Loading...
Cloud's list price: maximum flexibility, zero commitment, highest hourly rate.
Quick Definition
On-demand instances are cloud compute resources billed by the second or hour with no upfront commitment. They offer maximum flexibility and are ideal for unpredictable or short-lived workloads, but carry the highest per-unit price compared with commitments or Spot.
On-demand instances are cloud servers paid for by the hour or second at the provider's list price, with no commitment in either direction. Start one whenever you want, stop it whenever you want, pay only for the time in between. It is the default way everyone begins using the cloud.
On-demand is the price ceiling against which every other option is measured. Reserved instances and savings plans discount it 30 to 60 percent in exchange for commitment; spot instances discount it further in exchange for interruptibility. A mature spend profile uses on-demand only where its flexibility is genuinely needed: spiky traffic, new workloads with unknown patterns, and the buffer above committed baselines.
Example. A company reviews its bill and finds 85 percent of spend on-demand, although two thirds of usage has been flat for a year. Committing that stable baseline cuts the bill by a quarter; the variable remainder stays on-demand, where it belongs.
Persistently high on-demand share is one of the clearest savings signals in any account. The Cloud Cost Management Guide covers building a purchase-option mix, and the FinOps KPIs guide covers tracking coverage.
Unpredictable or short-lived workloads, brand-new services without usage history, and the variable layer above your committed baseline.
Typically 30 to 60 percent above committed rates and several times spot rates, for identical hardware.
Mature teams often land around 20 to 40 percent, with stable usage committed and fault-tolerant work on spot.