Where uptime meets
capital allocation.

Bankability is tested at underwriting; durability is tested years later, at renewal and refinancing. A financed data centre runs three contractual clocks — debt, lease, and the tenant's own customer contracts — that rarely align. Read through APAC's 1 GW quarter, Johor's doubling capacity, and the lease-tail maths, the risk is not whether a megawatt gets financed, but what the lease and tenant look like when the debt comes due.

APAC's data-centre pipeline is repriced by bankability, not demand. Introducing the Uptime Brief Bankable MW Framework: announced MW → technical MW → bankable MW → operating MW, with conversion set by five underwriting gates. Read through Korea's stalled Seoul pipeline, Ansan's powered-land deals, and the SGX REIT yield dispersion, the premium is shifting from plans to proof.

Oracle's credit spread eased when it added equity, then widened to a new high weeks later. The fix worked once and did not hold. Long-duration offtake is downstream of the funding window — and beyond technical and economic obsolescence sits a third: financing obsolescence, where the asset still computes but the debt behind it becomes harder to roll.

A US$35B deal financed Google's TPUs the way the market once financed aircraft and fibre — but it needed a vendor to stand in for the one thing compute lacks: a residual value the market can see. The deal did not need Broadcom because compute is now infrastructure-grade collateral. It needed Broadcom because compute is not — yet.

GPU rental prices are rising, but the real signal is in the financing underneath. Technical and economic obsolescence are diverging — and the risk has migrated from demand to asset life. Compute scarcity is the signal. The capital structure is the risk.

Two assets can look identical and be worth radically different multiples depending on whether the cash flows can be detached and sold. The moat may earn well. But you can't exit a feature.

When the Philippines grid failed in May, there was no intermediation layer operating at sufficient scale to buffer the disruption. The capital structures that close that gap are not interchangeable.

DFI capital finances the wires through utility balance sheets under sovereign guarantee. But technical openness and commercial openness are not the same thing — and the gap is where the next layer of value sits.

With 1,650GW queued across APAC grids, the entity that controls grid access controls the economics. Co-located generation and storage are no longer a premium feature — they are the workaround for a queue that no longer clears.

Hormuz repricing arrives at the meter as the Philippines declares a national energy emergency. The firming stack decision can't be deferred — and who controls baseload is now a live capital allocation input.

When a buyer acquires 1,100MW+ across three markets, they don't use the fragmented green electricity access layer. They effectively internalise the market function. The energy opportunity bifurcates — it does not collapse.

The fragmented procurement layer that mid-market buyers depend on is structurally at risk as the largest buyers begin bypassing it entirely.

How storage economics reshape the firmed clean energy value chain — and what the debt tenor and augmentation timing implications mean for project finance.

The largest renewable procurement in Australian history — and what the storage-backed structure signals about how hyperscalers are rewriting the energy access playbook.

Iran-linked mapping of Persian Gulf undersea cables as strategic pressure points — and what connectivity risk means for data centre underwriting in APAC.