From XaaS Boom to GenAI Bust? The Changing Fortunes of “Everything-as-a-Service”
The Rise of XaaS and Why It Dominated
Over the past two decades, “XaaS” (Everything-as-a-Service) became the default model for software delivery. Instead of selling boxed software or one-time licenses, companies embraced cloud-based subscriptions for Software-as-a-Service (SaaS), platforms, and infrastructure. This shift was driven by several advantages: customers got instant access to updates and cloud storage, while vendors enjoyed steady recurring revenue and stronger piracy protection.
For example, once Adobe moved its Creative Suite to Creative Cloud subscriptions in the 2010s, users could no longer simply pirate Photoshop; the software’s value was tied to online services and updates. Likewise, enterprise software like Salesforce CRM was only accessible via the cloud, effectively eliminating piracy by design. The cloud also made global distribution easier and lowered upfront costs for customers. This convenience and the promise of automatic backups and collaboration features helped XaaS offerings steamroll traditional software models. Privacy prevention was the primary driver initially, but the subscription model had so many other staggering advantages that XaaS spread far beyond software to revolutionize entertainment, manufacturing, professional services, and numerous other industries.
Several trends converged to make XaaS the dominant paradigm, from the maturing of Cloud computing infrastructure to the spikes in broadband coverage. This “rent vs. own” mindset enabled another tailwind in instantaneous global reach. With hindsight, it was not a surprise that by the late 2010s, XaaS was king. Enterprise IT budgets pivoted to cloud-first, and consumers grew accustomed to paying monthly for apps and services. The model was so successful it spread to almost “anything-as-a-service” – from platforms (PaaS) to infrastructure (IaaS) and beyond. Even hardware-heavy businesses started selling outcomes or usage instead of products (e.g., “device-as-a-service”).
The Great Unbundling: Breaking Apart Products and Services
In the XaaS era, a significant trend was the unbundling of products and services. Unbundling refers to breaking down integrated offerings into separate components, each of which can be provided as a service on a subscription basis. Prominent examples in software include companies like Dropbox, which expanded from a simple file storage service to offering additional features like Dropbox Paper and Dropbox Sign as separate subscriptions. Salesforce unbundled its CRM functionalities into distinct clouds such as Sales Cloud, Service Cloud, and Marketing Cloud, allowing customers to subscribe to specific modules.
This unbundling extended beyond technology and entertainment, reshaping everything from financial services to retail. Traditional banking functions were separated into specialized fintech offerings, with companies focusing exclusively on payments, lending, investing, or other previously integrated services. Car ownership was unbundled in transportation into car subscription services and subscriptions for features like driverless modules. Retail followed suit, with specialized subscription boxes replacing the department store experience. Unbundling offered compelling business advantages: one-time selling, higher profit margins on specialized services, more targeted marketing opportunities, and the ability to capture different market segments with varying price points.
Unbundled AI Selling Is Not Working
Today, however, the very factors that made XaaS dominant are starting to weaken. We are in an era of instant copiability, where new software features or even entire products can be replicated almost immediately by others. This is a massive departure from the previous era of software – traditionally, a killer feature or a patented algorithm could buy a company years of market leadership. Now, if a SaaS app unveils a cool AI-powered feature on Monday, a competitor (or an open-source community) can have a comparable feature out by Tuesday. This dynamic erodes one of the SaaS model’s key defenses: uniqueness.
Lower switching costs compound this effect. SaaS vendors historically relied on a bit of friction – moving all your data from one platform to another was non-trivial, discouraging churn. But AI is even chipping away at that. With no lock-in by file formats or data wrangling, SaaS customers can jump ship more quickly if they find a better or cheaper tool. Generative AI can write the “glue code” or adapters to move data between services, undermining the old strategy of trapping users with proprietary formats.
Finally, it is the revolution brought about by open source. The explosive rise of open-source AI models has profoundly disrupted standalone AI selling models. Meta's release of Llama models, Mistral AI's approach to free base models, and hundreds of other open-source initiatives have created a marketplace where fundamental AI capabilities are increasingly free and accessible. This has left pure-play AI companies scrambling. Anthropic's Claude and OpenAI's ChatGPT continue releasing new capabilities at breakneck speeds yet struggle to justify premium pricing when comparable open alternatives proliferate.
The rising tide of open source capabilities is creating unprecedented pressure on established SaaS giants like Salesforce, Adobe, and Intuit, as their once-proprietary features increasingly face free alternatives with comparable functionality. Adobe's creative tools now compete with open-source alternatives like GIMP and Inkscape, while Salesforce faces challenges from open CRM solutions that offer core functionality without subscription fees. When Adobe introduced generative AI features in Photoshop, open-source alternatives like Stable Diffusion quickly offered similar capabilities without subscription costs. From Zoom to Slack, or even earlier disruptors like Jasper AI or Stable Diffusion, are feeling pressure from open-source, non-subscription new competitors. Intuit's TurboTax, long a subscription staple for tax preparation, now faces competition from government-backed free filing initiatives and open tax preparation tools.
While all the software giants are battling with their unique AI-enabled products and features, one trend is clear: charging separately for GenAI is generally not working. Microsoft’s Copilot is perhaps the best example, followed by Adobe's efforts.
Signs of Saturation: Slowing Growth for SaaS Giants
The consequences of these shifts are starting to show up in the numbers. After years of heady growth, many leading XaaS companies are seeing revenue growth deceleration, even as they continue churning out new features, and in the earlier era, introducing significant new features guaranteed new revenue streams and growth through a higher number of users. Even as SaaS companies are announcing features at a pace never seen before, their growth has stagnated, if not started going in the negative territory.
Shifting Strategies: From Direct-to-Consumer to B2B, Partnerships and AI-Bundling
One does not need Sam Altman’s recent statement, like, “We have been on the wrong side of history here and need to figure out a different open source strategy,” or the recent Baidu’s recent decision to make its chatbot free to realize how everyone is pivoting to find new business moats. DeepSeek is not only breaking new ground in what it means to be an open source but also in partnerships. In a few short weeks, it has unveiled the successful execution of a strategy of integration with leading cloud platforms rather than exclusively offering standalone subscription services. Baidu Cloud, Alibaba Cloud, Tencent Cloud, and Huawei Cloud have all integrated the DeepSeek AI model into their platforms. This approach differs from the premium feature subscription model pursued by OpenAI and Perplexity. Chatbox as a Service will survive in some form but will unlikely live up to the earliest promises.
At Google, X, and Amazon, AI capabilities are becoming integral to current services like Google Drive, X Premium, and Amazon Prime. They all aim to charge more for extensive usage. Still, their approach is distinct from the one adopted by not only the pure GenAI chatbox providers but even those like Microsoft and Adobe, who had none of the advanced AI-related features with existing product lines.
Perhaps the most striking example of this potential shift comes from BYD, which announced it would offer cutting-edge self-driving systems on all its vehicles—including its cheapest Seagull model, priced at just $9,555—for free. This approach starkly contrasts Tesla's strategy of charging $8,000 for its Full Self-Driving software in the United States or $99 monthly on a subscription basis.
Conclusion: AI as An Invisible Layer Like Electricity
As things stand, GenAI is likely to become an invisible, ubiquitous layer, not a standalone category. Just as electricity is built into most products rather than sold separately to consumers by GE or Siemens, AI will infuse everything. The money might not be in selling “an AI” but in selling whatever complementary goods or services leverage that AI. One might be able to do this by making better devices like vacuum cleaners, cars, or smartphones with on-device AI capabilities that provide unique selling points. Or, better, if one invents new products with new capabilities. Physical embodiment could become increasingly crucial for AI differentiation as we move further into the robotics era.
The history of selling shows a continuous evolution from product-centric to service-oriented approaches, with the XaaS model representing the pinnacle of the service-based paradigm. However, the GenAI era introduces new dynamics that may drive a partial return to product-bundled services, particularly for AI-powered capabilities. Firms that cling to the old playbook of locking users in with subscriptions without substantial improvements face certain disruptions, but even those with material improvements may have a lot to consider.
