Software vendors once promised flexibility through cloud and subscription models. Instead, many enterprises now face higher costs, complex billing, and shrinking control over their own systems. Rimini Street CEO Seth Ravin argues that this trend signals the end of traditional ERP dominance, and the rise of open, AI-driven architectures.
In an interview with Frontier Enterprise, Ravin discusses why enterprises are resisting forced upgrades, how vendor pricing models evolved into today’s “tollbooth” economics, and what the next era of enterprise software might look like.
Why aren’t SAP customers migrating to S/4HANA despite multiple deadlines?
The product has been around for a decade. People keep saying, “SAP says it’s new.” No, it’s like a car built in 2005 that’s been sitting on the lot. It may be new to you, but it’s still a 20-year-old car with 20-year-old technology. Why would you buy that car? That’s why tens of thousands of customers aren’t moving.
We’ve now entered the era of agentic AI for ERP. The transactions codified into ERP software over the last 30 to 40 years mirror the same business processes humans have followed since the dawn of organised trade. I sell you something, I issue an invoice, I collect payment, I pay employees, I manufacture goods, and I ship them. These are fundamental processes that have existed since the invention of the wheel. ERP software simply codified them through technology. Now, as we move into the AI age, and through our partnership with ServiceNow on ERP modernisation, we’re reconstructing these processes in an entirely new paradigm.
If you’re already doing this, as we are, why would you upgrade 20- or 30-year-old software that’s already outdated? What would you really gain from it? Would it reduce costs, increase sales, or change your business dynamics? No, it’s mainly a project for the large consulting firms — Accenture, Capgemini, EY — to stay busy. They encourage companies to pursue these upgrades to nowhere, because at the end of the day, you’re still replacing 20-year-old technology with another 20-year-old version.
Instead, we’re building new technology over the top of existing systems, advising customers not to waste money on upgrades. Spend it on innovation instead, and leapfrog to the next generation of technology. We’re developing processes that will run businesses in a completely new paradigm for the AI era.
Why are enterprises returning to custom-built solutions after years of standardisation?
Everything shifts; it goes one way, then the other. Everyone said, “Let’s go back to standard software because these customisations are hard to maintain.” That’s true. The more customised your system, the harder it becomes to move it forward in a packaged format.
What people often forget, though, is that those customisations are part of a company’s intellectual property — the reason it can outperform others. Take Hyundai, for example. We support them worldwide, across their full Oracle infrastructure in 100 countries. What makes Hyundai better than other carmakers? The company has a clear mission: to build cars of equal quality at a lower cost. Hyundai believes it can manufacture any car for US$2,000 less than anyone else. That’s their strength. To achieve this, they’ve invested heavily in a custom procurement system called Hyundai AutoEver, which they view as a critical differentiator in how they source parts and build cars. It’s entirely custom-built, not an out-of-the-box system. This is where custom development meets packaged services.
Now we’re creating custom agentic ERP processes to replace the legacy ones embedded in ERP systems today. It’s a complete transformation, and it’s genuinely exciting. I often tell people I haven’t had this much fun in years. I’m a former coder, a tech guy through and through. When the internet first emerged and software had to be redesigned for the web, the interfaces and user experience changed dramatically. People forget how different it was. Even simple things, like how the cursor moved and how screens were organised, were new. It was complicated in those early days, and few understood how to make it all work.
Will the software industry become more fragmented as innovation slows?
We’re already the largest third-party provider of VMware support in the world, and there’s a reason for that. Broadcom came in and said, “We’re not going to honour your perpetual licences. We’re not going to provide support for them.” So customers were told: either move and buy a new SaaS licence, or lose support altogether.
The reality is, these companies already have an asset that still works, and will continue to work for years. Yet they’re being told they won’t receive support unless they move to SaaS, start paying rental fees for something they don’t need, and accept price hikes of up to 10 or even 12 times.
They’ve turned to us. We offer support for years to come, and we provide security they don’t have.
How will billing complexity and lack of transparency among major software vendors play out?
When we moved to SaaS, it was marketed as a rental model. Even then, we warned everyone that the total cost, however you account for it, would multiply. All you had to do was look at SAP’s and Oracle’s investor decks. They said, “Congratulations, we’re going to SaaS. It’s going to generate over three times more money from each customer because they’ll be in rental mode forever, instead of owning their house.” We saw it, we warned everyone, and it all came true. The thing to remember about this industry, like every industry, is that greed tends to rule the day. Vendors will always push the limits until something breaks. They’ve done this every time.
Look what happened with cloud. It was supposed to give you the ability to scale resources up and down without having a fixed amount. Imagine a retailer running a big sale: website traffic goes through the roof, so they need extra processing power to handle it. Then traffic drops again. Without cloud, that retailer could face a denial-of-service situation during the peak unless they buy enough (infrastructure) to always be at the highest level.
We’re entering another cycle. This one moves away from user-based, revenue-based, or company-size metrics for determining cost. Everyone now sees data as the new gold. In our model, we’ll be moving vast amounts of data between applications as we integrate more systems than ever before. So what do vendors want to do? They want to put tollbooths on those data highways, charging a transaction fee every time information moves from one place to another. Now imagine when all these companies set up their tollbooths. How many tolls can you afford to pay? And what happens when they raise the toll?
This emerging world of data-movement charges will change system architectures. For the past 30 years, ERP software has typically sat at the centre of a hub-and-spoke model. Data flows in and out of that hub to hundreds of connected systems. Companies like SAP have realised they can monetise that movement, effectively charging people for extracting their own data. Imagine paying a toll to use data you already own. You put it into the system, and now you’re charged when you take it out.
Indirect usage was their first attempt at this, followed by digital access rights. You can see where it’s heading: They’re afraid these revenues will disappear, so they’re doing everything they can to preserve them. For example, if you use SAP’s payroll platform SuccessFactors, you can move data freely within that system. But if you use Workday instead, SAP will allow the integration but will charge you for every data transfer between the two systems. They’re building ecosystems where staying inside the vendor’s walls means no tolls, but moving data beyond them incurs a fee. This is the direction the market is taking, and we believe the traditional ERP model will eventually disappear.
I’m advocating a completely new approach built on open source. Let’s create open-source hubs; we already have data hubs with Oracle, customer hubs, and vendor hubs. Now we need an open-source platform hub where data is stored freely and belongs entirely to the enterprise. You own the highways in and out, and no one can impose tolls. The ERP system should sit as one of many satellites, not at the centre. We’ll feed it only the data it actually needs, not everything. That way, you’re not paying to extract irrelevant data. At the same time, we’re going to remove that whole ERP software within the next five to 10 years. That way, we’ll have the data hubs down here, all of our applications above them, and we’ll end up sitting over that entire layer.
