Instead of investing in speculative AI applications, we invest in the essential computing infrastructure that every AI company requires — regardless of which applications succeed.
📈
Goldman Sachs projects AI compute demand will exceed supply through 2030. Microsoft has $80 billion in unfulfilled Azure orders — GPUs sitting idle in warehouses because they cannot find enough power.
🏗️
Hyperscalers cannot satisfy enterprise demand, creating opportunity for emerging "neoclouds." Power transformer lead times have stretched to 128 weeks — a physical constraint that widens the supply gap.
⏱️
Enterprise GPUs maintain economic viability for 6-8 years. NVIDIA chips show 5-year peak popularity cycles. We conservatively underwrite 6-year life and target return of capital in under 3 years.
| GPU I LP | AI Stocks / ETFs | Data Center REITs | |
|---|---|---|---|
| Returns driven by | Contractual lease revenue | Market sentiment | Share price + dividends |
| Correlation to S&P 500 | None | High | Moderate |
| Hard asset collateral | Yes — own the servers | No | Indirect (building, not equipment) |
| Capital return timeline | ~3 years | Unknown | Perpetual hold |
| Cash flow visibility | Multi-year contracts | Quarterly earnings | Long-term leases |
"We've entered the virtuous cycle... $1 trillion in orders through 2027."
— Jensen Huang, CEO, NVIDIA (GTC 2026)
"This isn't some sort of quixotic top-line grab."
— Andy Jassy, CEO, Amazon (Q4 2025 Earnings)
"GPUs are sitting idle in warehouses because we can't find enough electricity."
— Satya Nadella, CEO, Microsoft
This website is for informational purposes only and does not constitute an offer to sell or solicitation to buy securities. GPU I LP interests are offered only to accredited investors pursuant to Regulation D under the Securities Act of 1933. Past performance is not indicative of future results. All projected returns are forward-looking and subject to risk.