◆ The neutral layer for sovereign intelligence

Neverstranded.

A hyperscaler sells its own capacity. A reseller is locked to one stack. sovrgn is the neutral layer in between — any model, under your own law, with the route guaranteed even when a provider drops.

OpenAI SDK · Azure · Anthropic · AWS — change one base_url, keep your code
🇦🇺 🇬🇧 🇪🇺 🇺🇸 ⚡ reading the market…
Sovereign instance ·

You're in 's instance.

Routing under its law · backstop:

--:--:--
Total population
Workers (employed)
Knowledge workers — the demand base
Est. national demand at 30% adoption
National inference demand by hour (UTC)
Normalised working-hours demand. The vertical line is now (UTC); shaded bands are night at this instance's hub (real solar position). Your peak lands in allied sovereigns' off-peak — that's the time-zone hedge: your overnight silicon serves their day, theirs serves yours.
Estimates per the SRP demand model: knowledge workers × 30% adoption × 50,000 tokens per active worker per day.
Start here

Pick your country.

Everything in sovrgn is decided by jurisdiction — so the first choice is yours. Tap your country on the planet above, or step in here. The flows you saw are real: live grid prices, real submarine fibre (routes approximate), and demand following the sun — because allied instances hedge each other, your peak is another sovereign's off-peak.

Each tile shows its side of the market — ▲ Seller (net exporter) or ▼ Buyer (net importer) of intelligence — from Sovereignty = Capacity × Energy × Control × Access. Cheap power and spare GPU make you a seller; a residency mandate without domestic capacity makes you a buyer. Australia's capacity is grounded in published figures (AirTrunk, 755 MW+); UK/EU/US capacity is modelled (~) pending confirmation. Illustrative — directional, not a settlement figure.

Two sides of one market

Every country is a buyer or a seller of intelligence.

The balance on each tile isn't decoration — it's your position. ▼ Importers are short: they buy cognition they can't make at home. ▲ Exporters are long: cheap power and spare silicon they can sell. sovrgn is the exchange where the two meet — sovereign, metered, under each side's law.

▼ THE BUYER OF INFERENCE

You need cognition you don't control.

Enterprises, agencies & import-short nations — running on AI that lives in someone else's jurisdiction.
  • One contract, the whole market. Pooled demand buys leverage no single desk has — on price, terms, and which vendors are even allowed in.
  • Firm what's critical, float the rest. Reserve guaranteed in-country capacity for the paths that can't fail; batch the rest at spot.
  • Never wedged. A provider's outage, price spike or cut-off can't strand you — failover is in the contract, in-jurisdiction.
How buyers stay covered →
▲ THE SELLER OF INFERENCE

You have power and silicon to spare.

Energy-rich nations, sovereign DC operators & capacity holders — long on the inputs the world is short of.
  • Refine, don't rent. Turn cheap overnight power into exported cognition — capture the high-margin end instead of selling raw joules.
  • Sell your off-peak. Your night is another sovereign's working day; spare capacity clears into their peak across the fibre.
  • Set terms, not take them. Aggregate national capacity and you price into the market — a seat at the table with the hyperscalers, not under it.
How sellers go long →

Don't see your country? We can launch it quickly. Bring sovrgn to your nation →

For your business

When your AI provider fails, you don't.

You've built your operations on AI. That makes one vendor's bad day — an outage, a price spike, a policy change, a data-residency breach — into your bad day. sovrgn is the standing fallback that sits behind your existing setup and takes over the moment it's needed, without your data ever leaving the country.

Stay running

If your primary provider goes down or throttles you, sovrgn fails over to the next sovereign-approved option automatically. One outage stops being an existential event.

Stay home

Your data and your inference stay inside your jurisdiction by default. No silent offshore fall-through — if nothing in-country can serve it, you get a clean, auditable refusal, not a quiet border crossing.

Stay in control

One contract, the whole provider market behind it. Pooled demand gives you price leverage no single buyer has, and the freedom to switch vendors without rewriting a line.

For your nation

Aggregate your economy's demand for intelligence.

Intelligence has become a metered, globally-shippable commodity — a second energy market stacked on the first. A nation that buys it raw, business by business, is a price-taker: it inherits everyone else's outages, terms, and laws, and has no leverage with the handful of firms that supply the world's cognition. Hedge or be wedged. Pool the whole economy's demand and a country stops being a rounding error in someone else's roadmap — and earns a seat at the table.

A seat at the table

Aggregate the demand of every business, agency, and citizen behind one national instance, and you negotiate with AWS, Azure, OpenAI, and Anthropic as a buyer they can't ignore — on price, on terms, on which vendors are even allowed in.

Continuity of cognition

An economy that runs on borrowed offshore intelligence can be switched off from outside. A sovereign instance — with a national silicon backstop — means the country keeps thinking when global supply is constrained.

Refine, don't rent

Keep the energy and the inference on the same soil and you export refined cognition, not raw joules — capturing the high-margin end instead of renting back the smart output of your own power.

⬢ A sovereign is configuration, not a rebuild. Each nation plugs in its own jurisdiction rules and its own silicon backstop behind the same protocol. AloomU is Australia's. Claim yours →
Why this becomes mandatory

Soon, "what's your AI fallback?" is an underwriting question — and a regulatory one.

Cyber-insurance once paid out on any breach. Then it started requiring the controls — MFA, EDR, backups — before it would cover you at all. AI dependency is on the same path, and regulators are walking it too: the rules taking shape worldwide already treat the model layer and the application layer as separate things with separate duties. sovrgn is the control both of them can point at.

A named, testable control

Not a promise — a configured fallback chain with a fail-closed residency boundary, the way backups and MFA are named controls. Something an underwriter can point at.

Evidence, not assurances

Every routing decision and failover is metered and logged. Continuity isn't a slide — it's an audit trail you can hand to a risk committee or a regulator.

Lower correlated risk

When a whole economy depends on one foreign vendor, the loss is correlated and uninsurable at scale. Sovereign aggregation breaks that single point of failure — at the firm and at the nation.

⬢ Separation of layers is the regulation taking shape. Your application talks to a capability — never directly to a vendor's model. That seam is what lets the model layer be swapped, kept in-jurisdiction, audited, or failed over without your application knowing or caring. A clean boundary a regulator can stand on — and sovrgn is that boundary.
How it actually works

Ask for a capability. We own the vendor.

Under the hood it's the Sovereign Routing Protocol — and it speaks plain OpenAI. Keep your SDK, your code, your tooling. Set model to a capability and sovrgn resolves the right provider behind it, under your jurisdiction's residency rules. The capability is the contract; the vendor is ours to manage — so switching providers, or losing one, never touches your code.

smart

Frontier reasoning — the strongest model for the job.

fast

Low latency — fast, cheap, good enough at scale.

sovereign

In-jurisdiction residency — routes in-country, or fails closed.

Works withOpenAI SDK · LangChain · LlamaIndex— anything that already speaks OpenAI
# Point at your country's instance — nothing else changes.
from openai import OpenAI

client = OpenAI(
    api_key="sk_...",
    base_url="https://au.sovrgn.ai/v1",
)

resp = client.chat.completions.create(
    model="sovereign",            # capability, not a vendor
    messages=[{"role": "user", "content": "Hello"}],
    extra_headers={"X-Sovereign-Residency": "au-only"},
)

print(resp.choices[0].message.content)
It works like an electricity market

Firm your inference, or float it.

Power grids have priced this for a century: firm capacity is reserved, guaranteed and can't be curtailed; interruptible capacity is cheaper because it yields when firm load needs it. Inference is now priced the same way — and sovrgn lets you choose per call, under your jurisdiction.

⬢ Firmed inference

Reserved. Guaranteed. Yours.

Committed, in-jurisdiction capacity with guaranteed throughput and an SLA — the baseload your critical paths run on. It can't be bumped when the market tightens. The price you pay for certainty.

Grid analog: firm supply — agreed capacity, highest dispatch priority, can't be curtailed except under unforeseeable conditions (71% of US power-plant gas, 2016).
Real today: AWS Bedrock Provisioned Throughputhourly commitment · guaranteed units
◇ Floating inference

Spot-priced. Deferrable. Cheap.

Batch and off-peak work that rides spare capacity at a discount — and steps aside when firm demand arrives. Run it overnight, on the cheap side of the planet, while the sun is somewhere else.

Grid analog: interruptible supply — lower priority, short-term, curtailable, priced below firm to reflect the risk of being bumped.
Real today: Bedrock batch −50%Bedrock Flex −50%Anthropic Batch −50%

The market mechanics are the real ones: merit-order dispatch (cheapest capacity first), locational pricing (price tracks where the work lands), and a day-ahead/real-time split — firm reserved ahead, spot cleared on a ~5-minute cadence. Provider tiers and discounts cited are vendor-published and current; they move.

The continuity machinery

We don't just route. We guarantee the route.

"Plan B" only means something if it engages without you. The 2024 Red Sea cable cuts made the lesson physical: routes fail, not just vendors. These are the three mechanisms underneath every sovrgn instance — the parts that turn a bad day for a provider, a cable, or a jurisdiction into a non-event for you.

Automatic failover

Every capability resolves an ordered chain of providers. A retryable upstream failure reroutes to the next candidate transparently — one inbound call, one billed unit.

Circuit breaker

A per-provider breaker trips open after consecutive failures, skips the target that's down, then half-open probes before closing. A bad provider day never becomes your bad day.

Fail-closed residency

Send X-Sovereign-Residency: au-only and sovrgn routes only in-country. No silent offshore fall-through — if nothing in-jurisdiction can serve it, you get a clean, structured rejection.

⬢ Sovereign by definition, not by configuration. The terminal node of every in-jurisdiction chain is a national sovereign-silicon backstop — accredited against the SRP standard. AloomU is Australia's. Each nation plugs in its own.
Straight answers

What buyers ask us first.

Is sovrgn a proxy?

Architecturally, yes — and that seam is the product. Your application calls your country's sovrgn instance; the instance resolves a capability to a concrete provider under your jurisdiction's rules, with failover, residency enforcement, and metering happening at that boundary. Your code never talks to a vendor directly, which is exactly what makes vendors swappable.

What's the latency overhead?

The routing decision is in-memory and adds single-digit milliseconds to calls whose inference time is measured in hundreds of milliseconds to seconds. The fast capability exists for latency-sensitive paths, and failover only engages once the primary has already failed — the happy path stays one hop.

What happens during a failover?

From your side: nothing. A retryable upstream failure reroutes to the next provider in the capability's chain within the same inbound call — one request, one billed unit, however many attempts it took. A provider that keeps failing trips a circuit breaker and gets skipped until a probe confirms it's healthy again.

What does it cost?

Usage-based — you pay for what you route through it, and commercial terms are set per jurisdiction as each instance opens. We won't invent a pricing table before the terms are real: ask us and we'll tell you exactly where pricing stands for your country.

Which providers does it work with?

The protocol speaks plain OpenAI, so anything that can point an OpenAI SDK at a base_url works unchanged. Behind the seam, capabilities resolve across the provider market — OpenAI, Azure, Anthropic and others — ending in the national sovereign-silicon backstop for in-jurisdiction chains.

Am I locked in?

The opposite is the point. You code against capabilities, not vendors, so you leave the same way you arrived — change one base_url back. Your prompts, your SDK and your code never change either way.

What if nothing in-country can serve a request?

With residency pinned (X-Sovereign-Residency), you get a clean, structured rejection — fail-closed. Your data never silently crosses a border to keep a request alive, and the refusal itself is logged: auditable evidence the control works.

Every business needs a Plan B. So does every economy.

Pick your country and step into its sovereign instance — the fallback that keeps you running, keeps your data home, and that insurers will one day require.

Or talk to a human — tell us what you need to keep running, and we'll reply directly.