10
Loading video...

By Paul Brody
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
The SaaSpocalypse is real but perhaps a little exaggerated. CIOs and corporate software buyers are threatening vendors with vibe-coding, but it’s much more about getting a better price than it is about actually switching.
The idea is appealing: get an agent to replicate your existing solution and run it in-house for a few hundred dollars in tokens and ditch the costly software license agreement. Never talk to another software salesperson again.
So far, it’s not working. Yet. One survey showed that 35 percent of teams have replaced at least one enterprise application with something they vibe-coded, but this is most often in categories like workflow automation and business analytics—these are additions and don’t point to actually cancelling the entire software license. Industry revenue retention—the rate at which customers renew—has fallen slightly, but there’s no agreement on the cause and earnings haven’t been significantly affected.
Read More on Essays
Having a vendor is useful because the true cost of vibe-coding software is much higher than a few hundred dollars in tokens. For critical enterprise systems, you need to build and hold on to expert human resources to guide the development and maintenance of new solutions, and you’ve got to manage all the traditional complexity of moving from one solution to another.
And never, ever, underestimate the appeal of having someone to blame if things go wrong.

...
So if vibe-coding is more costly and complex than it at first appears, why is the SaaSpocalypse still gaining steam? The answer is that AI is radically reshaping the software business itself and the result is going to be a flood of new entrants. CIOs won’t need to vibe code new solutions because incumbent software vendors are about to face a tsunami of vibe-coded competitors, all coming up with competitive solutions and rock-bottom pricing.
Traditional enterprise software companies are costly to build and maintain. In addition to teams of developers, as the software grows, maintaining it becomes a monumental challenge as software bloats with features only one client uses. Throw in a huge sales force and a high cost of sales and you end up with products that have multimillion-dollar price tags for enterprise customers that use a fraction of the functionality available.
This can work when you need scale to survive, but AI is changing just about everything in the software business. First and foremost, you can build and maintain software at a much lower cost than in the past. But the transformation doesn’t stop there. AI is also radically lowering the barriers to finding and integrating software into the enterprise.
Once you have a product that does its specific function, say procurement or logistics planning, the two biggest historical challenges when it comes to getting new software into companies have been integration and user acceptance.
Related Story

Exclusive: China Urges US Cooperation on AI to Avoid Human Extinction
Pulling data out of one system, say customer addresses from order entry, and putting data into another system, say logistics, was not trivial. Just to get lists of customer addresses synchronized, you needed to hire expensive consultants to build a custom integration. A rule of thumb in the early days of development and implementation was $1 of consulting spent for $1 of software license.
Now, while many companies might not be comfortable vibe-coding a whole software application, giving integration tasks like this to AI agents looks straightforward: ETL tasks (extract, transform, load) are simple, finite tasks that are easily verified. In other words, a perfect, controlled and contained task for AI.
The second big obstacle has always been user acceptance—will company staff accept a new system? Change is hard and I’ve seen tears and shouting at meetings when new software has been introduced. Again, AI is transformational because the answer, increasingly, is that nobody is going to actually use this new software. More than one CIO has told me they don’t want to use any new software. It’s getting integrated into existing systems and hidden behind AI chatbots.
The era of building custom charts with specialized commands is over. The future is plain-English commands for AI agents: “Show me a map of product distribution requirements this month.” Specialized software will all ship with AI-agent friendly access points—documented APIs, Command-Line Interfaces and Model Context Protocol (MCP) servers that make it simple for AIs to understand and use the software.
Once you get past the big three questions—does your software work, can we integrate it and will anyone agree to use it—there are still a host of challenges for nimble new firms, but they may not be as insurmountable as they look.
So far, there’s not much evidence of a vibe-coding SaaSpocalypse if you look at software company earnings. However, it’s only in the last year that AI agents have become good enough to start building relatively sophisticated enterprise software. It takes time for developers to absorb these changes, quit their jobs and start new companies.
The tsunami of new software start-ups is still out in the ocean gathering strength. Enterprise adoption is also slow—just because you have a better product doesn’t mean that everyone has time to adopt it right away. Pilot customers can take months to find and months more to start testing, so it can be 12-18 months between when a start-up gets founded and when they show up in public with an initial product and some early reference customers.
Bigger software companies are not going to sit still, but they face a much bigger transformational challenge. Change takes time, and when it comes to organizational change, subtraction is much harder than addition. Traditional software firms will also have to carefully balance price cuts with the rate at which their customer base is changing, something that could prove difficult.
At the earliest, it will be another year or more before anecdotes about dramatic price cuts and threats of vibe-coding turn into stories about displaced incumbents and falling earnings. One area that probably won’t be affected is core enterprise systems. These big engines (ERP systems) from vendors like SAP, Microsoft and Oracle, are the systems of record for the company’s core financial performance are going to be treated much more carefully and conservatively by CIOs.
Connected to these core systems, however, are hundreds of specialized solutions from customer service to procurement to logistics. These systems will be the first to face the new competitive onslaught. The evidence will start to show in late 2027, and by 2028, the market for enterprise systems will look very different.
Paul Brody is the founder and CEO of Nightfall Networks, a blockchain technology company focused on building robust privacy systems for enterprise users.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines