Compute and facilities
- Vendor financing and structured leases for GPUs and networking gear.
- Project‑like SPVs combining offtake with long‑term hosting contracts.
- Private credit for gap financing; convertibles for growth balance sheets.
Industry Research
Capital pathways for AI infrastructure
Assessing equity, debt and partnership models enabling hyperscale build‑outs across compute, data and application layers.
Financing the AI stack
Data center expansion, power procurement and model commercialisation each require different capital structures and risk allocation. We outline instruments and counterparties playing a role in the current cycle.
Power and sustainability
- Virtual PPAs and tolling agreements to secure baseload at predictable prices.
- Hybrid equity for on‑site generation and transmission upgrades.
- Green financing frameworks aligned with disclosure requirements.
What this means for operators
Capital intensity and technology cycles demand balanced structures that protect downside while preserving upside participation.
For hyperscalers
Blend long‑dated leases, tax equity and supplier commitments to hedge supply risk while maintaining flexibility across GPU generations.
For emerging platforms
Consider alliances with utilities and real‑estate partners; keep leverage at levels consistent with utilisation ramp and pricing uncertainty.