Trusted computing is gaining an increasing acceptance in the industry and finding its way to cloud computing. With this penetration, the question arises whether the concept of hard-wired security modules will cope with the increasing sophistication and security requirements of future IT systems and the ever expanding threats and violations. So far, embedding cryptographic hardware engines into the Trusted Platform Module (TPM) has been regarded as a security feature. However, new developments in cryptanalysis, side-channel analysis, and the emergence of novel powerful computing systems, such as quantum computers, can render this approach useless. Given that, the question arises: Do we have to throw away all TPMs and loose the data protected by them, if someday a cryptographic engine on the TPM becomes insecure? To address this question, we present a novel architecture called Sustainable Trusted Platform Module (STPM), which guarantees a secure update of the TPM cryptographic engines without compromising the system's trustworthiness. The STPM architecture has been implemented as a proof-of-concept on top of a Xilinx Virtex-5 FPGA platform, demonstrating a test case with an update of the fundamental hash engine of the TPM.