The software industry, traditionally built on the idea that creating software is expensive, time-consuming, and requires specialized skills, may be experiencing a major shift. This belief has long been used to justify high prices, subscription models, large teams, and high company valuations. However, as the cost of producing software declines, especially with the rise of artificial intelligence, the balance of power could be shifting from software publishers to their customers. A recent example from a major French e-commerce company illustrates this change. The company replaced a product discovery and recommendation solution — which previously cost around 360,000 euros annually in subscriptions — by rebuilding it internally using AI in just 48 hours. This case reflects a growing trend: more companies are evaluating whether it is more cost-effective to develop their own software rather than continue paying for external subscriptions. Annual Recurring Revenue (ARR), a key metric for Software as a Service (SaaS) companies, has been central to the industry’s growth for two decades. It helps establish company valuations and supports long-term planning. However, if customers increasingly choose to rebuild essential software functions themselves, the traditional SaaS model could face challenges. At renewal time, customers may come with prototypes, asking publishers to either offer something they cannot replicate or significantly lower their prices. The impact of this shift will likely vary by type of SaaS offering. Systems that are deeply integrated into a company's operations, such as enterprise resource planning (ERP) software, are less likely to be replaced quickly. In contrast, tools designed to automate specific tasks may be more vulnerable to disruption. Startups that offer autonomous solutions to niche problems could be particularly affected. As the industry evolves, the focus on value creation is becoming more important. Startups are being judged not only on the urgency of the problem they solve but also on the measurable value they provide and the uniqueness of their solution. If customers can rebuild significant parts of a product using AI, SaaS providers must re-evaluate their value proposition. Software publishers may need to move beyond simply adding AI features to their products and instead ask more fundamental questions: What would customers retain, remove, or continue to pay for if they could rebuild the product themselves? These insights could help shape new value propositions centered on services, data, interoperability, and new economic models. For Chief Information Officers (CIOs), the growing ease of building custom applications could lead to more strategic decisions about which tools are truly necessary. CIOs may increasingly act as architects of a capabilities portfolio, deciding whether to buy, build, assemble, or even delete certain applications, rather than managing a list of pre-packaged solutions. While this transformation may not happen overnight, it could lead to a gradual rethinking of how companies approach software. Starting with small tools, businesses may begin replacing subscriptions to save costs, prompting further scrutiny of the value of existing software. Governance, security, and maintenance challenges may slow some projects, but the trend could eventually lead to a rediscovery of the benefits of standardization and interoperability.