$IREN is building a layered capital structure in which each completed project increases the company’s borrowing capacity for the next wave of growth. The first layer of that model is already visible. For the Microsoft contract, IREN said the GPU financing plus the customer prepayment covers about 95% of GPU-related capex, and that the facility is expected to be secured by the related GPUs and contracted cash flows. That shows the compute layer is increasingly financed at the asset and contract level rather than through equity. The second layer begins once construction risk is removed. After a data center shell starts generating stable cash flow, it becomes a fundamentally different credit asset than it was during development. At that point, IREN can borrow against the infrastructure itself, whether at the project level, portfolio level, or eventually corporate level. In the same Q2 FY26 materials, IREN explicitly said its financing workstreams include GPU financing, data center financing, and select corporate-level initiatives. Management is clearly building this stack in layers. The third layer is where the model compounds. As IREN’s asset base grows and cash flows diversify across more tenants, sites, and geographies, its credit profile strengthens. A broader and more diversified pool of hard assets and recurring revenues commands better financing terms than a single project standing alone. In other words, a larger and more diversified IREN gets to refinance and raise debt on better terms over time than a smaller, more concentrated IREN. That is the core insight: debt capacity is not static. It expands as the platform matures. First, debt is secured by GPUs and contracted project cash flows. Then, debt is layered onto completed and stabilized infrastructure. Later, broader corporate or portfolio debt is supported by the combined cash flows and hard assets of the platform. So each completed project does not just add revenue. It adds borrowing capacity for what comes next. IREN is essentially arbitraging cost of capital across different layers of the stack, using its vertical integration and growing asset base to move from narrow, asset-specific financing toward broader, lower-risk platform-level debt over time. This is also where the $6 billion ATM fits best. The ATM should not be viewed only as a dilution threat. In IREN’s case, it is optional firepower, not urgent need, especially given that the company is already cash flow positive and clearly has access to other forms of capital. Within this framework, a large unused ATM still strengthens the financing machine. Even if it is not tapped immediately, it increases balance sheet flexibility, improves counterparty confidence, and strengthens IREN’s hand when negotiating the next layer of debt. Lenders like knowing the borrower has an equity backstop available, even if management intends to use it only opportunistically. Put simply, IREN is building a self-reinforcing capital flywheel. It finances the compute layer with customer-backed and asset-backed structures. It then de-risks the infrastructure layer and adds debt against stabilized assets. As the portfolio matures and diversifies, it gains access to broader and cheaper corporate or portfolio debt. The ATM sits on top of that structure as optional equity capacity, supporting negotiations and giving management flexibility to issue shares selectively when the cost of equity is attractive. In that model, the same equity dollar funds multiple growth cycles instead of being trapped inside a single project. That means lower dilution than the market assumes and better returns on equity over time as mature assets are refinanced into new growth. If management keeps turning contracted GPUs into asset-backed debt, completed shells into refinanceable infrastructure, and a growing asset base into lower-cost platform-level borrowing, then IREN’s capital structure becomes one of its biggest competitive advantages.