Will build turn to buy? Consolidation, capabilities in the defense-tech autonomy landscape.

Markets little moved as rates and geopolitics dominated. The S&P and Nasdaq started the week out strong on Monday, but pared gains as higher interest rates around the world created jitters around growth and funding costs for the AI buildout. The 10-year Treasury reached its highest level since 2007, and continued instability in the Middle East compounded inflationary and budgetary concerns. The S&P ended the week flat and the Nasdaq was up 1%. Our NPM Private Market Tracker*, which shows the average price performance of the 50 largest names in our internal NPM Price data, is up ~88% YTD versus the S&P 500/Nasdaq up ~12%/~16%. (Bloomberg, Wall Street Journal, NPM)

The field feels crowded

There has been a ballooning number of private defense tech companies working on mass-produced autonomous vehicles — drones for the air, unmanned surface and underwater vessels for the sea, autonomous ground vehicles for terrestrial operations, and the software that runs them all. For the most part, the startups have stayed within their own domains – as examples, Shield AI is building autonomous drones, Saronic is building autonomous maritime vessels, Overland AI is building autonomous ground vehicles, and each of them has their own in-house autonomy software. As the largest private defense tech company, Anduril stands out as having the most cross-domain capability.

Public defense companies, including the primes (Lockheed Martin, RTX, Boeing, Northrop Grumman and General Dynamics) and others such as L3Harris, Textron, Huntington Ingalls, BAE, Kratos and AeroVironment are also developing (or have developed) autonomous hardware and software. In short, the field feels crowded.

Who is doing what in autonomy now

The chart below shows which companies are competing in which areas of autonomy at the moment. One key takeaway is that there are numerous drone and autonomous maritime start-ups, but only two focused on land-based autonomy. There is also only one company focused only on software with no hardware offering (Applied Intuition is primarily a commercial autonomy company, with partnerships with companies like Deere). Another takeaway is that the larger legacy defense companies tend to have broader footprints, but they may still have gaps in their capabilities or benefit from the complementary technology of a startup. Note that we focus on US companies here; we include BAE Systems because its US subsidiary works extensively with the US military, and Ukrainian company UForce because it now has a US subsidiary and is top-of-mind in the sector. Also, this is not an exhaustive list of companies involved, it’s meant to be an indicative subset.

Momentum favors staying independent…until it doesn’t

A growing number of defense tech companies working on autonomy makes sense at this point in the cycle given the momentum in the sector – valuations are increasing and capital is available, so many management teams want to go it alone. Instead, companies that have wanted to fill strategic gaps or leverage another’s technology have entered into partnerships. Some examples are (though there are many more):

Anduril and Overland AI: Overland will integrate its ground autonomy software with Anduril’s Lattice software to enable coordinated air/ground counter-drone capabilities (company press release, 10/10/25).

BAE and Forterra: BAE will integrate Forterra’s autonomy on its Armored Multi-Purpose Vehicle (Breaking Defense, 9/30/2025).

General Dynamics and Kodiak: GD will use Kodiak’s autonomy software on its Leonidas Autonomous ground Vehicle (Yahoo Finance, 5/7/2026). Kodiak has historically focused on commercial truck autonomy in Texas’ Permian Basin.

L3Harris and Shield AI: L3Harris integrated Shield’s Hivemind software onto its Green Wolf uncrewed aircraft (company press release, 7/16/2026).

Leidos and Havoc AI: Leidos will integrate Havoc’s collaborative autonomy software into its sea Archer USV (company press release, 4/20/2026).

Still, the current level of fragmentation and competition made us start thinking about M&A. We think consolidation could take a variety of strategic forms; it could be within verticals (i.e. consolidation within maritime), cross-domain (i.e. an air-focused autonomy company acquiring a maritime autonomy company to leverage its software system), hardware-software, or larger public companies acquiring smaller private companies to augment their portfolios. We see several potential drivers of M&A that could materialize over the next few years.

Industry demand downturn. One typical driver of M&A is an industry shock, as this can create willing sellers out of companies who do not have the wherewithal to survive and willing buyers out of well-capitalized companies that see attractive valuations for their targets (Journal of Financial Economics, DealRoom). In defense tech today, we think demand downturns could stem from the resolution of some key conflicts globally (this would not reduce the need for stockpile replenishment or new, less-exquisite technology, but it could take some of the steam off of the sector), or the failure of appropriations to materialize under the Trump administration’s $1.5 trillion FY2027 defense budget request.

Industry fragmentation. There might also be a supply problem in the form of too many companies with similar offerings that drives some companies to sell their product portfolios because they are not winning enough business. Consolidation tends to happen in fragmented markets once participants see value in combined buying power and other synergies, and for defense specifically, in achieving the scale needed to compete for large DoD contracts.

Revenue gaps. In addition to filling capability gaps in their portfolios, it is possible that some of the larger private defense tech companies make acquisitions in order to add revenue. Private defense tech companies are raising at revenue multiples that are higher than legacy players based on future expectations (over 10x versus legacies well below 10x, Source: The Merge), and for the moment the prospects for revenue growth are keeping investors comfortable with this. It is possible that, if the sector nears a turning point, some companies may look to acquire smaller peers who are generating revenue today in order to maintain valuation momentum.

Market shock. A broader economic shock – perhaps one spurred by higher global borrowing costs – could make venture capital and debt for building out manufacturing facilities harder to come by. Companies facing a down round or a failed raise could choose a sale over trying to survive independently.

IPO window closes. This year’s space/defense tech IPOs have had a mixed performance, with Hawkeye360, York Space and Aevex all trading below their IPO prices (Bloomberg). Drone manufacturer Swarmer is trading above its IPO price, but it is a microcap that raised <$20 million and we do not believe it is reflective of the market’s broader receptivity. As we note in this week’s news below, several energy-related companies have delayed their IPOs given market uncertainty, and while this is not defense tech, a softening of IPO prospects could lead companies to opt for strategic sales as an exit rather than an IPO.

Technology cycle compression. The new mantra in defense tech is cheaper, mass-produced attritable technology over exquisite systems. Exquisite systems lend themselves to building versus buying because they are complex and huge in terms of both time and dollars. The need to compete with novel technology quickly in this modern defense tech paradigm could lead some companies to buy over build. In fact, year-to-date M&A capital in Aerospace and Defense has flowed toward capability, capacity and execution certainty rather than legacy scale (Source: PwC US Deals 2026 midyear outlook).

Scale up while the music is playing. Anduril (most recently valued at $61 billion) has made several acquisitions in order to fill some gaps and cement itself as the large private neo-prime. For example, it’s acquisition of Blue Force Technologies in 2023 gave it the basis for its Fury unmanned fighter jet; it also acquired Klas (2025) to aid its tactical edge computing in GPS-denied environments and American Infrared Solutions (2025) which manufactures infrared cameras and optical components (Source: The Merge). Anduril could remain acquisitive, though it already has broader capabilities than other private companies and is growing revenue organically.

We think an interesting question is what companies in the ~$10-15 billion valuation range (the next tier down from Anduril) might do. Drone/autonomy company Shield AI (most recently raised a $1.5 billion Series G, or $2 billion including debt, at a $12.7 billion post-money valuation) has already grown through acquisitions – its Series G was used to fund the acquisition of Aechelon, a defense simulation company, and prior to that it acquired Martin UAV (which is now its V-BAT platform) in 2021. Maritime autonomy company Saronic (most recently raised a $1.75 billion Series D at a $9.25 billion post-money valuation) bought the Gulf Craft shipyard in Louisiana in 2025. These companies are not short of capital as proven by their recent rounds (and Shield has been reported by The Information as seeking to raise a next round at a ~$20 billion valuation). These companies could look to use their current capital and cache to make additional acquisitions and cement their size/capability moat ahead of a potential industry shakeout.

Overall, M&A is not happening at scale now, but we believe the sector seems ripe for it with the right impetus.

Based on our proprietary NPM Price data, the best performers of the large cap names in the private market thus far in 2026 have been:

Top performers continue to be driven by up-rounds that rerack their valuations. For example, Castelion closed its $1 billion Series C in August at a $13 billion valuation, up from its previous valuation of just under $3 billion. We think it’s notable that three space/defense names have now entered our top performer list – Hadrian, Castelion and Stoke Space.

Based on our proprietary NPM Price data using weighted averages for each sector, the top performing sectors YTD taking into account the top 100 largest names are AI and Machine Learning (127%), Consumer (+104%), Healthcare/Life Sciences (+68%) and Industrials (+61%). The fact that many of the top performers are in the AI and Machine learning category helps explain the sector’s performance.

1. Amazon blocked Meta’s Muse AI agent from accessing its shopping site (The Information, 9/22), though Muse surpassed 500,000 users after its first week (The Information, 9/23).
2. Kalshi is being investigated by federal regulators after nearly one million trades in a single ether market were placed in nearly identical amounts (Wall Street Journal, 9/22).
3. New York sued Polymarket for allegedly running an illegal gambling operation (The Information, 9/24).
4. Google’s Gemini model unexpectedly breached the networks of three outside companies during safety evaluations (Wall Street Journal, 9/18). Meanwhile, an agent developed by OpenAI hacked an Australian government website in June (Bloomberg, 9/24).
5. CoreWeave priced a $4.2 billion 7-year convertible bond at a 2.875% coupon; this compares with a 1.75% coupon on a 6-year convertible bond it sold in April (The Information, 9/22).
6. OpenAI expects to burn $278 billion by the end of 2030, according to a Financial Times Report (The Information, 9/21).

1. Nscale, a London-based AI cloud provider, filed to go public on the NYSE (StrictlyVC, 9/18).
2. Data center operator Firmus is seeking to raise $5 billion in its Australia IPO (Pitchbook, 9/23).
3. Accelevation, an Ohio provider of power distribution and other data center infrastructure, set IPO terms that imply a market cap of $44.9 billion at the midpoint (Axios, 9/24).
4. Meanwhile, SB Energy, a data center platform backed by SoftBank, is delaying its IPO (StrictlyVC, 9/21), though Nvidia is buying $1.5 billion of additional shares in SB Energy ahead of its IPO (Pitchbook, 9/23).
5. Nuclear reactor manufacturer Holtec delayed its IPO indefinitely citing “uncertainty of data center development” and Aggreko, which provides energy and temperature solutions to industries including data centers, has slowed down its process as well (New York Times, 9/21).
6. Westinghouse Electric, which makes large nuclear reactors, is seeking a valuation of over $50 billion for its IPO (Bloomberg, 9/18).
7. London-based fintech Revolut is exploring a dual listing in London and New York (Pitchbook, 9/21).
8. Oura officially launched its IPO and is looking to raise ~$2.2 billion (TechCrunch, 9/21).