STATUS OF THIS MEMO
This memo solicits founders. It identifies capabilities a sovereign state is expected to possess which the United States has lost, never built, or quietly outsourced. It is a follow-up to the original Sovereignty Gaps piece [0]. Distribution of this memo is unlimited.
[0] x.com/AbhayVenkatesh1/status/2082252761929044140
ABSTRACT
A sovereignty gap is a function the American state should be able to perform but cannot. Some gaps are visible from orbit: we do not enrich our own uranium at scale. Some are visible from the street: patients in kidney failure kept alive on dialysis technology from another century. Some are invisible until they are catastrophic: we do not actually know how many people there are.
Historically these gaps were closed by the state itself -- ARPA, the interstate system, the census. Today the state's capacity to close them is itself the gap. We believe the fastest remaining path is a company: build the capability privately, sell it to the sovereign, and become load-bearing.
These questions should be viewed through one lens: what does the regime want? Why, for instance, does the state need LLMs at all? Because LLMs are a form of GDP. Outside of AI, there is no growth.
What follows is a list of gaps we would like to see closed. It is not exhaustive. If you are working on one of these, or on a gap we have missed, we would like to talk to you.
A warning before the list: by solving one sovereign gap, you may induce another. Immigration patches fertility and creates its own gap; weather modification that raises your yields lowers someone else's. This is not a reason to stop; it is the game. Whack-a-mole, played seriously, is called statecraft.
REQUESTS
1. WEATHER
Weather understanding and weather modification. Why does a state need the ability to change the weather? Because China already can -- you are matching an adversary in their capacity -- and because agricultural yields change through increased rainfall, which is geopolitical advantage over other sovereigns. NOAA was attempting to upgrade sovereign weather systems before attempting became the wrong tense.
WANTED: Better sensing and data for the American atmosphere. Credible, measurable weather modification.
2. MANUFACTURING
Advanced manufacturing matters for resilience, and the honest version of this request admits the tension: it may lead to inflation, capital flows wherever returns are, and cynicism about manufacturing has often been correct. The physical world is being repriced -- whether on underlying business quality, or simply because returns are flowing out of software into physical-AI companies, is an open question. Intuitive Surgical was a DARPA project.
WANTED: Manufacturing capacity in categories where dependence on an adversary is intolerable. Not nostalgia. Chokepoints.
3. NUCLEAR
The United States lacks domestic enrichment capacity and is slow to field new reactor designs. Companies like Antares (new reactors) and General Matter (domestic enrichment) also answer a resilience question that cheaper clean sources like solar cannot.
WANTED: More of this. Fuel cycle infrastructure the state can rely on in a crisis, at costs that survive contact with solar.
4. ORGANS
There are not enough organs for the patients who need them. The kidney process is orchestrated through government, and the fallback -- dialysis -- is extremely outdated technology paid for by Medicare. The state already owns this problem; it is one of the few places where biotech sells directly into a sovereign obligation.
WANTED: More organs, by whatever means proves out. Dialysis made obsolete.
5. POPULATION
Fertility is collapsing and the tax base collapses with it. A gerontocracy cannot afford itself, votes itself richer, and does not innovate. Current interventions are oversold: egg freezing works out to roughly a 16% path to an actual child, and may function as a permission slip to defer indefinitely. Immigration patches the gap and induces its own.
WANTED: Reproductive technology that actually works, priced for deployment. Population management as infrastructure, not punditry.
6. CRITICAL MINERALS
Extraction and processing of critical minerals is concentrated in the hands of a strategic competitor. Companies like Wetstone help solve this.
WANTED: Domestic extraction and, especially, midstream processing. Speed matters more than elegance.
7. IDENTITY
There is a basic question of "we don't know how many humans there are." A surprising number of official population figures are fake [1], and we need to distinguish between humans and AIs. Is there a version of Worldcoin built for the sovereign rather than the token?
WANTED: Identity and census infrastructure a state can trust. Proof-of-personhood a state would actually adopt.
8. SOCIAL CREDIT
Every state runs a reputation system; most run it badly, on credit bureaus and watchlists built decades ago. Credit assignment is a sovereign function whether or not anyone wants to say so out loud.
WANTED: The version of this that is compatible with a free society -- built deliberately, before the alternative is imported.
9. CYBER
In August 2026 the White House authorized federal law enforcement to conduct cyber operations against foreign transnational criminal organizations -- and invited private companies to gather threat intelligence and propose operations under government direction [2]. This is the letter of marque, reissued for cyberspace: privateers undertaking the risky adventures a cautious state cannot, against an enemy that took $20.8 billion from American consumers last year alone.
WANTED: The cyber privateering company. Offense under sovereign sanction, built to outlast the administration that chartered it. Yes, you are inviting nation-states to hack you back.
10. BIOENGINEERING
The cat is out of the bag on bioengineering and genetic editing in humans. Enhanced Games is the primer: a very big deal, deliberately framed as a toy. States that engage with this seriously will not announce it.
WANTED: The picks and shovels of human enhancement -- screening, editing, verification -- built where the sovereign can see them.
11. AGI AND LABOR
AGI could lead to mass unemployment. The early indicators will show up first in the countries we offshored to, and the data is not yet clear -- though the Philippines and India already run social orders designed to keep people busy. There are a lot of bullshit jobs anyway; the market partially cleared them out post-ZIRP (Docusign with 10,000 employees), but it is still true, and it may remain true post-AI. Mass use of ChatGPT may be doing cognitive damage to the workforce -- though if AI grows the economy for you, is mass dulling even a problem?
Does China have more productive knowledge-work capacity than we do simply because it has more humans? Its companies collude on pre-training because they lack GPUs, the state coordinates them, and it holds more grid capacity than we do while remaining a net importer of energy. Whether the Chinese are truly scaling-pilled is unclear. Outside these two countries, almost no state has a coherent AI policy at all: what is Australia's? Sweden's? Germany's? Does France have some Yann LeCun copium?
WANTED: Companies that add frontier intelligence to the national labor pool and raise total knowledge-work throughput. Grid capacity built at the speed compute demands. Early-warning instruments for the labor shock.
12. ELITES
What happened to the elites? If you went to Harvard before, you worked at the CIA. East coast elites remain more interested in philanthropy and institutions than west coast elites, and the great mid-century instruments -- Sloan, Kettering -- have few successors beyond experiments like Renaissance Philanthropy. Meanwhile the mimetic power of socialism persists, even as technocapital appears to be winning.
WANTED: Institutions that put elite talent and elite capital back to work on sovereign problems -- venture philanthropy in the Sloan mold, long-horizon moonshot R&D, a reason for the ambitious to serve.
RESPONSES
This solicitation has no deadline and no program office. We also hold open questions we cannot yet close: what are the sovereignty gaps in the data space? In the AI space? Should central intelligence give data to private robotics companies?
If you are building against one of these requests, have an answer to an open question, or believe the list is missing a gap, submit a response:
https://forms.gle/NhXdy9zx8HBCnBkm9
Abhay Venkatesh
Aashay Sanghvi
Distribution of this memo is unlimited.