Disrupting the Pentagon
America’s military supremacy is in jeopardy
To win future wars it needs new weapons, new suppliers and a new system of procurement
ON THE FRONT lines in Ukraine, war is not nearly as foggy as it used to be.
Satellites and drones equipped with many kinds of sensors are always scanning every inch of the battlefield, while artificial intelligence (AI) instantly interprets the data they gather.
It is far easier than it would once have been for either side to spot and attack anything that moves—one reason why big, old-fashioned offensives have made so little headway.
America has played a big part in these changes.
It has helped Ukrainian forces build drones that are more capable yet cheaper than those deployed in Afghanistan and Iraq, for instance, and develop an AI “kill chain” in which targets are identified and munitions guided to them, often deep behind enemy lines.
American firms, too, are in the vanguard of this new era.
They visit Ukraine regularly to observe how their weapons are performing and adapt them accordingly.
Private capital is flooding into American companies that aim to disrupt the conventional battlefield.
In December Palantir, a data firm, became the world’s most valuable defence contractor, supplanting RTX, an aerospace giant.
Palantir’s market capitalisation was a fraction of RTX’s when Russia invaded Ukraine in 2022.
Now it is much bigger (see chart 1).
Asleep at the cutting edge
Yet for all America’s expertise and innovation, not to mention its vast defence budget, its own armed forces are struggling to adapt.
Mark Milley, until late 2023 America’s most senior soldier, argues that the military machine that he oversaw for four years is, in essence, unfit for purpose.
In an article he co-wrote last year with Eric Schmidt, a former CEO of Google and backer of a fund investing in military technology, he notes that American firms make the best AI systems, but its armed forces struggle to absorb and deploy them.
Troops lack the equipment and training to cope with a drone-saturated battlefield.
It takes America years to buy weapons that are evolving month by month.
The Pentagon, the pair conclude, needs “a systemic overhaul” in how it fights, what it buys and how it does its shopping.
America’s new president seems to agree.
This week Donald Trump announced that he was directing Elon Musk, his woodchipper-wielding agent of change, to take on the Pentagon.
Pete Hegseth, the defence secretary, complained in his confirmation hearing last month that the Pentagon is “too insular [and] tries to block new technologies from coming in”.
The new national security adviser, Mike Waltz, recently told an interviewer, “We do need great minds and we do need business leaders to go in there and absolutely reform the Pentagon’s acquisition process.”
Mr Trump is appointing tech types, such as Emil Michael, a former executive at Uber, to senior defence jobs.
They have their work cut out.
To remain a world-beating military power, capable of waging and winning a war with China, they will need to change three things.
The first is the armed forces themselves: how they fight and what they fight with.
The second is the defence industry that supplies them, which needs rebalancing towards newer, more innovative companies.
The third is the least understood and the most resistant to change, but vital to fixing everything else: the pork-barrel politics of defence spending.
There is a heated debate about how much change is needed within the armed forces themselves.
Some observers believe that technology is hyped and that orthodox measures of military strength, such as sound tactics, troop numbers and ammunition stockpiles will remain more important.
AI will struggle with complex tasks and commanders will not trust it, they argue; difficult terrain will flummox robots.
America, this camp thinks, should be pumping money into a bigger army, more shipbuilding and more high-end missiles.
At the other extreme are radicals who believe that everything has changed.
An outspoken exponent of this view is Mr Musk.
He recently responded to a Chinese drone display by harrumphing, “Some idiots are still building manned fighter jets like the F-35”—implying that swarms of cheap drones could easily defeat a vastly more expensive warplane.
A similar, if less abrasive assessment can be found in “Unit X”, by Raj Shah and Christopher Kirchoff, in which the authors describe their struggle a decade ago to establish a forward-looking arm of the Pentagon called the Defence Innovation Unit (DIU).
They, too, would do away with entire classes of weaponry.
“In a world where hypersonic weapons and anti-ship missiles can easily destroy a navy vessel,” they write, “it no longer makes sense to spend billions of dollars building destroyers and battleships.”
In the middle are more cautious modernisers.
They believe it is too early to shred America’s existing force structure entirely.
AI and autonomy have not yet advanced to the point where software can handle all the tasks of a human pilot, they insist.
Uncrewed surface vessels have worked well in the Black Sea, but would find it hard to cross the Pacific or navigate mid-Atlantic storms.
The Chinese drone swarm, they point out, was a choreographed light show which would have fallen to pieces in the face of jamming.
Physics imposes hard limits: a small drone will never be able to carry serious firepower across oceans while evading Chinese air defences.
“Not everything needs to be gold plated and have all of the capabilities,” says Aditi Kumar, who was DIU’s deputy director until January 20th, “but some systems do.”
In this view, America needs what experts call a “high-low mix”, in which a modest number of complex and high-end platforms and munitions operate alongside a much larger number of cheaper, simpler and mostly uncrewed weapons.
Fancy weapons are still essential—American ATACMS and Anglo-French Storm Shadow missiles have made mincemeat of Russian air defences in Crimea, for instance.
But they can be used to clear the way for cheaper ones.
Or both sorts can be used in tandem, with “loyal wingmen” drones flying alongside crewed jets, for example.
The Goldilocks army
The optimal mix of forces is hotly debated.
The question of which high-end weapons to scrap and which to keep is especially delicate.
Speaking last year, Mr Milley proposed doubling the number of submarines.
These are some of the most complex and expensive weapons in America’s arsenal, but their stealthiness means that they can survive on a sensor-saturated battlefield.
But he cast doubt on aircraft-carriers.
“We’re going to have a ship that sails with 5,000…sailors on it and the thing will be dead in the water in 20 years, for sure,” he argued.
“We can do all the electronic magic [to jam missiles], but the bottom line is: it’s a big piece of steel that saw its best day at Midway.”
Mr Milley was similarly doubtful about the F-35: “do we really think a manned aircraft is going to be winning the skies in 2088?”
The risk of this approach is that if America finds itself at war before 2030, it could be caught short—equipped with too few legacy systems like the F-35 and not enough new tech to compensate.
Regardless of what America buys, the next question is who should build it.
American defence contracts typically involve one buyer, the government, which sets the requirements and carries all the costs of research and development (R&D), which helps explain its conservatism about what it buys.
This set-up allows private firms to take on projects with huge costs and long timelines, which would otherwise be too risky: think aircraft-carriers or bombers.
But it is a hopelessly unsuitable approach to buying smaller, software-infused gear that has to be updated constantly to remain effective.
In October Shyam Sankar, the chief technology officer of Palantir, published a 4,000-word manifesto titled “The Defence Reformation”. He lamented that the number of big defence firms selling weapons to the Pentagon has shrunk from 51 in 1993 to five today (see chart 2).
“Consolidation bred conformity,” he argued, “and pushed out the crazy founders and innovative engineers.”
It also changed the identity and ethos of arms-makers.
Mr Sankar, who turned down a top job in Mr Trump’s Pentagon, notes that before the fall of the Berlin Wall only 6% of American defence spending went to specialist arms-makers.
Most contracts went to companies that had both commercial and military arms.
Ford made satellites until 1990, he notes, just as General Mills, better known for its cereal and cookies, made guidance systems for ICBMs.
The commercial world kept them competitive and forced them to invest in research and development at their own risk.
Today specialist defence contractors account for 86% of defence spending.
Michael Brown, a former head of DIU who is now a partner at Shield Capital, a venture-capital fund, notes that two-thirds of the business of the top ten suppliers to the Pentagon is defence only; in China, the equivalent figure is 30%.
“Think about the difference in your mentality as a company.”
As weapons become increasingly reliant on data and code, both the Pentagon and its suppliers are under pressure to absorb the tech world’s risk-taking mindset and rapid development in the form of frequent upgrades and increasing use of AI.
“We’re taking cheaper, commercial off-the-shelf components, using software provided by the government,” says Ms Kumar, referring to DIU’s drone development, “and integrating the two and fielding a capability that is substantially lower cost than what the department has been able to do before.”
Big Silicon Valley firms are coming around to working with the Department of Defence.
On February 4th Google reversed a long-standing policy barring the use of its AI tools for military purposes.
Tech giants and top AI model-makers, such as OpenAI, Anthropic and Meta, have also embraced military business.
“They didn’t want to do anything with DoD,” says Yili Bajraktari, who used to work at the Pentagon and now leads the Special Competitive Studies Project, a research group.
“Now you have the tech companies all in.
They are gung-ho about DoD.”
So are investors.
Venture-capital deals in the defence industry have grown 18-fold over the past decade, from $500m in 2014 to about $8.7bn in 2024, according to Bain & Company, a consulting firm.
Money can’t buy you readiness
It will be difficult, however, to harness corporate America’s enthusiasm using the government’s procurement system.
America’s defence budget, at over $800bn a year, is far and away the world’s biggest, but allocating it is a ludicrously slow and political process.
In theory, the administration lays out its defence strategy, the service chiefs tell the secretary of defence what they need to fulfil those goals and the administration then requests the sums necessary from Congress.
In practice, things are not nearly so simple.
Service chiefs sometimes lobby Congress directly to approve pet projects.
Lawmakers often prevent the Pentagon from retiring obsolete weapons if their home state will be harmed.
And Congress micromanages, allowing the Pentagon to move around no more than $6bn within the budget, and even then only with the approval of senior members of Congress for each slice of $15m or more.
The most baleful consequence of all this is interminable delay.
Drones in Ukraine have their software, sensors and radios swapped out every six weeks or so.
Year-old AI is archaic.
Yet the gap between the start of the Pentagon’s budget process and any money appearing is—at a minimum—two years.
Political deadlock means that budgets are rarely passed on time anyway, leading to “continuing resolutions” in which new programmes cannot be started.
Last year saw a six-month delay. Simplifying and accelerating all this, says Mike Horowitz, until recently a deputy assistant secretary of defence, is “the secret to unlocking the innovation problem”.
It is possible to persuade Congress to pay for innovative new schemes, but it is hard work.
In August 2023 Kathleen Hicks, the deputy secretary of defence at the time, announced that the Pentagon planned to buy “multiple thousands” of easily upgradeable drones to be ready within two years.
This “Replicator initiative” marks “huge progress”, argues Mr Horowitz, who helped run it.
“Compared to ‘Pentagon standard’,” he argues, meaning overpriced and slow-to-arrive kit, “Replicator is delivering a lot of capability fast at a low relative price point.”
Mr Brown notes that drones with a pricetag of $17,000 apiece during his tenure (2018-22) now cost less than a tenth of that.
But to talk Congress into allocating just $500m to Replicator—about half of one percent of the defence budget—Ms Hicks and her team had to conduct nearly 40 briefings.
Moreover, this sort of streamlined procurement is not catching on across the Pentagon, notes Mr Brown: “That’s the part that’s been disappointing.”
Another example is Other Transaction Authority (OTA), a procedural innovation that allows departments to buy things without getting bogged down in the Federal Acquisition Regulation, a 2,000-page Talmudic set of rules that has spawned a priesthood of procurement officers.
The Pentagon has spent $86bn via OTAs to date, mostly over the past five years, notes Austin Gray, who runs a defence startup.
But their use has now “plateaued”, laments Mr Brown.
The Pentagon’s lawyers are wary of rocking the boat, Mr Brown says.
“That kind of risk-averse mentality applies to so many things at DoD: don’t take the risk and stick your neck out, because it could get chopped off.”
Mr Sankar of Palantir recalls that when ChatGPT was released to public acclaim in 2022 his firm offered to include a similar chatbot free of charge in a product it was making for the army.
The army refused because it had not included a formal requirement for such a feature in the original contract.
The future is already herePhotograph: Getty ImagesRocking boats is a speciality of Mr Trump and Mr Musk.
Many in their orbit have big ideas about reform.
Mr Hegseth has promised “to hire a lot of smart people”.
Even some former officials from Joe Biden’s presidency believe that Mr Trump could shake things up for the better at the Pentagon.
“If the Trump folks pursue the innovation that some of them say they want,” says Mr Horowitz, “there’s a real opportunity—if they can effectively operate the Pentagon bureaucracy.”
Destruction +/- creativity
That is a big if.
At the moment the new administration’s focus seems to be cutting costs rather than reforming the process of procurement.
“We’re going to find billions, hundreds of billions of dollars of fraud and abuse,” Mr Trump said of military spending this week.
Mr Hegseth, for his part, seems concerned chiefly with rooting out wokeness in the ranks, although he is said to have recently met the ceo of Scale AI, a data firm.
But even if he harnesses the brightest minds in IT, they may not have a winning formula for wrangling obstreperous members of Congress.
Yet the task could scarcely be more pressing.
“I don’t think we understand the sense of urgency,” warns Mr Bajraktari.
“In Washington there is not a sense that there’s a war in continental Europe…and we might likely have a war in Asia.”
America’s armed forces will need a drastic overhaul if America is to remain the world’s pre-eminent power.
The sort of innovation that is required should come easily, given corporate America’s strengths.
Yet politics, as usual, is getting in the way.
“Everybody understands the problem we’re in right now,” says Mr Bajraktari.
“We’re the number one global power in software—and our military is not able to use it.”
0 comments:
Publicar un comentario