America’s Missing Military Revolution

Imagine buying new iPhones for yourself and your family. You visit a store to see the latest model, the iPhone 17, and then go to Apple’s website to buy yours. But before finalizing your purchase, you have to spend two to three years experimenting with your pick. By then, Apple will have released two newer models, perhaps an iPhone 18 and iPhone 19. Yet you are not allowed to upgrade your order—you have to stick with the model you tested.

Even then, your already two-year-old iPhone 17 cannot be delivered for two more years, at which point the iPhone 20 and 21 have been unveiled. And only your phone will be delivered at that time—your kids and spouse won’t get their iPhone 17s for another year or two. You may also be locked into whatever operating software was available on the iPhone 17 when you first tested it.

This is how the Defense Department has long been forced to buy technology, including missiles, drones, and the software that controls them. When Pentagon officials identify a problem, the process of finding a system that can solve it and purchasing it typically takes years. It then takes years more to acquire enough of that solution to meet the military’s requirements. By then, the solution is outdated. Relentless private-sector competition has spawned new technologies, but the military cannot buy them, much less do so at scale, because they have not gone through the same cumbersome testing, evaluation, and purchasing processes.

Having lived at the cutting edge of one side of this equation, as an Apple executive, and then at the other, as the official charged with driving defense innovation for the Pentagon, I’ve seen both sides up close and personal.

The entrenched, byzantine system for acquiring things is why the most sophisticated country in human history is at risk of no longer having the world’s most sophisticated military. The U.S. private sector sets the global standard for creativity, nowhere more than in tech hubs such as Silicon Valley. Yet the way the United States buys and deploys military technology is, like the rigid, Soviet-style five-year plans that even China and Russia have long since jettisoned, mind-numbingly bureaucratic. And despite real progress, the United States is still not changing this process fast enough, or at sufficient scale, to keep America safe in an increasingly dangerous world.

Battlefields from Ukraine to the Middle East have demonstrated with brutal clarity just how much warfare has changed. The breakneck pace of technological innovation makes it possible to roll out new autonomous weapons, AI algorithms, and other sophisticated systems that are updated constantly and easy to scale. The Russians are demonstrating every day in Ukraine that their army of developers can crank out new drones while updating hardware and software, sometimes daily, to match Kyiv’s constant advancements, which continue, in turn, to allow Ukraine to keep Russia at bay even though its armed forces and economy are far larger. China has embarked on a massive military buildup, including increasing its production of ships, satellites, missiles, and drones, and has demonstrated the ability to adopt and integrate the latest technology into them. Even Iran has successfully used new technologies to effectively close the Strait of Hormuz this year, despite being completely outmatched by American military power.

The United States faces simultaneous challenges from China, Russia, Iran, North Korea, and elsewhere. Chinese leader Xi Jinping has told his military to be prepared to defeat the United States in the Taiwan Strait by 2027. War with China is neither inevitable nor imminent. But Washington and its allies must be able to deter Xi, and they must be ready to win if forced to fight. To do so, they must harness the power of commercially derived technology.

Americans understand that the acquisition system is broken. The Trump administration, the Biden administration before it, Democrats and Republicans in Congress, military leaders, and the tech sector’s founders and funders have all called for dramatically changing the way the Department of Defense buys and deploys technology. This consensus represents progress. Ten years ago, only a small group of people were arguing that the Pentagon needed to make such a transformation, or that Silicon Valley’s commercial technology was pivotal for military power. Now, it is central to every defense-related conference, to policies originating in the Pentagon and on Capitol Hill, and to the investment theses of an ever-growing, dynamic defense tech sector.

But Americans would be foolish to think that Washington is anywhere near solving this problem. It is not. The United States is not yet delivering on this strategic imperative or making the permanent transformational changes necessary to do so. Progress has been relatively isolated. Successes have captured the news, but improvements have taken place largely at the military’s margins. Even the herculean efforts to acquire capabilities for the conflict with Iran largely amounted to a collection of waivers, special actions, and brute-force leadership.

To deliver real, enduring change at the scale required, the entire system needs to be transformed. The Pentagon, Congress, and the private sector need to build on the momentum of the last five years, and shed, once and for all, their mutually reinforcing pathological caution when it comes to purchasing technology. They need to get comfortable, together, taking risks, making mistakes, and quickly become adept at innovating at scale—just as private-sector companies, investors, and customers do every day. Otherwise, the changes thus far will amount to little more than rearranging the deck chairs on the national security Titanic.

RUNNING IN OATMEAL

The origins of the current military procurement system date to 1947, when Congress passed a law meant to standardize and simplify Washington’s chaotic purchasing process from World War II. The new rules worked well as the country scaled up to meet the Soviet threat through the 1950s. But they also gave rise to new risks as the military, defense contractors, political leaders, and science and technology institutions forged an increasingly tight relationship, creating the “military-industrial complex” that President Dwight Eisenhower warned about in his 1961 farewell address. The complex, Eisenhower said, threatened to self-deal, overspend, and ultimately misdirect U.S. policy. His predictions proved prescient. Over the next several decades, cost overruns and defense scandals prompted Congress to create ironclad rules emphasizing competition, auditability, documentation, and compliance and to add procurement regulations aimed at safeguarding fairness, frugality, and transparency.

These laws were well intentioned and served (and still serve) an important purpose. But as they accumulated over decades, they resulted in a stunning morass of regulatory restrictions, policies, and learned behaviors, reinforced by a web of entrenched interests both inside and outside government, that have slowed procurement and adoption.

The problems begin before anything is ordered, when requirements processes yield hundred-page documents that dully describe the precise product the government wants built rather than letting the tech sector’s experts determine how best to apply new technology to solve military problems. These documents might be fine for crafting orders for a legacy capability that military experts are familiar with. But they do not work when applied to commercially derived technology that moves at the speed of a market driven by billions of global consumers. The most innovative companies, not their customers, typically know how technology might leap ahead. If the engineers at Apple, where I worked from 2009 to 2023, had asked average consumers in 2007 what they wanted a smartphone to be, the answer would likely have been a faster flip phone, in brighter colors. The world would never have seen an iPhone, an App Store, or all the other innovations that followed.

The Pentagon’s risk-averse culture, reinforced by rules established by Congress, encumbers every stage of the military development, procurement, and integration process. Military leaders who understand the need for speed often feel as though they are running in oatmeal. The Pentagon takes the same experimentation rules it uses for crewed aircraft that fly above populated areas and applies them to drone experiments held far out at sea, meaning that drone testing is constrained by procedures designed to protect human pilots and people on the ground. Minute software updates and bug fixes that the commercial world might complete several times a day in the run-up to a product launch can require multiple meetings and, ultimately, sign-off from a three- or four-star officer. Perhaps most damaging of all, the budgeting process established by Congress and the Defense Department locks in specific product designs at least two years in advance. The integration of that product into military planning, training, and staffing doesn’t even begin until after the acquisition process is complete and the Pentagon is ready to field the product. As a result, it is nearly impossible for the Pentagon to make bets on truly new technologies or adapt at the constantly accelerating speed of war.

A drone inside a manufacturing facility, Ashville, Ohio, March 2026 Megan Jelinger / Reuters

The myriad processes governing acquisitions are implemented by an army of patriotic, hard-working officials. But they are incentivized to never color outside the lines. If a system has defects or major cost overruns, procurement staffers, their bosses, and their bosses’ bosses face congressional inquiries and outside investigations. They might lose their jobs. Many fear they could wind up in prison if their actions are misinterpreted as corrupt. But if soldiers die or the U.S. military loses because it doesn’t have the most up-to-date innovations, no one goes back five years to find the staffer whose conservatism led to that result. Little wonder, then, that procurement teams exercise extreme caution. Even after Secretary of Defense Pete Hegseth’s November call to reward outcomes and risk-taking rather than adherence to process, the Pentagon’s performance incentives remain structured around compliance, prescribed milestones, and avoiding cost overruns.

All these problems are aggravated by the dysfunctional codependence of procurement officials and traditional defense contractors. Companies that have grown up meeting increasingly complicated bureaucratic needs are compliance machines. As much as ten percent (and perhaps even more) of their staff work on law-related, contracting, and cybersecurity teams. Another up to 20 percent are experts dedicated to helping navigate the Pentagon’s bureaucracy. At normal tech companies, legal, compliance, and government affairs teams make up an average of two to five percent of headcount—and much less at tech startups. Those companies are not keen on hiring an army of employees or consultants to manage relations with the military—especially when it is far from clear it will buy their products. Some do so anyway. But they then risk becoming like many existing contractors: more focused on how to work with the government than on how to create breakthrough inventions.

It is thus no surprise that the American military acquires capabilities at a speed completely out of touch with the pace of modern change. Ukraine will pump out roughly seven million drones this year. It updates its drone software weekly and modifies its hardware every few weeks to adapt to changes on the ground. The United States, by contrast, purchased just 4,000 small drones in 2024. It has publicly claimed that it purchased roughly ten times that figure in 2025, but many of those drones are older models that have not been updated in response to lessons learned in Ukraine or recent conflicts in the Middle East—for example, the need to cope with GPS jamming. The department has pulled out all the stops to buy more drones for the war in Iran. It has requested funding from Congress to purchase up to 340,000 modern drones over the next two years for its Drone Dominance Program and has contracted for the first 30,000 so far in 2026. The army is aiming even higher, hoping to procure a million drones over the next couple of years. But these aspirational numbers are still nowhere close to Ukraine’s, and reaching them—much less converting them into fielded military power—depends on massive, unrealized changes in the military, Congress, and the private sector.

NEED FOR SPEED

The Defense Department knows it has a problem, and it has spent the last decade taking steps to solve it. In 2015, Secretary of Defense Ash Carter established the Defense Innovation Unit (then called DIUx) specifically to help deliver commercially derived technology to the military. The organization had a slow start. In early meetings, tech leaders wearing jeans and T-shirts sat across from DIU defense officials wearing suits, each side speaking a language the other couldn’t understand. But after Carter reset the unit later that year by bringing in new leaders who were native speakers of both the tech world’s dialect and the language of warfighting, the organization succeeded in building a bridge between the two communities.

This revamped unit, DIU 2.0, proved that real military problems could be solved with commercially derived technology. It also showed that the commercial tech sector could create those solutions quickly, delivering prototypes within weeks or months rather than the government’s usual five to ten years. The team achieved this by pioneering the use of “other transaction” acquisition authorities (OTAs), which allow the department to bypass most of its tedious procurement rules when including nontraditional suppliers. These authorities, legally available since the space program began in the late 1950s but almost never used, helped DIU drive a rapid, competitive acquisition process that responded to real needs. Its procurement team featured both frontline military operators and technologists and focused on speed as well as quality. Between 2017 and 2023, that team delivered solutions at a success rate of over 50 percent—meaning the prototype was accepted by the military—across hundreds of projects and helped innovative companies such as Anduril get off the ground. It brought in mission-planning tools, drones, counterdrone technology, and space launch capabilities.

But prototypes neither deter nor win wars, and they do not deliver the kinds of financial returns companies need to make sizable investments. Thanks to the Pentagon’s bureaucratic processes and entrenched bias toward traditional sources, DIU struggled to help the innovations it introduced go from prototype contracts to fielded capabilities. The military did not get the strategic impact it needed. Investors grew antsy. Katherine Boyle, of the venture capital firm Andreessen Horowitz, warned in 2021 that “time was running out with Silicon Valley.”

Then, in 2023, Secretary of Defense Lloyd Austin elevated DIU so that it reported directly to him instead of to the undersecretary of research and engineering. I was the new director, and Austin charged us with inventing a fresh strategy. The resulting DIU 3.0 was based on five pillars. The first was embedding DIU teammates with the staff of each four-star combatant commander—the commanders who oversee U.S. military activity in a given region or function, such as cyber-operations—to keep DIU focused on the highest-leverage problems. For example, the commercial tech lead for Admiral Samuel Paparo, who helms U.S. Pacific Command, was an embedded DIU teammate charged with helping close critical operational and strategic gaps in the Pacific and mobilizing the rest of DIU and the Pentagon to do so.

The second pillar was to partner directly with each of the service branches (army, navy, air force, space force, Marine Corps, and Coast Guard) that man, train, and equip the military, to ensure that initiatives conceived through the first pillar aligned with their priorities and had what the new strategy called a “pathway to scale.” The third pillar called for turning the more than 200 Defense Department entities focused on innovation into a cohesive community. The fourth pillar focused on ensuring DIU’s depth, its connectivity to the tech sector, organic tech expertise, and knowledge base, in each of its tech portfolios—such as autonomous weapons, AI, space, human systems, cyber-capabilities, and energy—so the Defense Department could access the latest and best innovations. This pillar also focused on removing barriers for firms working with the military, to broaden the base of suppliers. The fifth pillar called for building relationships with allies and partners so that American and allied militaries could grow stronger together.

Congress then codified Austin’s organizational switch into law, making the DIU director a “principal staff assistant”—the same level as an undersecretary of defense—with a seat at the Pentagon’s leadership table. It increased DIU’s budget to nearly $1 billion and changed the nature of the budget itself. Beginning with the 2024 National Defense Appropriations Act, championed by Ken Calvert, the chair of the House Appropriations Subcommittee on Defense, Congress gave DIU groundbreaking flexibility. DIU could move money within its portfolio to meet the most critical needs and fund ideas all the way from conception to the initial fielding of operational units. This facilitated real-time experimentation and optimization. DIU could now help partners in the military services buy new technology, quickly prototype and test it, and begin deploying it in a single year. Then, in the second year, DIU could help the services begin scaling that new technology while continuing to update it. Now, in the third year, even the cumbersome current budgeting dynamic allows true scaling by the services, assuming they plan for that expansion when first collaborating with DIU in year one. DIU and the services, working together, finally have the tools to make the iPhone analogy obsolete.

To see how this works in practice, consider Saronic, a young defense tech startup. In January 2024, DIU, working with the navy, announced it would field proposals for a new model of highly capable, inexpensive maritime surface drones, similar to those the Ukrainians had famously used to sink parts of Russia’s Black Sea Fleet despite having only a tiny navy of their own. Over 100 companies responded to that solicitation, from established defense contractors to startups. In July, DIU awarded prototype contracts to three applicants, including the venture-backed Saronic, a small company led by a former navy SEAL. By December, Saronic’s prototype boats had passed rigorous testing by navy operators, DIU technologists, and others, earning a DIU “success memo” for scaling. DIU used its new flexible funding to buy the first operational units for a freshly created navy squadron so that its sailors could train with these boats and develop tactics for them as they were produced. The navy then awarded Saronic a $392 million contract in mid-2025, enough for hundreds of vessels. Already, Saronic’s boats have helped the U.S. military in Iran, including by rescuing helicopter pilots stranded off the coast of Oman earlier this year. The company’s valuation reflects this rapid acceleration: Saronic raised $600 million in Series C funding in February 2025 and $1.75 billion in Series D funding in March 2026 at a $9.25 billion valuation. None of this would have been possible without DIU, its flexible funding, and its partnership with the navy.

Walking through a military manufacturing facility in Louisville, Colorado, June 2026 Cheney Orr / Reuters

Congress paired the unprecedented trust and leeway it granted DIU with new reporting requirements for transparency and oversight. But these are reasonable: DIU must advise Congress of its spending plans before it executes them. It can still implement them as it sees fit, choosing the best technologies from the best companies. DIU has taken transparency even further, providing Congressional appropriators with updates as often as every two weeks on every project, including news both good and bad. This has unsettled some Pentagon officials who fear the scrutiny will invite meddling. But although Congress has asked hard questions, it has not intervened even once. Instead, openness has built trust, preserving and bolstering DIU’s freedom of action in support of the military’s most pressing needs. DIU has put that freedom to work across a broad array of critical fields, including drones and counterdrone technologies, AI-based command-and-control systems, AI-driven biological threat detection, and many others.

The Pentagon’s current leadership has taken steps to build on that progress. Hegseth has called for a “Warfighting Acquisition System” to propel acquisition reform more broadly and mandated the use of DIU’s OTA process for all software purchases. On priorities such as the Golden Dome missile defense system, submarines, and drones, he has centralized decision-making. Congress has issued multiple legislative proposals, and the White House has launched a steady stream of executive orders, signaling support for innovation.

Meanwhile, commercial firms are investing far more in defense innovation than at any time since the early Cold War. In fact, it is one of the few areas of venture capital that has grown in recent years. In 2025, venture firms invested between $30 billion in defense technology alone and $50 billion if dual-use companies are included, up from less than $10 billion (some estimates suggest less than $2 billion) in 2020. That number has continued to rise in 2026, with over $35 billion invested in defense technology and dual-use goods in the first half of the year alone by both established firms and new actors dedicated to the industry.

The American military now has more uncrewed systems in the air, on the sea, and under the sea. It has used them to strike targets in the Middle East. It is adapting tactics worldwide to reflect new capabilities. It has access to more AI models and is making better use of them, including for improving its defenses against adversaries’ drones. It is developing new tech solutions across the spectrum, and it is getting better at doing so.

ROOT AND BRANCH

Still, these steps are not nearly enough. Even with all the expanded authorities, executive orders, new policies, and public pronouncements, somewhere between one and three percent of the U.S. defense budget is spent on truly innovative technologies. DIU’s own budget is around one-tenth of one percent of total defense spending. And only about one-third of one percent of contracts written are OTAs, whether from DIU or elsewhere, representing slightly less than three percent of new funding commitments. The rest remain mired in the old way of doing things, and the vast majority still go to traditional defense contractors. The reforms of recent years have created positive, accelerating momentum, yet they remain workarounds rather than changes to the system. Washington needs to take much more fundamental steps toward real, systemic change that will outlast this administration, and the next.

First, the United States should reinforce and scale its early successes in accessing, prototyping, delivering, and integrating commercially derived technology. This means strengthening DIU’s role in leading the dual-fluency world—people who understand both tech and the military—by more thoroughly embedding its personnel across the Department of Defense, through service partnerships, and by deepening outreach to the tech community. It should stay focused on the country’s most strategic needs and be put in a position to dismantle remaining barriers between the Pentagon and the commercial sector. Congress should also increase its flexible budget. DIU should remain small, but it should be empowered to serve as the vanguard of change.

But taking risks and bringing in cutting-edge tech is the job of the entire department, not just one unit. And without revamping the way the Pentagon’s money works, current reforms designed to move the department’s organizing construct for procurement from traditional programs to more flexible portfolios won’t matter, and the innovative rewiring underway across the military will not have the effect it must. Congress should thus carve out a portion of the budget—say, 20 percent of today’s research and procurement budget—for each of the services and require it to be spent on technological solutions in areas such as autonomous systems, artificial intelligence, energy, biotech, space, cyber-technology, advanced manufacturing, and quantum computing. Within these portfolios of record, the services should have the freedom that today’s DIU has to switch from product to product, version to version, and from prototype to operational fielding without having to go back to Congress.

Alongside this new flexibility, the services should be required, just as DIU is, to provide detailed, real-time reporting that shows how they use this money and offer frequent updates, including when the news is bad, ensuring appropriate oversight for Congress and the American people. Many in the Pentagon believe such information sharing will be too risky or too hard. But project-tracking technology that is already standard in private industry makes this kind of reporting increasingly a governance standard, not a unique burden. Congress has already made clear that it will not grant the services the same flexibility DIU has enjoyed without the same transparency and mutual trust.

The department also desperately needs to reduce the cost and increase the speed of acquiring more traditional platforms, such as ships and planes. This problem might seem unrelated to drones and AI, but the commercial sector can assist here, too, by providing components, materials, and advanced manufacturing techniques that could help the military produce these systems at greater scale and lower cost. This is already happening with the F-35 fighter jet’s software, the navy’s use of the manufacturing firm Hadrian’s software for submarine production, and experimentation by several services with advanced manufacturing firms for cruise missiles and other platforms. Success on major platforms will require partnerships between nimble commercial companies and traditional contractors with the capacity to manage large programs. The Pentagon needs to use its muscle to facilitate that teamwork.

To take full advantage of new technology, American officials will also need to look beyond their borders. The best innovations typically draw on talent, hardware, and software from many places—mostly U.S. allies—and require global markets to compete against China’s manufacturing ecosystem. Washington needs its allies to have the best technology, both to be effective against common adversaries and to be interoperable with American forces. It needs allied tech companies to have opportunities in the U.S. market, and for U.S. tech companies to have opportunities in allied markets, so all can compete without relying on Chinese money or parts. Although pressure from U.S. President Donald Trump is finally unlocking increased allied defense spending, friction over issues such as his stated desire to annex Greenland, tariffs and trade imbalances, and the war in Iran is simultaneously undermining allied confidence in Washington. Decisions like the one in June to cut off allies’ access to Anthropic’s latest AI models have exacerbated concerns in many allied capitals about the risk of being overly dependent on U.S. technology. The United States has to focus on its own critical needs, but with the full understanding that true strength in technology and national security comes from partnerships, not from any one country alone.

A U.S. Marine Corps drone, Pocheon, South Korea, August 2026 Kim Hong-Ji / Pool / Reuters

Driving change and capitalizing on it will require the Defense Department to hire, develop, and retain far more dual-fluency talent. Despite its talk about bringing in staffers who understand the commercial world, the Pentagon has done a terrible job of actually doing so. This is largely because its hiring processes are infuriatingly slow. Although qualified, patriotic tech warriors will take a 90 percent pay cut to do important work for the country, it is hard to convince them to wait six to 18 months before starting their new jobs—the time it typically takes from first discussions to hiring. That does not even include the additional waits associated with security clearances for those who don’t already have them. Exceptions exist, but like so many things in today’s defense innovation world, they result from direct pressure from the top, often only for individual cases. Solving this will require a major overhaul of the way the Defense Department conducts hiring and thinks about talent. Highly bureaucratic processes reminiscent of those for procurement must be replaced with more efficient ones like those in the private sector. In addition, the Pentagon will need to create new job categories, compensation structures, incentives, and career paths to retain and develop dual-fluency uniformed and civilian personnel. Today, too many of these people see little opportunity for advancement in the military and leave to join the private sector’s defense tech boom. The vast majority of the changes needed both for acquiring new talent and retaining tech warriors are within the Pentagon’s control. Congress is eager to help with the rest.

Perhaps most important, the Pentagon must do more to unlock its teams’ latent potential. It must change performance incentives for acquisitions, contracting, and adoption personnel, rewarding appropriate risk-taking rather than punishing every failure. It is not enough for defense leaders to simply demand that Pentagon staffers change their behavior; they must change the underlying criteria for evaluation and promotion that govern officials’ careers. The effective balancing of risk and opportunity across workforces as they scale is at the heart of what has made Silicon Valley so successful, and it is achievable in government. Congress has a role to play, too, by balancing the censuring of officials whose risks don’t pay off with awards for those who take heroic gambles for the right reasons.

The tech sector’s founders and funders must change, as well. Defense technology is experiencing a gold rush, spurred by the combination of massive need, enormous potential, real steps toward government reform, growing defense budgets, and positive examples from an ever-larger number of successful companies. The sector spans both increasingly saturated areas—such as simple aerial drones—and nascent ones still wide open for new entrants, meaning some areas are ripe for consolidation and scale, whereas others remain fertile ground for startups. Programs like the Pentagon’s Replicator initiative, which funded over 100 companies, two-thirds of them small businesses, while turning away over 300, illustrate that fierce, Darwinian competition is both inevitable and healthy in this sector. Out-innovating China demands the Silicon Valley mindset of constant churn—launches, failures, acquisitions, and redeployment of capital and talent—rather than the traditional contractor reflex of taking their losses to Congress or the press.

THE RIGHT RISKS

Fundamentally, what the defense world needs is a sea change in its approach to risk. Officials in both the Defense Department and Congress, working with the private sector, must take appropriate financial, process, and reputational risks today to avoid taking more dangerous ones—risks to military missions, to the troops, and to the nation more broadly—later. Americans are very good at driving dramatic change in response to terrible shocks, such as Pearl Harbor, September 11, or the COVID-19 pandemic. This time, the United States should not wait until tragedy strikes.

Fixing the military’s procurement and adoption systems cannot, of course, guarantee that Washington will deter U.S. adversaries or win a major conflict. But the current administration and Congress have a historic opportunity to work with an eager tech sector in support of the best warriors in the world, to bring real change right now, and to make the United States and the world much safer in doing so. No matter where individual Americans stand on the political spectrum, and no matter how different their views on domestic or foreign policy choices, none can afford for the nation to miss that chance.

Loading...