It used to be that every company faced a specific choice: buy or build. For most people, the answer was to buy. Software as a Service (SaaS) dominated because it was cheaper, faster, and easier than building proprietary software from scratch. But things have changed now as many tasks within SaaS companies and many features within SaaS products can now be completed by your AI.
Today, more companies are realizing that construction, once viewed as too expensive, is not only no longer viable, but the better option, offering more customization and significantly lower labor costs.
Klarna and IBM: Companies’ shift from business to AI has already begun
Look at Klarna. The fintech giant recently replaced large parts of its workforce with AI-driven automation. IBM did something similar, cutting hiring in roles like customer support agents, data entry clerks, financial analysts, software testers, and even entry-level developers, which they knew would soon be automated by AI. But it’s not just these notable examples.
This shift is happening across industries like finance, healthcare, retail, and manufacturing, driven by the simple fact that AI reduces the complexity of software development.
The main reason SaaS took off was that building software was difficult, requiring large teams of engineers, months (if not years) of development, and significant ongoing maintenance. Now, AI is dramatically reducing these barriers, making development easier while still requiring constant maintenance and updates. Companies no longer need to license software for every function because they can dynamically create their own AI-powered solutions.
For example, AI-driven automation is eroding the market value of Salesforce subscriptions as companies build their own custom CRM tools, reducing reliance on third-party SaaS providers. Reliance on third-party software vendors is diminishing, and organizations that realize this early are the ones moving forward.
The debate over new purchase versus construction
Purchasing SaaS made sense because it allowed companies to focus on their core business rather than wasting time building and maintaining internal tools. Salesforce, HubSpot, and Workday grew by convincing organizations that outsourcing software to specialists was a smart move. But artificial intelligence is changing the game. Now, construction is not only cheaper, it is better. Here’s why:
-
Customization: AI enables companies to precisely design software to fit their needs rather than forcing workflows onto rigid SaaS platforms. For example, companies like Netflix are leveraging AI to create highly personalized content recommendation engines rather than relying on off-the-shelf solutions.
-
Cost savings: Once AI-powered tools are set up, they can operate with minimal human intervention, reducing ongoing licensing and staffing costs. For example, Tesla has reduced its reliance on third-party software vendors by developing AI-based automation for manufacturing and quality control.
-
Faster iteration: Organizations can update and improve AI-driven tools on demand instead of waiting for vendors to release updates. Amazon is constantly improving its AI-driven warehouse and logistics management systems to improve efficiency without waiting for external software updates.
-
Data ownership: Companies no longer need to send valuable proprietary data to third-party vendors. Apple has focused on data privacy by developing its own AI models for on-device processing, reducing reliance on external cloud-based AI services.
For decades, companies have made peace with the limitations of SaaS, but AI reduces the need for those compromises.
Infrastructure advantage: who will win?
So, if AI makes custom software development easier and more attractive than SaaS, who benefits most? The answer lies in infrastructure. In the early days of the cloud, companies realized that managing their data centers was inefficient. AWS, Google Cloud, and Microsoft Azure have commoditized infrastructure and become the backbone of the Internet. Now a similar battle is taking place in the field of artificial intelligence.
The winning companies will be those that control the AI computing infrastructure. OpenAI, Anthropic, and other model providers are vying for dominance, but their fate ultimately depends on access to the real power brokers: those who control chips, power, and data centers.
Nvidia, Intel, and AMD own the GPU market. AWS, Google, and Microsoft rule hyperscale cloud computing. Companies that realize this are increasingly avoiding third-party AI providers and improving their own AI stacks. The closer they are to the infrastructure layer, the greater their influence.
SaaS vendors have a big problem
SaaS companies should be worried. The economic model that has made it dominant – recurring revenue through subscription licensing – is based on the assumption that building software is too expensive for most companies. This assumption is crumbling. Take enterprise automation tools like Zapier or UIPath. Five years ago, companies paid them exorbitant fees for automation.
Today, those same companies can use AI to write scripts that handle on-demand automation, tailored to fit their workflow. CRM platforms, enterprise resource planning (ERP) systems, and customer service software face the same challenge: AI enables companies to replace generic, expensive SaaS solutions with custom-built, AI-powered tools that are cheaper and more effective.
This is not just a small shift. It’s an existential crisis for SaaS vendors who have spent years building walled gardens around their platforms. When companies no longer need to play within those walls, the foundations of the SaaS business model begin to erode.
The new playbook for enterprise software development
Forward-thinking companies are already adapting. They reduce reliance on third-party SaaS vendors and invest in internal AI-powered tools. Here’s what a winning strategy looks like:
-
Control account: Companies are moving from renting AI services to controlling their own computing resources. For example, OpenAI has begun building its own AI supercomputers to reduce reliance on external cloud providers.
-
Indigenous development of artificial intelligence: Instead of SaaS licensing, organizations use AI to dynamically create and maintain their own software. Shopify, for example, has implemented AI-driven code generation to simplify backend development and reduce reliance on third-party software solutions.
-
API-First Thinking: Companies are integrating AI into their core workflow through APIs rather than relying on bloated SaaS interfaces. Stripe has invested heavily in AI-powered financial APIs to enhance fraud detection and transaction efficiency.
-
Security and compliance: By keeping AI workloads in-house, organizations avoid the compliance risks of sending sensitive data to third-party SaaS platforms. JPMorgan Chase has developed proprietary financial analytics platforms based on artificial intelligence to maintain data security and regulatory compliance.
The smartest companies aren’t just replacing SaaS, they’re designing their entire tech stacks around native AI architectures. The ultimate goal is not simply to reduce costs; It’s agility, security and ownership.
AI for SaaS consumption and beyond
SaaS won’t disappear overnight, as the global SaaS industry’s market cap currently exceeds $3 trillion, but its dominance is declining. AI is changing the economics of software development, making in-house solutions not only viable, but preferable. Companies that realize this now will gain an advantage. But beyond just taking SaaS market share, AI is expanding the overall market by creating efficiencies that reduce work hours and open up new economic opportunities.
Instead of simply shifting money from software subscriptions to internal AI-enabled solutions, companies are using AI to improve operations, automate routine tasks, and augment human capabilities. This transformation not only reduces operational costs, but also increases productivity, allowing companies to scale faster and compete more effectively.
Tesla has developed AI-powered internal automation to simplify manufacturing, JPMorgan Chase has built its own AI-driven financial analytics platforms, and Netflix has created its own AI-powered content recommendation system, each reducing reliance on third-party SaaS providers.
Those who continue to rely on SaaS vendors may find themselves locked into solutions that are more expensive and less adaptable. The lesson of past technological revolutions is clear: real power belongs to those who control the infrastructure. Just as cloud computing has reshaped enterprise IT, AI is reshaping enterprise software. The winners won’t be yesterday’s SaaS providers. These will be companies that own their own AI stacks, control their computing, and move faster than the competition.
If you’re still debating buy versus build, the answer is already clear: AI makes building the better option.