Shyam Sankar has had a busy couple of years. As the chief technology officer and executive VP of Palantir Technologies ($PLTR), he has helped drive the company’s meteoric rise to the top of the defense industry (and the stock market). Now, he’s set his sights even higher: he’s calling for a total Defense Reformation. Simple stuff, really.
Sankar drew inspo from Martin Luther when he published 18 theses of his own in December outlining how he would rethink the defense industry. The United States, he said, is “in an undeclared state of emergency” and has failed to deter its most dangerous adversary: China.
Tectonic sat down with Shyam to talk about the Defense Reformation, the success of Palantir, and why he prefers a First Breakfast over a Last Supper (the secret Pentagon dinner in 1993 that accelerated consolidation in the defense industry). Plus, he digs into what a second Trump presidency might mean for defense innovation.
Note: This interview has been edited for length and clarity.
Tectonic: What made you write the 18 Theses for the Defense Reformation?
Shyam: I’ve spent the better part of 19 years kind of crawling across the Valley of Death and playing some tiny role in bringing innovation to defense. The 18 Theses are my diagnosis of what has made defense innovation particularly challenging in the present moment, what used to make it work in the past, and what we can actually learn from that past. I wrote it right now because I think we are in a state of undeclared emergency. The West has lost deterrence. We had a pogrom in Israel. We have the militarization of the South China Sea. We have unprecedented phase-zero gray-zone operations. North Korea is only getting worse and more volatile. We have North Korean soldiers dying by the thousands in Ukraine. One day we may look back in disbelief and say, “How did we not realize World War III had already started?” I think we are in the final hours of preventing a broader conflict from emerging, and we have to get serious about doing that. And I think the 18 Theses lay out part of the roadmap to getting serious.
I want to dig into your diagnosis a bit more. Why do you think the US defense innovation system has stagnated?
The root of what ails us is monopsony. That is what the Defense Department is: a single-buyer market. At some very fundamental level, you either believe in the free market or you don’t. Everyone, including the Chinese and the Russians, has given up on communism, except for Cuba and the DoD. We persist with these five-year centralized plans. We think we can go through a centralized planning process that’s going to deliver defense innovation. But if we really take stock of the things that we’ve done in the past that are truly innovative, we didn’t take this approach. We actually had multiple competing efforts. Innovation is chaotic and messy. There is no process for it. The second you try to create a process, or de-risk it, or make it more predictable, you’ve ruined it. You know, innovation is artistic.
There are two parts to innovation. One, it requires competition. Often, when people say competition, they think of competition amongst the industrial base. Yes, sure, let’s have that. But that’s actually missing the point. You need competition inside of the government in order to start to approximate market forces. When we were building the submarine-launched ballistic missile, Vice Admiral William Raborn had four competing programs inside the Navy to build it. Today, we would look at that with wild disbelief, as wasteful or duplicative. Why wouldn’t we just have one thing?
Well, the obvious answer is that there is so much fundamental engineering uncertainty. Back then, we wanted to compete multiple approaches off each other and see what was going to work. That made us better. We got there faster and cheaper. That shouldn’t be surprising, but somehow it is surprising. It shouldn’t be surprising because that’s the free market working, right?
That’s how we did intercontinental ballistic missiles (ICBMs). Today, we just think, “Oh, the Air Force built the Minuteman!” But when we were going through it, every service threw their hat in the ring. It wasn’t a monopoly that was given to the Air Force. The Army tried to do it, the Navy tried to do it, but the Air Force won. Edward Hall and Bernard Shriever won.
I think that the idea of Edward Hall really matters because the other part that we’ve kind of forgotten from the past is the importance of these individuals. We call it the Apollo program, but you might as well call it the Gene Kranz program. Behind all these efforts, there are really unique people who have a founder personality, whether they’re inside government or outside government. There is no way we could enjoy the asymmetric advantage we have today without the 30 years Hyman Rickover spent building the nuclear Navy.
There’s a reason we did Operation Paperclip, the effort to bring Nazi scientists to the US. I think we might struggle with that today, because, oh my gosh, how could we bring these horrible people over? But we were much more clear-eyed then about the sense of competition and the state of declared or undeclared emergency we were in. We understood that winning mattered and we understood that those people were irreplaceable. We needed the right people to solve these problems.
Now, I think the real consequence of the consolidation of the defense industry after the Last Supper is that it drove out the heretics. It drove the creative founders and the creative engineers from the industrial base into other parts of our economy. And that’s what we see now.
Of all of the nations in the world, we understand that there is something special about founders—there is a reason we call them the Founding Fathers. Part of the reinvigoration of our defense industrial base is reminding ourselves that it was actually an American industrial base. It wasn’t Northrop Grumman. It was Jack Northrop. It was Leroy Grumman. It was Glenn Martin, not Lockheed Martin.
At the present moment, we have those founders showing up again. That’s the other reason now is the moment to write the 18 Theses. We have the creative engineers, people who are working in the national interest, who can really do these things again.
What would it take for the current system—the DoD—to be overhauled?
It’s really a matter of will. You know, we have all the authorities. We don’t need any kind of congressional action. We don’t need a whole new way of doing things. You could start it by just saying, “Okay, we’re going to have more competing efforts.” We’re going to start doing that by thinking, “Where can I start to approximate market forces?” Well, if you think about the combatant commands, it’s a strange setup. The DoD has totally bifurcated supply and demand. Demand in DoD is really the combatant commanders, the warfighters, and the people who service the real-world events that happen. The supply side is the services that man, train, and equip. But there is very little that can happen on the demand side that meaningfully changes what is happening on the supply side.
This system was more or less survivable when cycle times for hardware platforms were very slow and long. We were in the Cold War. Of course, we were much faster then—we built the U2 in 13 months. It was definitely survivable after we’d won the Cold War and there was really no near competitor. But now we’ve had 30 years of sclerosis that has set in as a consequence of not having a peer adversary.
The first step is understanding that pace really matters. You know, one of the things that Dr. Karp, our founder, says is, “If you really want a project to succeed, take the timeline, cut it in half, and then cut it by 5x again.” We tend to do the opposite in the DoD. Bill Gates has a quote where he says we overestimate the amount of progress that can be made in two years and underestimate how different things will be in 10 years. In the DoD, it’s the opposite. We underestimate what we can get done if we focus on the next two years. And we overestimate what we can do over the next 10 years. We’re always thinking about the force of the future and the timelines are too long. It’s just unserious. If you start having multiple competing efforts and demand signals coming from combatant commanders, these things start playing to the natural incentives of what we’re really good at. Part of the will comes from embracing the fact that innovation is messy and chaotic, that there is no straight, clean line between here and there.
What about your experience at Palantir and “crawling across the valley of death” has taught you that this is necessary?
Literally everything about the experience. Part of it would be internal. As the CTO, I’ve got to build the future, the technology, right? How do we do that? I can tell you zero times it has been pleasant or predictable. It’s always been interpersonally rife with conflict. It’s always had an unpredictable path and unpredictable timeline. It’s been incredibly hard, even if you’ve done it 10 times. When I was younger and less experienced, I would think, “We must be doing this wrong.” Why is this so hard? Why is everyone so upset? Surely, the next time I do this, I can make this more predictable. Surely I can add process to this. And the more I did those things, the more I screwed it up. You begin to understand that this is the nature of the beast. And I think it takes a few cycles to realize that what you’re optimizing on is winning, not a pain-free process. We kind of conflate the two. The process ends up being an opioid, not medicine. I’m not saying that all opioids are bad, but it’s not a medicine.
How has Palantir’s technology helped US warfighters? How has it changed the nature of warfare?
One of the things I like saying is that data is not the new oil. It’s the new snake oil. Data is not inherently valuable; what really matters are decisions. That is a frame that John Boyd would recognize—it’s his OODA loop. Our software is geared toward helping operators, from sensor to shooter, go through their OODA loops faster and get into a continuous learning loop that allows them to maintain a better OODA Loop than their adversary.
50% of what we do is commercial, 50% of what we do is government. And those can seem like they’re completely unrelated things, but they’re not. In the commercial world, you call it a value chain, from the hand of your supplier to the hand of your customer. You’re trying to make a bunch of decisions, from procurement, to manufacturing, to production, to quality, to sales and operations planning. You’re trying to get better and better at those things, to win against your adversary or competition. In the government world, you call it a kill chain instead of a value chain, from sensor to shooter, but it’s literally the same thing, One of the things we’ve been very focused on with this concept that we’re calling “The First Breakfast,” as the antidote to the Last Supper, is how we enable America’s best companies to participate in defense in the same way that Chrysler used to build cars and missiles and Ford built satellites until 1990. While we may no longer be the best at mass production, America is the best at software by a yawning margin. You can’t field that many B21s between now and 2027—the Davidson window—but you can field a lot of software. Our David’s slingshot is going to be software. So, how do we leverage two decades of experience delivering modern American software into air-gapped environments, from secret to Special Access Programs (SAP), to enable other companies to do that? What we’re excited about with Warp Core is really creating an offering that enables American software and defense companies to come from unclassified, Secret, and Top Secret to Special Access Programs with their software, inherit our accreditations, get access to the mission data, and actually add value in the context of integrated workflows for the warfighters. Ideally, we are shrinking a process that would have taken two to three years to something that might take two or three weeks.
What is Palantir focused on right now?
The main effort is the work that we’ve been doing with the Chief Digital and Artificial Intelligence Office (CDAO) and Project Maven, bringing together and integrating the global warfighting combatant commands with the services. We’re playing our part in the realization of the vision of Joint All Domain Command and Control (JADC2).
Where is the US currently weakest? Where would you be focusing your efforts if you were in the Pentagon?
If I had a magic wand, what is the biggest missing national asset that I think we need? I think we need a digital twin that goes from the foxhole to the factory floor. How do we give our policymakers, our elected officials, our warfighters, a pane of glass to understand the end strength scenarios needed in the foxhole, on these dates, for these conditions in the world, and these threats that we’re concerned about? Given we need these things, how do we work backward and understand whether, for the capital we’ve allocated and the money we’ve appropriated, we are building the right things on the factory floor? Will they get there in time?
Look at some of our major weapon systems, like the Joint Air-to-Surface Standoff Missile (JASSM). How many have we ever made at peak in a year? I think the number is something like 500. How many years would it take us to make 10,000 of them? Too long, right? A lot of this stuff that we have is very exquisite, but if you have an eight-year lead time to build a Patriot, or a two-year lead time to build a Long Range Anti-Ship Missile (LRASM), something’s wrong. We need to really be thinking about production as part of the kill chain. We measure how long it takes us to close the kill chain, but we have to include the time it takes to produce, to resupply, the logistical chain, and the constraints around that because that’s the reality in a pure fight. I think we have found ourselves over-invested. We got confused. We thought that the stockpile was the deterrent when it was actually the ability to make the stockpile. What we can never give up is our ability to produce. We need that constant demand signal in the industrial base to force people to be more efficient, to get better, to get faster.
How do you think the coming US administration, the second Trump administration, will perform in terms of defense innovation?
I’m very optimistic. I think we have leaders who see the problem with clear eyes, people who understand that deterrence is the goal here and that you need to deter through strength. As someone who believes in the importance of founders, and who is himself not a founder, what I see in the coming administration is a lot of founder personalities. These people understand that there is no playbook. You really do need an entrepreneurial and innovative approach to get these things to work.
What is the biggest threat that the incoming administration is going to have to face in terms of defense readiness and defense innovation?
My advice to people is that you can either fix the system or you can win. They’re not the same thing. If you spend all your energy trying to fix the system, you’re going to get bogged down in trench warfare and you’re probably not going to make that much progress. We can look at someone like Bill Perry, who delivered stealth and GPS. He did that around the system. He had a choice. He could either fix the system or he could deliver stealth and GPS. He went around the Planning, Programming, Budgeting & Execution Process (PPBE) and just got those things done. I think there’s a lot to be said about picking three to five lines of effort and just getting it done.
Why are partnerships so important for non-traditional defense companies?
Who wants to be the only person winning on a sinking ship? We have to remember that the goal is to win as a nation. And what is that going to take? It’s going to take us working together just like it did during World War II. It is going to be a whole-of-nation effort. It’s going to require mobilization. It’s going to require voluntary civil-military fusion. No one company can do this alone.
Palantir, of course, is now more valuable than the defense primes. What does that say about the defense industry and the status of primes?
I think it says we’re back. Look, I’m one of these people who believes we need our primes. They do really exquisite things. They are real engineering companies and we should be, as a country, incredibly proud of them. We also need to help them evolve their business model, which is not really their fault, to one where they are remunerated for investing their own capital in research and development, rather than investing the taxpayers’ capital.
I think [Palantir’s valuation] is a wildly positive statement for the future of defense tech and America because it shows that the market is rewarding our model. You know, we were the first defense company in 46 years to join the S&P 500 without mergers or spinouts or any of those shenanigans. And we’re not going to be the last. That’s the exciting part, that it’s really helping pave the way for a new golden age in defense innovation.
So you think there will be others that will follow in Palantir’s path?
Absolutely.