Shyam Sankar speaks at the AI Expo For National Competitiveness on May 7, 2024, in Washington, D.C.

Shyam Sankar speaks at the AI Expo For National Competitiveness on May 7, 2024, in Washington, D.C. Tasos Katopodis / Getty Images for Palantir

Eighteen ways Palantir wants the Pentagon to change

A conversation with the CTO of the self-described software prime contractor.

The Defense Department is a sclerotic monopsony whose communist approach to acquisition has the United States on a precipice, writes Palantir’s chief technology officer, who prescribes a “painful” but “necessary” reformation based on competition and software.

“I think we're just scratching the surface,” said Shyam Sankar, whose company calls itself the first software prime contractor. “The other huge opportunity is really on using AI to drive efficiencies,” including using technology to supplant human workers and processes that slow innovation and adoption. 

Sankar recently penned 18 theses that could reform how the Pentagon does business, in explicit comparison to Martin Luther’s criticisms of the Catholic Church. He highlights several well-documented problems often studied by congressionally mandated commissions—such as a belabored budgeting-and-planning process, the perils of requirements, and how cost-plus contracts remove incentives to innovate. 

The 18-page document also manages to say the quiet part out loud, demanding the Pentagon and Congress to change its practices with colorful curtness. One of Sankar’s main themes is the need for competition—whether it’s companies vying for Pentagon contracts or customers in the Defense Department looking for solutions. 

He also criticizes a system whose major suppliers sell only to the Pentagon and other militaries.

“Chinese primes only earn 30-40% of their revenue from the PLA; the remainder is commercial. Those cheap products your neighbor is buying on Amazon are subsidizing lethality which could be used against our men and women in uniform, much the same way that during the Cold War your purchase of an American car, camera, and cereal subsidized America’s lethality against her enemies,” he wrote.

Instead, Sankar writes, the Pentagon should vastly accelerate its purchases from largely commercial companies—like SpaceX, whose CEO Elon Musk has been tapped to lead a Trump administration government-efficiency drive. 

In September, Palantir’s stock joined the S&P 500, increasing  software and tech companies’ share of the index. In recent weeks, Palantir’s market value of $137 billion edged out Lockheed Martin’s.

Defense One spoke with Sankar. 

What are Palantir’s goals for 2025, especially amid a new administration?

Our government business has re-accelerated dramatically. We're pretty excited about the product road maps that we've built. [Maven Smart System has] become the [Joint All Domain Command and Control] MVP. It's getting a huge amount of adoption, so we're pretty excited about that, and continue to invest in that.

What defense areas will you expand into?

I think we're just scratching the surface. So I mean, a big focus, obviously, Maven is at the pointy end of integrating AI into the kill chain. So how do you do that across all the services? It's coming down from OSD right now. The Army has started to adopt it…and the opportunities for that to happen with the [maritime operations centers] with the Navy and with the [air operations centers] with the Air Force, that are definitely top of mind. [We’re] continuing to push into space…we're at the very beginning of that. 

And then the other huge opportunity is really on using AI to drive efficiencies. There's more civilians in the DOD than there are uniformed service members. There are more acquisition professionals than there are Marines. And so have we just crushed ourselves with process? And I'm sure most of that process is needed, but can we use AI agents instead of humans to drive efficiency and speed through that? 

We've been able to automate the foreign-disclosure process, where it used to take humans three days to collect intelligence and then rewrite it to be releasable to partners, taking that down to three minutes. So the obvious benefit there is lethality and being able to share with allies and partners. But I think the other part of it is it just makes for a much leaner force.

If you look at [Georgetown University’s Center for Security and Emerging Technology] report on Maven. With Maven, you can perform targeting operations with 20 people that used to take us 2,000 people in Iraq. So that sort of efficiency really drives the ability to hide in a pure conflict. You're much smaller, [there’s a] much smaller footprint, less support infrastructure, and [that] drives lethality.

Would shrinking the Pentagon workforce, particularly on the civilian side, actually drive efficiency?

It does in the U.S. commercial market every single day. We automated [an] insurance underwriting process where one of America's largest insurers, with 78 AI agents, took a process that used to take three weeks down to hours. We have this challenge in government, generally—it's not just the Department of Defense—where we view people as abundant, and technology is unaffordable and expensive. But that's the exact opposite of how the commercial world looks at technology. They view people as very expensive, and technology as cheap as the way of achieving efficiency and effectiveness. 

I think a lot of that is because of the monopsony. That's what really speaks to the heart of the “Defense Reformation.” Like, why do we have the system we have? Well, I don't envy the position of being the department; being the sole buyer for a thing is very hard. At some very fundamental level, you believe in free markets or you don't. And what we have is a centrally, unplanned process. So it's like the worst of communism, really: a central planning process is not actually planned. We have a [Future Years Defense Program] that takes five years, two years to do [program objection memorandum]—you can't really change your mind anywhere along there. You're gonna have unitary efforts: we have one F-35 for three services. I'm not sure we should have 30 years on $2 trillion

When we were building submarine-launched ballistic missiles under [Vice Adm. William] Raborn in the [1960s], we had four competing programs running concurrently. Today, we look at that and we say, ‘Oh, isn't that wasteful?’ But actually, CBO zone analysis shows that it was cheaper and faster because we had competition. And we've gotten too fixated, since the end of the Cold War, on how much competition is there in the industrial base. That's not actually the problem. The problem is there's not enough competition inside of government. 

Will Congress go along?

Absolutely. I mean, I don't want to be too Pollyannaish about it, but I think Congress recognizes it’s part of the problem. I think what makes it really hard is they do have an actual oversight function to perform. So how do you balance the need for oversight with the necessary maneuver room that you're going to need when you're doing something innovative? That's just reality. 

If I think about all the projects in the company that I'm overseeing, I have to give folks the room to pivot, to change, to move. I think some of the recommendations coming out of the PPBE Commission are exactly right, like a capability element approach. Rather than a program-element approach, we're going to resource counter-UAS as a category and you're going to have the freedom and flexibility to move and adapt within that. 

If making a change, if changing your mind, if reacting to new data is major brain damage and requires things to go up and down the building up and down the Capitol Hill, you're not going to make changes, which means you're just going to accumulate error. Because the only thing you know for sure is that whatever plan you came up with yesterday is wrong today. The question is: can you fix it every single day as you observe that error, or can you just let all that error accumulate? And our five-year FYDP, our two-year POM—how we oversee these things is not fit for the moment.

What do you most want the Pentagon and Congress to hear?

The first one is that monopsony is the problem. You can look at your industrial base as a monopsonist. You get the industrial base you deserve. It really is all your fault. So you have to change your behaviors, what you're incentivizing and how you're organizing yourself to get what you need. Stop blaming the prime, stop thinking like some process is going to fix this. We need to approximate market forces as much as possible. “We need more competition inside of government” would be the pithy way of saying it.

The second one is that cost-plus makes us dumber, poorer, and slower. The only two industries that do cost-plus are general contractors on remodels and the government. I don't know if you've done a remodel before, but are you happy with how that went and the price predictability and the timelines and stuff? It's just not the way to do these things. You've already lost by thinking about requirements to begin with. Share the problems with a competitive American market and let them come to you with solutions and then buy the best ones you like. Buy multiple if you want, but you're going to have to rely on the American industrial base, not the defense industrial base to provide compounding price performance. And you see that with the SpaceX example. It used to cost $50,000 per kilogram to get to orbit with the shuttle program. With Starship Heavy reuse, it's $10 to $20 per kilogram—not missing any zeros there. You can't get that with cost-plus, you can only get that because Elon [Musk] cares about getting to Mars. 

But ​​the DoD is kind of a unique customer. It has specific needs, like lethality and security.

But that's really a response from our present world. Before the fall of the Berlin Wall, only 6 percent of [Pentagon spending on major weapons] went to defense primes. The rest went to dual-purpose companies. Chrysler built cars and missiles. Ford built satellites. Until 1990, General Mills, the cereal company, built torpedoes and artillery. So we had a very different American economy, where people were invested not only in economic prosperity, but also in freedom. And this is important, because commercial innovation drives price-performance improvements. You just can't see anywhere else. If you work at a car company in America and you're not able to decrease price for 4% a year, you're gonna get fired, you know. iIn the industrial base and the defense industrial base, if prices aren't increasing all the time, people would be shocked. So, like, the fundamental mechanisms are missing. And I think the reason we have what we have is that when we lost competition; when the Berlin Wall fell down, we didn't have a pure competitor anymore. The Soviets were basically done, and we went from 6 percent spending on unique defense primes to 86 percent today. 

It's not that I'm saying the products are dual-use, like missiles are single-use, but the companies are dual-purpose. I built the Operation Warp Speed supply chain for the U.S. government in two weeks because two years earlier, I had solved a structurally similar problem for BP and oil and gas production.

Anything else?

The most important thing is: “How did we get here?” We are in this state of undeclared emergency. We've lost deterrence as the West in the world, we've spent trillions of dollars. And how did we get here? 

My argument to you is really this: we went from spending 94% of our money on dual-purpose companies to creating a unique class of defense specialists that we put on the Galapagos Islands, and we created a great schism between commercial innovation and defense. And that has made our country much less safe, much less innovative. It's created many of these problems, and we did that because of the nature of monopsony. And the answer is to fix that. That's where it starts.