Ex-Microsoft Engineer Details Azure's "Trillion-Dollar" Legacy of Rushed Launches and Talent Exodus
Axel Rietschin, a former Azure Core Compute engineer who spent a year on Azure and eight years on the Windows Base Kernel team, published the fourth installment of his Substack essay series detailing what he calls "how Microsoft vaporized a trillion dollars." His account describes a company that rushed Azure to market in 2008 to compete with AWS, then spent years fighting fires caused by "rushed decisions and wishful thinking." Rietschin argues the "post-launch talent exodus," lack of software quality discipline, and poor architectural vision created "a sophisticated system perpetually on life support." He points to OpenAI's $11.9B compute deal with CoreWeave in March 2025 as a "vote of no confidence" in Azure's ability to meet demanding AI workload requirements at scale — and notes Microsoft laid off roughly 15,000 people during May–July 2025. The credibility of these accusations comes from Rietschin's insider perspective. Having worked on both Windows Kernel and Azure Core Compute, he witnessed firsthand the architectural decisions and engineering culture that shaped Azure's development trajectory. His description of a system "perpetually on life support" resonates with reports from enterprise customers about Azure's reliability challenges and the complexity of managing large-scale Azure deployments. Federal cybersecurity evaluators reportedly dismissed Microsoft 365 Government Community Cloud High (GCC High) as inadequate in 2024, and ProPublica documented government dissatisfaction with Azure security compliance. These external validations lend weight to Rietschin's insider account, suggesting his critique isn't just isolated complaints but part of a broader pattern. The timing of these revelations matters coming shortly after Microsoft's announcement of MAI models and their increased focus on AI self-sufficiency. If Azure's infrastructure is indeed as fragile as Rietschin describes, it would explain Microsoft's urgency to build alternative compute solutions and reduce dependence on both Azure and OpenAI simultaneously. *This is the most credible insider account of Azure's structural problems I've seen. It explains a lot about why Microsoft is so aggressively building MAI models and hedging its OpenAI dependency — they may not have a choice.*