Giant Pool of Money

Research evidence excerpt

Giant Pool of Money

s to have from contracts vs GPU pool (US$m)

a finite useful life of 5 years, after which we assume no further revenue. If there is demand for 600

500these chips after 5 years, there will be upside to our forecasts. 491

400 379

258 246 217

192 181 172Even more upside through gearing. We currently forecast MP1 to fund its future capex of ~ 300200

55$200m pa (mainly for GPU Pool) using operating cash flow. If MP1 decides to gear up, say to 100 31 0 0 114

2x Net Debt/EBITDA, FY28e EBITDA of $470m suggests MP1 could secure ~$940m in debt 0 FY26E FY27E FY28E FY29E FY30E FY31E FY32E

funding to purchase further chips and servers, translating to ~$590m in annualised EBITDA at Contracted Capacity GPU Pool

.

full chip utilisation (based on EBITDA payback periods of c.19 months for chips) and ~$740m Source: Company data, Jefferies

in ARR.

Figure 3 - Bloomberg Neocloud Rental Indexes

Risks to earnings. We acknowledge there may be counterparty risks with Latitude's unnamed (Compute Desk Neocloud Blackwell Index US,

Compute Desk Neocloud Hopper Index US,

customers for the fixed-term contracts. The GPU pool also presents contract length risk and Silicon Data H100 Rental Index, and Compute

exposure to spot GPU rental rates. These characteristics may limit lender's appetite to fund the Desk Neocloud Hopper Index EU)

projects. However, as the Compute division scales (with an ARR of A$680m for the contracts $6.00

and GPU pool announced thus far), and MP1 shows its ability to execute, MP1 may be able to $5.00$4.00

tap the debt market rather than rely on equity. $3.00

Earnings and valuation. We upgraded our EBITDA by 20% in FY28 to reflect our confidence that $2.00

the GPU Pool would experience strong utilisation (80%) in FY28 onwards. MP1 is trading on $1.00$0.00