The 'pay-as-you-go' pricing model is the correct answer because it allows the startup to avoid substantial upfront investments in physical hardware and data center infrastructure that are typical in on-premises environments. Customers pay only for the computing resources they consume, converting what would be capital expenditure (CapEx) into operational expenditure (OpEx). Global infrastructure pertains to the AWS's worldwide data centers and is related to global reach and low latency rather than upfront cost savings. 'Go global in minutes' is more about the rapid deployment and scale worldwide rather than cost. While resource optimization does help save costs over time, it's not directly related to minimizing initial expenses.