strategy requires, or change the strategy and accept the risk. “Cutting the force while leaving the strategy unchanged is not a solution.”
His proper sequence is worth stating in full. First identify national security objectives. Second, determine the military strategy needed to achieve them. Third, determine the force structure and capabilities required to execute that strategy at acceptable risk. Only then work out what it costs. If political leaders judge the price unacceptable, that is their prerogative, but it should prompt an honest discussion about adjusting the strategy or accepting more risk. Reversing the sequence, starting with the budget and then constructing concepts suggesting that the resulting force is sufficient, is what he called “rationalisation”, not strategy.
Brown tested the point by asking what single design assumption Western air forces should change. Deptula’s answer was to stop letting arbitrary budgets drive defence strategy and to fund the national defence strategy according to the demands it places on the force. That requires leaders willing to make the case. They work within the budget they are given, he said, but they can still say what the strategy asks for, and the gap between the two is risk: how much are they willing to accept?
His closing message on autonomy follows from this. A force multiplier multiplies something. If decades of underinvestment keep shrinking the number of penetrating fighters, bombers, tankers, electronic warfare aircraft, weapons, bases and crews, eventually there is not enough underlying force left for autonomy to multiply. His bottom line: embrace autonomy, accelerate it, test it under the most demanding conditions, then field it at scale, but use it to build a larger and more capable combat air force, not to explain why a shrinking one will somehow be enough.
Gaps, risks and opportunities
The session was framed around gaps, risks and opportunities, and Deptula’s material maps onto that triad cleanly.
The gaps are capacity and proof. Capacity has been eroded by decades of retiring aircraft faster than they were replaced and truncating production. Proof is missing because the operational evidence for CCAs under electronic attack, degraded communications and denied GPS is still being gathered. Brown added a third, which is advocacy: air forces are not yet good enough at explaining what airpower takes to the people who fund it.
The risks are three. One is the substitution trap, in which the promise of autonomy becomes the justification for cutting proven capacity before it has been demonstrated. Another is the attrition trap, in which a force built to absorb losses ends up competing on the industrial ground where an adversary is strongest. The third is the dependency trap, in which an apparently inexpensive system relies on a web of tankers, nodes and links that an adversary can attack.
The opportunities are real and substantial. CCAs can extend sensing, add weapons capacity and accept risks that commanders would not impose on crews. For a middle power such as Australia, they may expand the reach and weapons depth of a modest force. The condition is that they are designed from the mission backward and integrated into an enterprise with the depth to sustain them.
One reading of the argument is worth pressing, and it is my own rather than Deptula’s. His framework treats cost per effect as the decisive test, and it is a good one. But the test is only as sound as the assumptions fed into it, and those assumptions are exactly where the debate between affordable mass advocates and their critics now lies. How survivable are crewed penetrating aircraft and their bases against a missile-heavy adversary? How quickly can a force regenerate sorties rather than merely count aircraft? How much of a package’s cost is truly fixed and how much is scalable?
His emphasis on regeneration, on factories, engines, crews and logistics, points towards a broader proposition: that airpower is an enterprise rather than a platform. Seen that way, the real contest is between competing enterprises, and the one better able to regenerate effects under pressure holds the advantage. That is a harder metric to apply than unit price, and a more honest one.
The measure that matters
Deptula’s Forbes article closes on a line that his seminar remarks reinforced. The measure of a combat aircraft is not how cheaply a nation can afford to lose it. It is whether it can accomplish the mission, survive where possible and return to fight again, so that the war can be won as rapidly and decisively as possible. What costs more than a first-rate, full-capacity air force, he concluded, is a second-rate one without the capability or capacity to win.
For air forces in the Indo-Pacific weighing how to build mass and depth with autonomy in the mix, the lesson is not to choose between crewed and uncrewed, or between high-end and low-cost. It is to start with the effects required to win, test what the new systems can truly deliver, and resource the force the strategy demands, or be candid about the risk of not doing so.
See the following for the Forbes article: