5 Things To Know About National Security Deeptech for Investors and Prospective Employees

Just like recruiting CIA assets, it's hard - but not impossible - to diligence deals and employment opportunities.

My time recruiting CIA assets, working with two deeptech companies, and helping raise $100 million+ in startup funding provides some perspective I’d like to share for those doing diligence on national security deeptech. If you’re looking to put your or your LPs’ money into a company, or looking to dedicate your time to as an operator/employee, here are some things to consider:

1. If you have an in-house physicist, don't use them for most of your technical diligence. Sounds crazy, right? The field of physics is vast; PhD expertise in one area does not confer knowledge in other areas. A PhD in physics isn't a panacea. A PhD in atomic, molecular, and optical (AMO) physics, for example, doesn't mean you know all the sub-specialities in that field. Further, all of us have to guard against our human nature of being deeply wedded to past decisions and future justifications of them. Bias can come from where one studied and who one studied under. This is amplified when on diligence calls one feels the need to disprove the other. Technologies can be dismissed out of arrogance rather than understanding.

Yet you won't know that unless you have an incredibly humble physicist on staff. Same goes for an employee without the specific technical expertise evaluating which company to join.

Instead: Find an academic, national lab, military service lab, MITRE, RAND, or Aerospace Company technical expert in the same exact technology and ask for 30 minutes of their time. They'll often meet with you - you simply have to find them on LinkedIn or find their email and ask. But don't make the mistake of telling a government-employed expert that you want to talk to them specifically as part of your diligence of company X, Y, or Z - it triggers a fear of a conflict of interest and they'll pass on the call.

Say you are an investor interested in learning more about the tech and its applications because you want to increase your investments in the national security space and give them a preview of some basic questions you'd like to ask. If you know what questions to ask and how to ask them (I’ll write on that more later), they will outline for you all the ways the tech may not work and you'll walk away with a solid list of questions to ask the company. They'll also give you companies verbally that you should compare.

Also ask them to give you the names and contact information of two other experts who agree and disagree with the tech approach.

If the company you’re diligencing won’t give you a couple of these references out the gate, pass. It should go without saying that you should go beyond the referenced contacts. Then find & offer to pay an academic third party expert to join you on further technical diligence calls so you can watch the exchange and interject with your own questions. Don't invest in a deeptech company without getting a third party expert involved. It's cheaper than you think.

2. Don’t fund a private version of a national lab. If you’re looking to join as an employee, know what you’re getting into. Are you joining a research environment or a company-building environment? They are very different.

Government funds basic research and industry funds products. Nobody funds the “in-between” except for those who are “venture” investors in the truest sense of the word. Perilous risk with phenomenal upside (as our Twitter feeds full of $NVDA news remind us) isn't for the faint of heart. The "in-between" is where the glory and returns are, yet the "in-between companies" aren't all equally as risky.

If a deck shows what is essentially a science project requiring 5-10 years (or more) of capital intensive investment before any revenue, with scientific breakthroughs scheduled on their “roadmap,” know you're wading into the deep end of risk even for deeptech. Gov’t grants or small gov’t R&D contracts they expect to win in the 5-10 year time period don’t count towards this idea of “revenue" though government "programs of record" do (more on this later). Be that as it may, R&D contracts are important as a start - if a company post seed claims to be “dual-use” but hasn’t achieved one of these yet or can’t speak to exactly how they're getting one soon, pass.

Instead: Find companies with revenue off-ramps along the way to their ultimate technology goal and market that amplifies the core tech rather than distracts away from it. Easier said than done, I know. But they exist and I’ll write more on hard questions to dig at on this in the future.

Ask and distinguish early between pure R&D money coming into the company and sales of devices/products/services to government customers if government is the pathway a company claims will lead them to commercial markets.

3. Scientific breakthroughs can’t be scheduled via slideware roadmaps. (I know, I know, "thanks Captain Obvious"). But it's a bit shocking how easy it is to overlook this and justify it because you like the founder or the concept. If the company needs a scientific breakthrough to make any money, unless you're really on the bleeding edge of deep tech investing and some of these bets are foundational to your strategy, see #2. This is hard because the initial pitch may lead you to believe “it’s just an engineering challenge at this point.” Most of all, watch out for any company that needs to overcome a material science challenge. (If a company has to create/discover a new material in order to provide value, that is what we call a material science challenge folks, not an engineering one!)

Instead: Look for companies that clearly distinguish between a science challenge and an engineering challenge and back the engineering heavy ones. It's not to say that you should never fund any science. After all, evergreen companies develop intense market leverage, but the science part shouldn't all be upfront in many national security tech cases (though biotech/pharma plays are outside this framework). This is the classic "exploration vs exploitation" model. The heavier the science, the more important #4 is below.

4. There are "engineering challenges" and then there are engineering challenges. If a company doesn't point out exactly where they are on the science>engineering spectrum, acknowledge where the hardest challenges are, and demonstrate how they obsess over solving these challenges daily, pass.

Instead: Look for companies that say the hard part out loud without being asked. More to follow in future posts on questions to pose and follow-up you can do in deeper diligence to drive to the answers you need.

5. Don’t pass on an entire industry because it’s "too nascent.” This sounds counter to #3 and #4, but it's not. Beware of conflating one piece of industry for the entire industry as Lux Capital did recently in their shareholder letter with quantum and fusion.

Most people hear "quantum" and the mind immediately jumps to “quantum computing,” which is admittedly further down the road for real-world value despite the hyped headlines and alluring McKinsey reports. Yet, quantum computing itself is one of the highest optionality bet you could make in computing if a quantum company also has other revenue off-ramps that don’t district, but rather build towards, this ultimate end goal.

Said more precisely in the quantum case - there's a lot more to quantum than quantum computing. There are optical atomic clocks, for example, well past the scientific research phase with near-term industry applications like data throughput speed-ups for data centers in addition to government assured Position, Navigation, and Timing (PNT).

If AI is going to scale, we’re going to need to rely on Sam Altman and $7T for chip fabs, (and Chinese money apparently), or we're going to have to find new ways to increase data density as we also scale domestic and "friend-shored" production. In the quantum clock case, it’s a matter of integrating photonic integrated circuits (PICs) to make clocks smaller and cheaper through manufacturing cost curves to expand sequentially to reach massive markets. This is an example of a non-trivial engineering challenge and a supply chain challenge.

Instead: Make room for the cognitive dissonance of holding two opposing thoughts in your head at once. We rely on mental models and shortcuts. Categorizing for a quick "no" makes sense. But beware of throwing the baby out with the bathwater. A good way to avoid this is to start with industries you believe are "too nascent" and do some research to find the POCs for those associated portfolios at MITRE, RAND, any of the military service labs (AFRL, ARL, etc), and ask them what they think are aspects of the market not typically captured in the news headlines or understood by most. They'll tell you. If you want to find hidden markets, much like CIA sources, they don't always come to you (even if you're a big name) and they're not always obvious at first.

Final thoughts: For those of you just starting to catch the wave in deeptech national security opportunities, you're not too late. Welcome, but take caution in jumping onto a bandwagon you know little to nothing about. At least buckle up. The road to national security deeptech is fraught with challenges and few outside these specific industry niches understand them. You'll never fully know what to ask & how to evaluate a company technically, but there are ways to leverage others for this and mitigate risk for massive upside. Much depends on your risk appetite and the company stage. Financial return remains paramount. But a sense of mission matters, too, and if you don’t have that, then it’s probably better for you to work on something easier and less consequential like building another food delivery app.