AI-assisted analysis. See our editorial policy.
Human editorial review not recorded
The problem with data sovereignty is that it has been treated as a question about location when it is a question about jurisdiction, and this panel describes the mechanism that separates the two.
A digital embassy extends a country's critical digital infrastructure beyond its borders while the sending state retains legal control over the data, the compute and the governance (9:48). The servers sit somewhere else. The law that governs them does not move.
The inversion that makes it interesting
The framing offered here is the useful part, and it starts from a physical fact.
Many countries lack what a large data centre requires — reliable power at scale, water for cooling, the grid to support both. And those things cannot be relocated: you cannot move a water supply across a border, and moving an energy grid is harder still (13:50).
So invert it. Rather than demanding that infrastructure be built where the sovereignty is, move the facility a country would ordinarily build at home into a host country that already has the resources, and carry the jurisdiction with it. The sending state gets sovereign control without having to manufacture the physical conditions for it.
This also has a sustainability argument attached that is more than decoration. A host country with genuine surplus power supports the facility without straining anything; the same facility forced into a country without that surplus competes with everyone else for it.
The precedent is older than AI
The concept predates the current buildout. Digital embassies were originally arranged to protect continuity of essential government services — a small state placing its critical records under its own law on another state's territory, so that a disruption at home does not erase the functioning of government (9:20).
What has changed is the driver. Continuity was a defensive requirement satisfied by relatively modest infrastructure. AI capability is an offensive one, and the infrastructure is enormous, which converts a narrow diplomatic instrument into a template for where the world's compute physically sits.
What a framework is actually for
The panel's argument for standardising this is not about principle, and it is better for that.
These arrangements are bilateral by nature — a sending state and a host state, with treaty-level questions about immunity, access, dispute resolution and what happens if relations deteriorate. Each pair could negotiate from scratch. The point of a shared framework is that the accumulated bilateral agreements stop reinventing the same technical, legal and governance answers every time (19:40).
The third element is the one investors will care about most: time. Anyone deciding where to place a facility wants to know, before committing capital, that a tried and tested governance framework already exists in that jurisdiction (26:40). Absent that, the negotiation itself becomes part of the project risk, and countries with the physical resources but no legal template lose to countries with worse resources and better paperwork.
What it does not resolve
The framework addresses law and process. It does not address leverage.
A sending state whose national AI capability physically resides in a host country has transferred something that no legal instrument fully returns. Agreements hold while relations hold. The scenarios that make sovereignty matter — sanctions, conflict, a change of government on either side — are precisely the scenarios in which a contract about physical infrastructure inside someone else's territory is tested, and the party holding the buildings has options the other does not.
That is not an argument against the arrangement. Every state already accepts comparable dependencies in energy, shipping and payments. But it does mean the framework is doing something narrower than the word sovereignty suggests: it makes the dependency legible and negotiated in advance, rather than removing it.
Which may be the honest version of what is on offer. For a country choosing between no AI infrastructure and infrastructure under someone else's roof but its own law, a well-drafted dependency is a real improvement on the alternative. It is just not sovereignty in the sense the term usually carries.
Talk chapters
Key takeaways
- 01
A digital embassy extends a country's digital infrastructure beyond its borders while the sending state keeps legal control over data, compute and governance. 9:48
- 02
The inversion rests on a physical fact: water supply and energy grids cannot be relocated, so the facility moves to where those resources already exist. 13:50
- 03
The concept predates AI, originally arranged to protect continuity of essential government services rather than to host capability. 9:20
- 04
The case for a shared framework is that accumulated bilateral agreements otherwise reinvent the same technical, legal and governance answers each time. 19:40
- 05
Investors want a tried and tested governance framework in place before committing capital, which makes legal readiness a competitive factor between host countries. 26:40
Entities mentioned
Related talks
Huang brings a diagram to Davos: AI as a five-layer cake running energy, chips, cloud, models, applications — with economic benefit landing at the top and every layer below it a precondition. His argument for why this is a genuine platform shift rather than a product cycle is the strongest part, and it does not rest on his commercial position: software was pre-recorded and worked on structured data, whereas a machine that reasons about unstructured input and inferred intent makes previously impossible applications possible. What the framing accomplishes is worth noticing separately. By presenting the layers as a chain rather than a portfolio, it converts infrastructure spending from a bet into a prerequisite, and the question of proportion between layer-two spending and layer-five value stops being askable. Read against the GTC keynote two months later, the same business gets two framings: one a case for choosing his product, the other a case for the category existing at the scale he needs.
Harari spends most of his address establishing terms before asking the question he came to ask. His preliminary work is to dismantle the word tool: a knife's use is decided by whoever holds it, whereas what is arriving decides for itself, and can also invent new kinds of knives. From there he argues that anything constituted by words — law, books, text-centred religion — is exposed, while drawing a firm line at feeling, where he says there is no evidence at all. The structural claim is that the ancient tension between letter and spirit has always run inside humanity and is about to be externalised between humans and the new masters of words. Only then does he arrive at personhood, and his handling is precise: corporations, New Zealand rivers and Indian deities hold legal personhood safely because the decisions are made by humans behind the container. An entity that decides for itself ends that arrangement. He does not answer the question; he tells the room it is coming.
A show of hands opens the session: nearly everyone has piloted, far fewer have scaled, and everyone who scaled hit problems they did not anticipate. What makes the panel useful is where the four answers do not point. None of the executives — running a healthcare manufacturer, a payments network, an energy producer and a consultancy — blames model capability, cost or data infrastructure. All four describe an organisational constraint. McInerney's account is the sharpest and is an account of failure: eighteen months of executive advocacy and democratised model access produced nothing, until three hundred senior leaders were put in a room for two days and made to build agents themselves. Jakobs supplies the mechanism worth copying, measuring returned clinician time against the three to seven minutes a patient currently receives rather than against cost. Nasser rejects the premise that acquiring compute produces value, and locates returns in operations rather than in the back-office functions most organisations automate first.
A year after their first joint appearance, the heads of Anthropic and Google DeepMind returned to a shared stage and disagreed mainly about speed. Amodei held to a horizon of one to two years for systems that outperform humans across most cognitive work, resting the claim on a self-improvement loop that runs through code; Hassabis kept to the end of the decade, arguing that verifiable domains like coding and mathematics automate far earlier than natural science, and that the capacity to pose a new question rather than answer an existing one is still missing. The exchange is most useful where they converge: both accept the loop is the variable that decides everything, both are sceptical of doomerism without dismissing the risk, and both want more time than the competitive dynamic allows. Amodei's chip-export argument and Hassabis's call for minimum international safety standards are the two concrete policy asks.
The distinction in this session that deserves to travel is a two-word change: the model is moving from an attention economy to an attachment economy. That changes what is measured and what regulation would have to address, because attention competes for time while attachment competes for relationship, and the two produce different products from identical technology. Attention is finite in a way people notice; attachment produces reliance that feels like preference, which makes it harder to regulate for the same reason it is harder to notice. The supporting argument is about incentives rather than intent: earlier engagement produces more data and longer relationships. The regulatory proposal — measuring well-being outcomes rather than asking for safety by design — identifies the right target without solving the measurement problem that made regulators settle for a floor in the first place.
The most useful sentence is a diagnosis of how clients are getting it wrong: they see a cost opportunity rather than a revenue one. That framing decides everything downstream, because an organisation treating AI as cost reduction measures success by what disappears and arrives at a smaller version of what it already was. The market number offered — seventeen per cent index growth in 2025 that nobody foresaw — is deployed against bubble anxiety and argues in both directions, since unforecast gains are poor evidence for confidence in any current consensus. The panel then locates the real source of instability outside technology entirely, in three geopolitical situations generating enough uncertainty to dominate planning, which is a corrective worth taking seriously at a technology conference. Their closing question about thriving across multiple futures is correct and sits uneasily beside the cost framing they diagnosed at the start.