By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
a16zCan the Military Move at Startup Speed? How the Army and Navy Are Rebuilding
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
- 0:00 – 4:18
Meet the Army and Navy CTOs
Steph Smith asks both guests to introduce themselves. Alex Miller describes his role as CTO for the Army's Chief of Staff and his earlier intelligence work, while Justin Fanelli describes the Navy and Marine Corps CTO role and its focus on scaling commercial innovation.
- •Alex Miller splits his time roughly sixty/forty between educating people and hands-on engineering
- •Miller says the Army's rules were written by enthusiasts but implemented by risk minimalists
- •Miller previously was science and technology advisor for the Army G-2, which runs intelligence
- •Fanelli says the government CTO role arrived about thirty years after the private sector
- •Fanelli's goal is outcomes and overmatch by scouting non-traditional partners and tying new buys to divestment
- 4:18 – 8:07
Pathways for Startups and the DoD Black Box
Steph Smith asks how startups can engage, given routes like SBIR, DIU and AFWERX. Alex Miller explains why the DoD is a black box and how the Army is trying to dismantle it. Justin Fanelli adds that more companies want to work on national security than ever before.
- •Startups focus on R&D to a minimally viable product, which is where DIU and the WERX groups fit
- •Miller says the black box dates from the Cold War-era rules, with everything since being duct tape
- •The Army wants industry help writing requirements and building consortiums using other transaction authorities
- •Fanelli calls this the best alignment of industry and national security in our lifetimes
- •Fanelli says Anduril, SpaceX and Palantir fought through the system, and PEOs are not a Horizon three company's first customer
- 8:07 – 11:04
What Programs of Record and PEOs Are, and the Horizons Model
Steph asks Justin to explain program offices and the path to a program of record. Fanelli walks through PEOs, the Navy's shift to portfolios, and how Horizon one, two and three capabilities each face different routes and funding.
- •There are 75 program executive offices, 18 of them in the Department of Navy, each with many program management offices
- •The Navy is converting programs of record into portfolios to buy capability per dollar, called military Moneyball
- •Horizon one products can go directly to PEOs with existing budgets, but only by replacing something
- •Horizon two products can use valley of death funds such as APIT, though there is not enough funding
- •Horizon three companies need guerrilla marketing and help from DIU, the Hill and requirements writers
- 11:04 – 15:44
Blowing Up Gold-Plated Requirements
Alex Miller argues that PEOs and programs of record are workarounds to a broken system and that requirements should be rethought. Both describe bloated requirements documents and examples of faster alternatives like capability need statements.
- •Miller notes PEOs are not in statute and the SecDef can reshape anything below law
- •Miller describes a counter-drone requirement whose statistics added up to a system that only had to work 51% of the time
- •Miller says programs of record create a perverse incentive to maintain and coast
- •Fanelli recalls a 4,000-page requirements document and says capability need statements cut an 18-month process to three months
- •Fanelli frames the goal as scaling what works, noting that these are still pockets of excellence
- 15:44 – 20:24
Making the Best Thing the Normal Thing
Prompted by Steph's comment on the pace of technology, Fanelli gives examples of enterprise reuse across services, and Miller argues that value and time to delivery should replace cost, schedule and performance. Miller also describes the Army's openness to fielding with units directly.
- •Horizontal capabilities like AI cannot sensibly be bought across 75 PEOs
- •A hyper-converged edge compute box let the Navy divest a billion-dollar purchase and deployed to the Red Sea
- •An electronic warfare solution was painted brown and put on a Humvee by the Army in three months for 300K
- •Miller says performance in requirements documents often does not reflect real war-fighting needs
- •Miller says the Army can modify its kit and offers startups a front door to Army Futures Command and units
- 20:24 – 27:03
Overmatch and Enterprise Services Over Vestigial Systems
Fanelli defines overmatch as order-of-magnitude leaps proven through A/B comparisons, and Miller gives the example of 17 mission command programs each with its own mapping stack. Steph compares the approach to venture capital, and Miller says companies should be assessed on what they are good at, like Lego pieces.
- •Fanelli wants power-law leaps rather than 15% improvements, using the Naval Identity Service as an example
- •A/B testing, such as a cloud simulation of 100,000 reps, translates breakthroughs into outcomes
- •Miller says 17 mission command programs of record each had their own map servers
- •The DoD is not building the best maps, data platforms or API gateways, so it should drop vestigial systems
- •Miller says vendors should be chosen for what they are best at rather than being forced to deliver everything
- 27:03 – 32:35
Enterprise Services, Common Concern, and Proximity to the Warfighter
The guests discuss breaking silos and sharing common services across the services while leaving truly unique problems to individual ones. Miller discusses LLM providers and unique data, and Fanelli advises startups to stay close to the warfighter and describes maverick risk-takers inside the system.
- •Miller says the enterprise should do enterprise things, using the intelligence community's services of common concern as a model
- •Miller says LLMs are all available, so what matters is the military's unique data and missions
- •Miller says this is a culture and process problem, not a technology problem
- •Fanelli tells startup CEOs to spend months near their users before hiring salespeople
- •Fanelli describes sharing AI and autonomy across the services while keeping narrow models for unique needs like command and control
- 32:35 – 37:24
Budget History and Why the Spending Cycle Is Rigid
Miller traces how naming programs in the budget since 1968 shaped acquisition, and explains the Army's limited flexibility within a three-to-five-year prediction cycle. Fanelli says capabilities must be undeniably better to justify divestment.
- •In 1968 the Minuteman III became the first named program in the budget, replacing broad pots of money
- •The Army is 21% of the defense budget and has flexibility over less than half of that
- •Thousands of NDAA line items dictate spending, forcing the Army to predict the future three to five years out
- •Fanelli says offerings must be undeniably better so a little CapEx can be justified by lower total cost of ownership
- •Fanelli notes DIU and innovation activities are still under 1% of the budget
- 37:24 – 41:08
Risk Appetite: Bridging Silicon Valley and Washington
Steph asks how to bring the risk cultures of venture and Washington closer. The guests say risk is measured on waste, fraud and abuse rather than on lost time, and they argue for measuring speed and outcomes instead.
- •Miller says the people managing risk measure waste of dollars, not the cost of spending a dollar over ten years
- •Fanelli cites In-Q-Tel's ratio of 25 outside dollars per government dollar as a proven model
- •Fanelli uses a quarterback analogy for being measured on avoiding mistakes
- •Fanelli proposes speed as an independent variable alongside the older cost as an independent variable
- •Miller says venture investors should engage with leadership about risk measures
- 41:08 – 43:26
Cross-Domain Fights, Marine Innovation Unit, and Being Ready for Windows
Using cross-domain solutions as an example, the guests describe solved problems still being fought over and ways the Navy is scouting and scaling technology. Fanelli urges successful companies to share their stories and be ready when windows open.
- •Miller says cross-domain as a service is just an API for binary XML between two networks and is a solved problem
- •Fanelli says solved problems should stay solved and the focus should move higher up the stack
- •The Marine Innovation Unit keeps reservists with venture jobs scouting technology
- •Fanelli wants success stories to build a flywheel and overwhelm the system with how much better it can be
- •CAPE's pilot in Guam became operationally needed, showing the value of being ready
- 43:26 – 47:47
Looking Ahead: What Each CTO Is Excited About
Steph asks about the year ahead. Fanelli expects scaled fielding into operations and maintenance funding, and Miller outlines Army programs including next-generation command and control, Project Fly Trap on counter-drone, and an autonomy consortium.
- •Fanelli says these things no longer need to be proven and sees the best chance in six to 12 months to get new capability into full production
- •Miller says the Army's next-generation command and control consortium is moving into its program phase this year
- •A tank commander at the National Training Center showed command and control on an Android phone replacing a 20-year-old tablet
- •Project Fly Trap in Europe applies lessons from Ukraine to automate counter-drone sensing
- •Miller plans a consortium for autonomy so the Army trades iron rather than blood in first contact
- 47:47 – 52:46
Quickfire Questions: Flags, Books, Misconceptions, and Movies
Steph asks lightning-round questions about startup red and green flags, a recommended book, misconceptions about each service, and favorite movies. The guests give candid answers and close the conversation.
- •Fanelli's red flag is working with bad, long requirements documents, and he points to capability need statements instead
- •Miller's green flag is building your own North Star product, and his red flag is relying on the DoD as sole revenue
- •Fanelli recommends The Hundred Year Marathon by Michael Pillsbury
- •Fanelli says the Navy has won 12 policy exceptions by bringing data, and Miller says the Army has not missed this fight
- •Favorite movies are Top Gun: Maverick for Fanelli and Starship Troopers for Miller
