Doug Bernauer | The a16z Show Summary

1. Key Themes

The Ocean as the Last Great Unclaimed Frontier for Technology

The ocean represents a massive technological opportunity that has been culturally and strategically neglected since the 1960s, creating a wide-open space for a "SpaceX moment" in the undersea domain. The co-founder of Ulysses traces this neglect to a specific confluence of historical events — a fatal accident in the Navy's Sea Lab program, Vietnam War budget redirections, Cold War classification of undersea engineering, and Jacques Cousteau's pivot from exploration to conservation.

"50 years later, we're still like working with the same tools in the ocean. And it's exactly this kind of like exploration spirit, this like desire to kind of conquer this great blue frontier that we're trying to reignite and be that like SpaceX from the ocean that get like young kids dreaming about being, you know, aquanauts as well as like astronauts." [00:07:01]

The Undersea Domain Is the One Area Where the US Can Still Achieve Decisive Military Advantage Over China

While China is dramatically ahead on surface shipbuilding (hundreds of ships per year vs. US single digits) and aerial drones (tens of millions vs. US single-digit millions), the autonomous underwater vehicle space is still roughly parity — hundreds on each side — making it an actionable window for US dominance.

"In the autonomous, like underwater vehicle, undersea drone space, we're looking at like single digit thousand autonomous underwater vehicles or underwater drones made every year. And, you know, hundreds of them are made in China and hundreds of them are made in the US. So it's not like a massive, like a total disparity here. We're like on the same kind of order of magnitude. So the undersea domain is definitely a domain in which like the US could create an advantage." [00:08:24]

Critical Minerals Are a Stack Problem, Not a Spot Solution Problem

The core insight from Mariana Minerals is that you cannot sell point solutions into mining — you have to own the entire vertical from exploration through refining. The industry's risk structure (every single change risks a multi-million dollar plant shutdown) means incremental tech adoption is structurally impossible, and the only way to capture tech efficiency gains is to be the operator yourself.

"We actually do believe you have to control every single piece of the entire journey, the entire life cycle of, you know, an atom of metal end-to-end to actually be able to build a tech company here. This is not about, you know, a point solution for one particular part of the process in order to actually capture the gains and efficiency and build a feasible business. You really have to own the entire process end-to-end." [00:48:32] — Erin Price-Wright

China's Minerals Dominance Is a Talent Story, Not Just a Policy Story

Everyone discusses Chinese state policy support for critical minerals, but the deeper and less-discussed advantage is a massive, skilled, experienced labor pool. A single Chinese nickel refinery in Indonesia had 13,000 people on site during construction and commissioning — a number Western companies simply cannot mobilize.

"I was in Indonesia in February and was kind of visiting one of the recent Chinese nickel refining operations. And so they had 13,000 people on site during construction and commissioning. And, you know, if we were building a refinery in the U.S., it's hard to mobilize, you know, a tenth of that realistically. And when you have, it's not just about the number of people. It's about being able to iterate on every individual work front as fast as humanly possible. And we just don't have that labor." [00:31:57] — Turner Caldwell

The 3-Week Data Lag in Megaproject Construction Is a Solvable Software Problem With Massive Leverage

Large-scale construction projects operate on a roughly three-week lag between what happens in the field and what leadership can actually see and make decisions on. The current workaround — daily stand-up circles — is a pre-digital hack. Collapsing that lag to real-time is Mariana's "Capital Project OS" thesis, and the leverage is enormous given that mining is effectively construction that never ends.

"There's like a three week lag generally for like really large construction projects where you are trying to aggregate data from all the different contractors, all of the different parts of the facility into a consolidated integrated schedule which you can then make decisions off of like how do I prioritize what I'm doing today. And the way you run in between those three weeks is people stand in circles every morning and they say what are you doing today, what are you doing today, what are you doing today." [00:37:14] — Turner Caldwell

AI/RL Applied to Process Control Can Compress Years of Commissioning Into Months

Google's DeepMind achieved 30-40% energy reduction in data center thermal systems with only nine control variables. A minerals refinery has a thousand control variables with 24-48 hour feedback latency. The opportunity for reinforcement learning to optimize these circuits — and to compress commissioning timelines from the Western standard of 2-4 years down toward the Chinese standard of 6 months — is the core technological thesis of Mariana's PlantOS.

"Refineries that were built recently are still not commissioned that were built like four, three years ago. But the Chinese companies are doing it in like six months. And a lot of Western companies, it takes two to four years. And that stacks up where we need to build like an insane number of mines and refineries. And if you are kind of four times longer or five times longer every time you build a refinery..." [00:42:49] — Turner Caldwell

Nuclear Micro-Reactors Are a Diesel Replacement Product, Not a Grid Product

The framing of nuclear as a grid-scale solution misrepresents what micro-reactors like Radiant's are actually competing with. The real market is the off-grid diesel generator market — anywhere diesel costs $6.50 or more per gallon, nuclear wins on economics. That includes Hawaii (80% diesel powered), islands globally, and remote industrial sites. This reframing dramatically clarifies both the addressable market and the go-to-market.

"You could run a diesel generator or you can run a nuclear reactor and you're really deciding between those two things. We don't beat super cheap diesel. We beat diesel at like $6.50 a gallon, that kind of a number. So that's where our initial customers need to be." [01:09:49] — Doug Bernauer

There Is No Nuclear Industry Yet — And That Is the Opportunity

Doug Bernauer makes the provocative and precise point that we are in pre-Kitty Hawk for nuclear. Most excitement is premature because the fundamental milestone — new-design reactors going critical at full power — has not yet happened. Radiant is the only new-design reactor currently permitted to go to full power, with a target of 2026.

"The one thing I like to say, it sounds a little sensationalist, is that there is no nuclear industry. We're kind of, it's almost like we're getting excited about flight before Kitty Hawk." [01:05:38] — Doug Bernauer

2. Contrarian Perspectives

The US Has No Rare Earth Resource Deficit — Only an Exploration Deficit

Conventional wisdom holds that the US lacks the geological endowment of rare earths compared to China. Turner Caldwell pushes back directly: the US has enormous natural resources, but the permitting burden for exploration on federal land (triggered at just five acres) has so suppressed exploration activity that we simply haven't looked for what's there.

"The USGS estimate for US rare earth reserves — that is tied to lack of exploration activity, not necessarily fundamentally like a lack of geo presence. We haven't looked for it. Either we haven't looked for it, or it's hard to find in like high concentrations that are mineable, which we're trying to kind of like drop the percentage requirement that makes something economical." [00:55:48] — Turner Caldwell

You Should Build Mining Infrastructure at the Bottom of Commodity Cycles, Not the Top

Counter to the instinct to invest when the headlines are loudest (and prices are highest), Caldwell argues that the right time to build minerals infrastructure is when prices are depressed — when no one else wants to. Lithium is explicitly called out as being exactly in this position right now.

"You actually want to be building infrastructure at the bottom of commodity cycles, not at the top of commodity cycles. You know, it's the Warren Buffett quote of, you know, invest when there's blood in the water. Lithium is like exactly in this position right now. And that's why we're focused on lithium." [00:53:30] — Turner Caldwell

Selling SaaS Into Mining Is a Death Spiral

The standard venture playbook — build a software layer and sell it into an incumbent industry — fails structurally in mining. The industry builds one large plant every five years, pilot programs never convert because the commercial-scale window has already passed, and operators will never trust external software recommendations over their own instincts. It is not a cultural problem to be overcome; it is a structural feature of the industry.

"There's a lot of folks that are doing SaaS products, which is kind of the lowest cost way to get into, to like generate uplift in a mining project or a minerals refinery. And the barrier there is ultimately how do you get the operators to trust the recommendations from this SaaS tool from this small company that is trying to kind of tell you how to run a plant. And the culture is typically like, don't touch my things, don't touch my cash register. I've been calling it a death spiral for a lot of the folks that are trying to sell into the mining industry because it's hard." [00:30:13] — Turner Caldwell

Seasteading Was Right but Just Needed Enabling Technologies That Now Exist

The mainstream view is that seasteading was a naive libertarian fantasy that failed. The Ulysses founder argues it was simply too early — the enabling technologies (Starlink for connectivity, autonomous cargo aircraft for logistics, underwater robots for maintenance, next-generation solar and wave energy) are only now converging to make it economically viable.

"What was wrong, what they didn't have at the time were the tools and technology to build it. It was like a good idea, but it was too early because the technological enablers weren't doing it. Starlink, you know, for the first time ever, you can actually have a habitation out there that can be connected to the rest of the world. You have your Ulysses autonomous underwater robots that go in and like fix the cables underneath it. So you're starting to see the preparation of tools now that I think can actually reduce the cost of preparation down such that the benefits do outweigh the costs in something like seasteading." [00:11:58] — Will O'Brien (Ulysses founder)

Nuclear Waste Storage Is a Political Problem, Not a Technical One — and It's Distorting Risk Everywhere

The US already has a working deep-bore nuclear waste repository (the Waste Isolation Pilot Plant in New Mexico), currently used for defense waste. The failure to build a civilian equivalent means gigawatt-scale plants store waste on-site at coastal locations far more vulnerable to natural disasters than a salt dome in the high desert. The problem is entirely political, not technical, and it is suppressing the entire sector.

"Because of that, these gigawatt scale plants are operating, generating nuclear waste, and they have to store it at the same site where they're making power. And in California, this is coastal regions that are risky, that you can have a tsunami or something instead of taking it and putting it in a salt dome structure in the high desert where there's no water, no risk of like certain natural disasters. So it's just a smarter, safer, better idea and we don't do it." [01:13:36] — Doug Bernauer

3. Companies Identified

Ulysses Autonomous ocean robotics company building a fleet of autonomous surface vehicles (Leviathan) and autonomous underwater vehicles (Mako) for conservation, commercial infrastructure inspection/repair, and defense applications. Described as building "the ocean company" — the full-stack platform for the ocean analogous to what Starlink is for space. Why mentioned: Portfolio company featured as a central subject; founder describes best-in-class specs for small-to-medium AUVs, Navy interest, and a planned manufacturing facility that would 2-3x global AUV supply.

"MAKO, our underwater vehicle is truly like best in class for its form factor, the small to medium size... That facility alone would, you know, 2 to 3x like international supply of autonomous underwater vehicles. And obviously all of that will be built in the United States." [00:02:00] and [00:10:09]

Mariana Minerals Vertically integrated, software-first critical minerals project developer and operator. Builds and operates mines and refineries using LLM-powered construction management (Capital Project OS) and AI/RL-powered process control (PlantOS). Came out of stealth with $85 million raised. Why mentioned: Portfolio company and central subject; thesis is that vertical integration is the only viable path to tech-enabled mining.

"We're a vertically integrated software first minerals project developer and operator. And so we focus on the back end of the minerals value chain which is actually doing the detailed engineering, getting through the permitting, building the asset, commissioning the asset and then operating the asset." [00:35:45] — Turner Caldwell

Radiant Nuclear micro-reactor company building trailer-mounted one-megawatt reactors designed to be mass-produced (targeting one per week) from an 80-acre Tennessee facility. Reactors deploy in 48 hours, run for five years (equivalent to a full oil tanker of diesel), and target off-grid diesel replacement markets. Founded 2019 by Doug Bernauer, formerly of SpaceX. Why mentioned: Portfolio company and central subject; only new-design reactor currently permitted to go to full power; on schedule for 2026 full-power criticality.

"We are now the only reactor permitted, of these new reactors, to go to full power... We're targeting one per week coming off of a production line from our Tennessee facility, which is an 80 acre site we just signed for in October." [01:07:00] and [00:08:27] — Doug Bernauer

SpaceX Launch and spacecraft company. Why mentioned: Multiple founders and investors use SpaceX as the archetype of what a new entrant can do to revitalize and de-calcify an incumbent industry; Doug Bernauer spent 12 years there working on Falcon 9, Grasshopper, Hyperloop, Boring Company, and Mars Colony design before founding Radiant.

"I joined in 2007. So I joined when they had two failed rockets, no successful rockets. And so I got to work on the first ones that worked." [01:03:20] — Doug Bernauer

Anduril Defense technology company. Why mentioned: Trae Stephens, co-founder, is cited as having debated the billionaire tax at a Hamilton Society event as "a real life billionaire, founder of Anduril." Also referenced as an exemplar of the talent pipeline (hard tech workers from Tesla, SpaceX, Anduril) now willing to enter the minerals space.

"We literally had, you know, Trae Stephens, a real life billionaire, founder of Anduril, debating the two guys who wrote the policy." [00:18:44]

MP Materials Rare earth mining and magnet manufacturing company. Why mentioned: Cited as the specific recipient of a major US Department of Defense demand-side support deal, including equity participation — held up as the model for government support that could mobilize private capital into critical minerals.

"They did this just now with MP Materials. And that, you know, ideally provides some stability on the revenue side of things so that investors, there's trillions of dollars of capital kind of like dry powder just sitting around waiting to be deployed." [00:58:53] — Turner Caldwell

DeepMind / Google AI research lab acquired by Google. Why mentioned: Held up as the proof-of-concept for AI-optimized industrial process control — DeepMind's work on Google data center thermal systems achieved 30-40% energy reduction with only nine control variables, validating the thesis for applying RL to mining refineries with a thousand variables.

"Google kind of proved this. They bought DeepMind in 2016, 2017. And one of the first things they did was throw the DeepMind team at automating and optimizing the data center thermal systems... just in that relatively simple system, they were able to reduce energy consumption by 30, 40%." [00:40:20] — Turner Caldwell

Starlink Satellite internet constellation operated by SpaceX. Why mentioned: Cited as the specific technological enabler that now makes seasteading and remote ocean habitation viable for the first time — connectivity that didn't exist when the idea was originally proposed.

"Starlink changes the game, right? Starlink, you know, for the first time ever, you can actually have a habitation out there that can be connected to the rest of the world." [00:11:58]

Pantalassa Ocean geothermal / wave energy company. Why mentioned: Named as a company developing technology to generate energy from natural heat in the ocean, cited in the context of enabling technologies for seasteading.

"We're also seeing interesting advances in ocean geothermal or basically just like taking advantage of natural heat in the ocean to generate energy with Pantalassa generating wave energy." [00:13:26]

Car Power Ship Modular floating power station developer. Why mentioned: Named as a company developing modular floating power stations on ships that can be brought to remote offshore locations — another enabling technology for ocean habitation and operations.

"We have Car Power Ship who developed these modular kind of floating power stations that you can put on a ship so that it can be brought out to locations." [00:13:26]

Waste Isolation Pilot Plant (WIPP) US deep geological repository for defense nuclear waste, located in New Mexico in a salt dome structure. Why mentioned: Cited by Doug Bernauer as proof that safe, centralized nuclear waste storage is technically solved and already operational for defense waste — making the failure to build a civilian equivalent a purely political failure.

"We already have a Waste Isolation Pilot Plant in New Mexico, which is like this deep bore hole down inside of a salt structure. This is where defense waste already goes and they just said they were going to build it and they built it." [01:13:07] — Doug Bernauer

4. People Identified

Doug Bernauer Founder and CEO of Radiant. Spent 12 years at SpaceX (2007–2019), working on Falcon 9 ground systems, the Grasshopper reusable rocket (4-person team reporting directly to Elon Musk), Hyperloop, Boring Company, and Mars Colony design. No prior company founding experience. Why mentioned: Central subject; founded Radiant in 2019, is on track for full-power reactor criticality in 2026, has secured an 80-acre Tennessee manufacturing site targeting one reactor per week production.

"In 2020, I said in 2026, I will put a full scale reactor and get it critical and get it up to full power. And we're on schedule to do that, which is kind of wild... We are now the only reactor permitted, of these new reactors, to go to full power." [01:06:30] — Doug Bernauer

Turner Caldwell Founder of Mariana Minerals. Background implied in deep process chemistry and metallurgy. Why mentioned: Central subject; came out of stealth with $85M raised; articulates a sophisticated first-principles view of where the mining industry is broken and how software-first vertical integration can fix it. Has personally visited Chinese nickel operations in Indonesia and brings direct operational insight to the geopolitical minerals competition.

"We want to build 10 projects in 10 years. Those projects will be at increasing scale over time. What will have demonstrated that the 10 year mission will have been accomplished, other than building those 10 plants, is that we will no longer be as worried about like our fundamental capability to go and build this complex infrastructure. We will have unlocked it." [01:02:24] — Turner Caldwell

Trae Stephens Co-founder of Anduril. Why mentioned: Cited for excellence and intellectual engagement — personally showed up to debate the two authors of the billionaire tax policy at a Hamilton Society event, and was credited with giving attendees the specific analytical tools to understand why the policy is structurally flawed.

"We literally had, you know, Trae Stephens, a real life billionaire, founder of Anduril, debating the two guys who wrote the policy. Right. And I think everyone came away with it like, oh wow. Okay, we now understand specifically why it's bad." [00:18:44]

Erin Price-Wright General Partner at a16z American Dynamism. Why mentioned: Articulates the investment thesis for why vertical integration is the only viable model for tech-enabled mining; describes years of failed attempts to find a viable point-solution mining investment before landing on the vertically integrated model.

"Over many cycles we've gone out and spent a lot of time meeting companies and, as I mentioned before, the challenge is how do you sell a point solution or a point piece of technology into this industry that is has very little incentive to adopt it." [00:47:34] — Erin Price-Wright

Elon Musk CEO of SpaceX and Tesla. Why mentioned: Directly credited with the insight that led Doug Bernauer to nuclear — Musk reviewed Bernauer's solar-powered Mars Colony plans, found them to require "four miracles on a single mission," and told him to look at nuclear instead.

"I was trying to do it with solar and getting totally stuck and showing Elon these plans that were like four miracles we need on a single mission. And Elon was like, you probably should look at nuclear. And that's really the jumping off point." [01:04:29] — Doug Bernauer

Jacques Cousteau French ocean explorer and conservationist. Why mentioned: Cited as the pivotal cultural figure whose shift from exploration to conservation was one of several forces that caused an entire generation to lose interest in ocean engineering and colonization.

"Jacques Cousteau, the biggest cultural icon for it, he was real about exploration. And then he kind of pivots more to like conservation. So there's a cultural shift, there's this like classification. And as a result, 50 years later, we're still like working with the same tools in the ocean." [00:06:34]

JFK (John F. Kennedy) 35th President of the United States. Why mentioned: Cited as having commissioned a subsea equivalent of the space program and tripled funding year-over-year to explore Earth's oceans — establishing that ocean exploration was a co-equal national ambition to going to the moon before being abandoned.

"JFK commissioned this plan that was effectively like the undersea equivalent of a space program. And he tripled funding like year over year to like explore the Earth's oceans." [00:05:15]

Peter Thiel Investor and entrepreneur. Why mentioned: Cited in two contexts — as an advocate for seasteading and the concept of the sovereign individual, and implicitly as an intellectual forebear of the Hamilton Society's ideas.

"Thiel kind of speaks about this, you know, the sovereign individual is probably one of the first books that really kind of coined, I think, a lot of these ideas back in the 90s." [00:10:51]

5. Operating Insights

The "One Miracle Per Startup" Rule as a Capital Allocation Framework

Doug Bernauer articulates a crisp operating principle: a startup should solve one miracle that leads to a product and revenue, then earn the right to pursue the next miracle. This is directly why Radiant focuses on off-grid trailer-mounted reactors rather than grid-scale deployment — each is a separate miracle and conflating them kills companies.

"One miracle that leads to then a product and revenue, that's the way. And then you have time to think about another miracle. You don't want to have too many miracles in a startup." [01:11:44] — Doug Bernauer

Treat Commodity Cycles as a Construction Scheduling Tool

Turner Caldwell frames the counter-cyclical build strategy not just as a valuation insight but as an operating directive: infrastructure funds won't show up at the top of the cycle, talent won't leave to work on projects that might not get built, and permitting and supply chains are actually easier to navigate when the industry is quiet. The instruction is to deliberately enter markets when they are depressed.

"You actually want to be building infrastructure at the bottom of commodity cycles, not at the top of commodity cycles. Lithium is like exactly in this position right now. And that's why we're focused on lithium." [00:53:30] — Turner Caldwell

Run Construction Projects Like Manufacturing Facilities by Closing the Data Feedback Loop

The operational insight is that the 3-week reporting lag in megaprojects is not inevitable — it is a data aggregation problem solvable with software. The implication for any operator running large-scale capital projects: collapsing that lag to near-real-time changes the decision-making cadence from reactive to proactive and removes the need for the "stand in circles" workaround that currently substitutes for actual visibility.

"We're really trying to accelerate and democratize access to data fundamentally and run construction projects like manufacturing facilities." [00:37:44] — Turner Caldwell

Use the Specific Context of Competing with 1,000 People to Force Automation Decisions

Mariana's founding logic offers a direct operating heuristic for any capital-intensive, labor-constrained business: start every process design question with "how do we enable 200 people to do what 10,000 people are needed to do today?" This forces automation and software decisions that would otherwise be deferred.

"How do you enable 200 people to do what 10,000 people are needed to do today? At least on the parent co side of things. And that comes from leveraging the recent advances in LLMs to automate workflows in the construction side of things and the engineering side of things and the procurement side of things." [00:36:15] — Turner Caldwell

6. Overlooked Insights

China Is Building an "Underwater Gray Wall" of Sensors Across the Indo-Pacific — and Almost Nobody Is Talking About It

Buried in a broader discussion of China's undersea capabilities is a brief but alarming disclosure: China is systematically deploying a networked array of low-power subsea sensors across the Indo-Pacific designed to detect submarines and other military assets. This is the undersea equivalent of a surveillance grid and would give China persistent intelligence on US naval movements in the region — potentially more strategically significant than any individual weapons platform. This was mentioned in passing and not picked up by the host as a thread worth pulling.

"They're also building what they call like an underwater gray wall of sensors across the Indo-Pacific. These would be sensors that you put down, pretty low power draw, and they can sense maybe submarines or other things moving and they're networked. And it gives them this kind of undersea kind of GPS or like navigating capability, which would obviously be very useful to them in the context of a war." [00:09:17]

The Industrial Tank and Equipment Supply Chain Is a Hidden Bottleneck Choking All Western Infrastructure Buildout

Turner Caldwell drops a single sentence — almost as an aside — that reframes the entire minerals infrastructure problem: the supply base for basic industrial equipment like tanks is broken in the West. A pump replacement in Australia takes 30 weeks; the same pump from China arrives in days. This is not a mining-specific problem. It is a foundational constraint on any company trying to build hard industrial infrastructure in the Western world, and it implies an enormous venture opportunity in industrial supply chain modernization that no one in the conversation chose to explore.

"The industrial supply base just for, like, manufacturing tanks is broken... A new pump in Australia could take, you know, 30 weeks, and getting that exact same pump with a mine in China, it shows up in, you know, a week or three days. That entire industrial equipment supply base we're going to have to look at at some point." [00:44:50] — Turner Caldwell