Cloud platforms sell computing to applications for a price. However, by precisely defining and controlling the service-level characteristics of cloud servers, they expose applications to a number of implicit risks throughout the application’s lifecycle. For example, user’s request for a server may be denied, leading to rejection risk; an allocated resource may be withdrawn, resulting in revocation risk; an acquired cloud server’s price may rise relative to others, causing price risk; a cloud server’s performance may vary due to external factors, triggering valuation risk. Though these risks are implicit, the costs they bear on the applications are not. While some risks exist in all Infrastructure-as-a-Service offerings, they are most pronounced in an emerging category called transient cloud servers. Since transient servers are carved out of instantaneous idle cloud capacity, they exhibit two distinct features: (i) revocations that are intentional, frequent and come with advanced warning, and (ii) prices that are low in average but vary across time and location. Thus, despite enabling inexpensive access to at-scale computing, transient cloud servers expose applications to risks, the scale of which were unseen in the past platforms. Unfortunately, the current generation system software are not designed to handle these risks, which in turn results in inconsistent performances, unexpected failures, missed savings, and slower adoption. In this dissertation, we elevate risk management to a first-class system design principle. Our goal is to identify the risks, quantify their costs, and explicitly manage them for applications deployed on cloud platforms. Towards that goal, we adapt and extend concepts from finance and economics to propose a new system design approach called financializing cloud computing. By treating cloud resources as investments, and by quantifying the cost of their risks, financialization enables system software to manage the risk-reward trade-offs, explicitly and autonomously. We demonstrate the utility of our approach via four contributions: (i) mitigating revocation risk with insurance policy, (ii) reducing price risk through active trading, (iii) eliminating uncertainty risk by index tracking, and (iv) minimizing server’s valuation risk via asset pricing. We conclude by observing that diversity and asymmetry in the creation and consumption of cloud compute resources is on the rise, and that financialization can be effectively employed to manage its complexity and risks.