Research

Research & perspectives

Notes from operating the market: where term paper is actually clearing, and how compute credit is getting structured.

CreditOctober 2026

Tier 2: Trading firms, revenue-stage AI apps, and the structure that gets them financed

Part 2 of the GPU Offtaker Credit Map: trading firms, AI applications, and the financing structures that make Tier 2 contracts bankable.

CreditOctober 2026

Tier 1: The benchmark deal

Eight private companies can back a multi-year compute contract on their own balance sheet. Who they are, how lenders structure the deals, and what the benchmark means for everyone else.

CreditSeptember 28, 2026

The GPU Offtaker Credit Map: 200 Private AI Compute Buyers, Ranked

A transparent framework for 200 private buyers of compute: 169 scored, 38 bankable with light or modest support, and what the tiers mean for structuring.

TradingSeptember 2026

The Tenor Trade

Buy capacity long, sell it short, and keep the difference if you can stay filled. What the trade pays, what you are underwriting, and how to hedge it.

TradingSeptember 2026

Hedging Compute Prices

If you buy compute, own GPUs, or lend against them, you are exposed to the price of compute whether or not you intended to be.

TradingAugust 2026

Compute Is Already Trading

The same GPU rents at three prices depending on how long you commit, and that ordering is a term structure people have already started trading.

CreditAugust 2026

Investing in the Middle Market for Compute

Why profitable, high-utilisation operators still can't get financed below the hyperscale tier, and what actually carries a compute credit structure.

SubstackAugust 2026

Public AI Credit Just Repriced. The Compute Market Is Next.

150 basis points moved in ten weeks without a single rate card changing, and the adjustment has already started.

Follow the desk

New research goes out on LinkedIn first.

Index prints, term structure, and how compute credit is getting structured. If you operate capacity or lend against it, that is the feed to be on.