Emerging Themes in Defense Tech: Investing in the Cost Curve
16 February 2023
By Arthur Karell, Partner at First In
We have seen a rush of investor interest in emerging defense technology companies going into 2023. That is no surprise. Publicly-traded defense industry stocks began to outperform the broader S&P index towards the end of 2021, as investors sought greater exposure to counter-cyclical industrials. War in Ukraine and commitments by Western allies for increased military spending added momentum to defense industry equities in a slowing economy. Stocks of Raytheon and Lockheed Martin were up by 40-60% in 2022, while the rest of the index was down 19% on the year. Investors seeking bargains also brought that bid to a wide variety of early-stage companies selling into the defense complex.
Private investment in technologies supporting national security is always a welcome development. From an investor perspective, however, risk/reward is becoming harder to discern. It’s certainly a more crowded market, but more importantly for those early-stage investors anticipating near-future spending and skating to the puck, the puck just got slapped across the ice in 2022.
While the military lessons being drawn from Russia’s invasion of Ukraine may reassure some that the U.S. is years ahead of competitors in technical capacity and training proficiency, it has also laid bare systemic underinvestment in the large masses of attritable systems required for any direct great power kinetic conflict lasting longer than a couple of weeks, including precision guided munitions, short-/medium-/and long-range air defenses, autonomous air, ground, and maritime vehicles, and even LEO space-based capabilities (which are vulnerable to interception).
As such, we see a renewed focus from defense customers on bending the cost curve of weapon systems, in order to be able to procure enough of them in sufficient numbers to be relevant in a near-peer fight. This is a central theme in First In’s approach to the defense market. We invest in companies that aim to provide existing or emerging defense capabilities at a fraction of the cost – driving better pricing for the customers (and taxpayers), along with margin advantages, across our defense technology portfolio. Two promising areas of innovation in particular are industrial automation of procurement priorities and products that drive down the cost of deploying computing capabilities across the enterprise.
Industrial automation reduces the expenditures necessary to achieve production of defense (or dual-use) articles, especially those that must be near-shored or onshored. It encompasses software and hardware enabling the design and manufacture of end-use weapon systems and platforms/vehicles, along with critical components like computer processors, power units, and communications systems. In either case of end-use items or components, industrial automation also enables the rapid ability to test continuously, which is a sine qua non of successful defense product development.
Not all production automation is relevant to our defense investing thesis; we only consider the industrial automation of product or systems categories that are priorities for defense customers. It’s not difficult, however, to determine what technologies are procurement priorities: they are listed as mission area categories, highlighted in wargame studies and Congressional testimony, and of course, discussed in requests for research and development proposals.
The cost curve heuristic can also be applied to the proliferation of computing across the defense enterprise and its distributed networks. We’ve written about distributed computing from the perspective of great power competition, and the urgency is equally clear through a procurement lens. The U.S. Department of Defense (DoD) is growing its total compute with every new cloud IT migration, weapon system, sustainment platform, and existing infrastructure retrofit; how quickly, cheaply, and effectively that compute can be provided is a major driver of vendor profitability. It’s no longer enough to point to the broad, double-digit CAGRs of cloud adoption across the U.S. Government (USG); agencies are now more likely to require tracking of (cloud-based) software deployment costs and savings.
Several aspects of this computing cost curve opportunity are of interest, including those employing artificial intelligence (AI) to achieve order-of-magnitude improvements in delivering and securing software products for defense. Just as one would expect emerging natural language processing, machine learning, and computer vision models, tools, and libraries to improve software development margins in the commercial sphere, so too will they increase the reach of early-stage defense ventures. We expect this trend to be especially true for cybersecurity at the cloud and infrastructure layers, and it will also require deployment of AI-based defenses against adversarial AIs.
Now that public market aerospace and defense equities are likely fully priced, the comps for privately-held defense technology companies are coming back to Earth. A cost-curve approach to early-stage investing is just as applicable, if not moreso, to a market that is facing top-line headwinds. Indeed, there is a small but material chance that Congressional Republican infighting in the near-term, and entitlement spending and debt service in the medium and long-term, may lead to the 2023 NDAA being a high-water mark for defense spending.
Whether or not the defense market maintains its 2022 momentum, investors cannot simply rely on a generalist approach of riding a rising tide, given the specialized needs of defense customers and the idiosyncratic nature of government contracting. First In is a team of former military and startup operators who are well acquainted with the interplay of policy, regulations, products, and market dynamics that constitute the inner workings of the defense industry. In an era of renewed geopolitical competition, we are eager to invest in early stage ventures that address core defense customer concerns of capability and cost.