NeuralPool is raising capital to acquire and build AI data centers in geographically sovereign locations — physical infrastructure positioned outside the reach of the geopolitical risks threatening centralized hyperscale compute.
This isn't just a bet on AI — it's a bet on where the world is heading and who will control its most critical infrastructure.
Whoever controls AI computing power controls the future — economically and strategically. Today that power is concentrated in the hands of a handful of hyperscalers, almost entirely inside the United States.
As the world shifts from a unipolar to a multipolar order, the massive hyperscale data centers underpinning the American economy and military become high-value targets — for cyberattack, and potentially physical disruption.
NeuralPool is building some of the first AI data centers in South America — politically neutral territory, outside the direct line of geopolitical conflict, with access to abundant low-cost energy and growing connectivity.
If first-world data centers are ever taken offline — through conflict, regulation, or attack — NeuralPool's distributed, internationally located infrastructure could become some of the most valuable computing real estate on earth.
The opportunity in plain terms:
NeuralPool is acquiring physical land, power contracts, and GPU hardware to stand up its own AI data centers. Capital from this raise goes directly into facility buildout — power infrastructure, racks, cooling, networking, and GPU procurement — not into marketing or speculative software.
Early investors aren't just buying exposure to AI infrastructure growth — they're positioning for a scenario where resilient, geographically sovereign compute becomes irreplaceable.
Every dollar raised is allocated to physical infrastructure and the operations needed to bring it online.
| Phase | Site | Target Capacity | Status |
|---|---|---|---|
| Phase 1 | South America — Site A | Pilot facility · ~1MW | Site selection |
| Phase 2 | South America — Site B | Production · 5–10MW | Planned |
| Phase 3 | Additional jurisdictions | Multi-site, 25MW+ | Planned |
NeuralPool's owned facilities sell GPU capacity directly into a global market for AI inference and training — at margins that decentralized supply alone can't reach.
Owned racks rent GPU hours to AI labs, agencies, and developers in USDC and fiat. Owned hardware means owned margin — every hour a card is utilized is direct top-line revenue back to the company.
The same hardware powers a metered inference API for hosted AI models — recurring usage-based revenue from developers who never touch a GPU directly.
Long-term capacity reservations with enterprise customers — predictable cash flow underwriting the cost of additional buildout.
Building in public with a clear, phased rollout.
neuralpool.io live · GPU mining + rental software stack · operational dashboard · early supply-side network
Strategic investor conversations · South American site evaluation · power and land term sheets
First owned facility energized · initial GPU racks installed · live capacity sold into NeuralPool network
Multi-MW production facility · enterprise SLAs · long-term capacity contracts with AI customers
Additional jurisdictions · redundant geographic footprint · institutional compute partnerships
Centralized GPU clouds and existing decentralized projects leave the table unset for us.
Lower-cost infrastructure in jurisdictions with cheaper energy and land. No account lockup, instant provisioning, geographically distributed with no single point of failure.
Decentralized GPU networks rely on consumer hardware that can disappear at any moment. NeuralPool combines that long tail of supply with owned, enterprise-grade facilities that customers can actually contract against.
Hyperscalers are doubling down inside high-risk regions. NeuralPool starts from the opposite premise — geographic diversification first, scale second. Resilience is the product.
We're having conversations with strategic investors, infrastructure partners, and anyone who sees what we see in the future of sovereign AI compute. We respond within 48 hours.
This page is informational and does not constitute an offer to sell securities. Any investment offering will be made only to qualified investors under applicable exemptions and through formal documentation. Projections are not guarantees of future results.