Have you ever heard of a company called Rimini Street?
01
Its business is incredibly bizarre. It's a company that doesn't write a single line of code or develop any software product itself. It makes over $400 million a year by maintaining and repairing old systems that others have phased out.
Its clients include Welch's (the grape juice brand), Japan's largest automotive aftermarket retailer AUTOBACS, and Khimji Ramdas, the largest private conglomerate in Oman.
Its business model can be summed up in one sentence:
You bought Oracle or SAP software and pay a sky-high annual maintenance fee. Rimini Street says, give me half that amount, and I'll do the same job.
Yes, you read that correctly. Helping you save money gave birth to a whole company. How did that happen? It's an interesting story.
Let's first do the math with Oracle.
You spend $1 million on an Oracle software license, thinking that's the end of it. Sorry, that's just the beginning. Starting the second year, Oracle charges you 22% of the license fee annually, which is $220,000, called the "Annual Maintenance Fee."
What's even tougher is that this fee increases by 4% to 8% every year.
Do the math: For a $1 million license, the first-year maintenance is $220,000, the second year is $237,600, the third year is $256,656... over five years, you pay a total of over $1.3 million in maintenance fees, more than the cost of the software itself.
You might ask, can I not pay? The answer is: You can choose not to.
But you will immediately lose three things: security patches (no one will fix system vulnerabilities), regulatory updates (your system won't know if tax rates change), and technical support (you're on your own if problems arise).
Auditors won't let it slide, neither will insurers. Essentially, what you're buying is an "uptime" insurance.
Let's be honest, what do we call this? Protection money.
SAP is similar. Globally, about 35,000 companies use its ECC system. In 2015, SAP released its new S/4HANA system, but even today, over half of its clients are still stuck on the old system and haven't migrated.
Why not migrate? Because a migration starts at several million dollars for a medium-sized enterprise and can reach up to a billion dollars for large multinational corporations, taking 18 to 36 months.
More importantly, the old systems are incredibly stable; an ECC system fine-tuned over 15 years can run finance, process procurement approvals, and manage the supply chain. Why spend hundreds of millions to stop and overhaul it?
So, no migration. Having paid maintenance for 20 years, they keep paying.
The CTO of the American beverage brand Welch's faced this exact situation. The maintenance fees for Oracle EBS were taking up an unreasonably high portion of his IT budget. Instead of asking the CEO for a bigger budget, he directly switched to Rimini Street.
According to Rimini's official case study, the money saved in the first year amounted to roughly a quarter of the company's net profit that year. A quarter, folks. One IT decision offset years of sales growth.
By now, you probably see the picture:
The maintenance fees for Oracle and SAP are essentially toll fees; you bought their road, so you have to pay them forever. Stop paying? You can still use the road, but you're on your own if there's an accident.
This business has been running for decades at Oracle and SAP, solid as a rock, because no one had ever broken this scheme.
There's a saying I really like: "Switching cost > dissatisfaction = lock-in." If the cost to replace your financial system is $1 billion, no matter how dissatisfied you are, you won't switch.
Oracle and SAP profited massively from this formula.
Until someone stood up and did one thing. He said software and maintenance are fundamentally two separate things.
You buy your software; I'll handle the maintenance for you. It's legally sound, I'll give you a 50% discount, and the service is better than the vendor's. This person is Seth Ravin. He did this because he knew which screw held the lock in place.
02
So, how exactly was this lock broken?
Back in the 1990s, Ravin was an executive at a company called PeopleSoft (later acquired by Oracle), responsible for client sales.
He did something many wouldn't think of: He created an "Extended Support Plan" within the company, specifically for clients whose systems were running stably and didn't want to be forced to upgrade.
In essence, he created an "anti-lock-in" scheme within the original vendor. This didn't become huge, but he learned one thing: What clients fear most isn't the cost; it's the hassle.
In 2002, he left to start a business, co-founding a company called TomorrowNow. He did the same thing: helping clients maintain old Oracle systems for half the price charged by the original vendor.
In 2005, SAP bought TomorrowNow. Ravin stayed for three months and then left.
What followed is known in the industry: Oracle sued SAP and TomorrowNow for copyright infringement. SAP ended up paying $359 million in damages, and TomorrowNow was shut down.
The first attempt to break the lock failed.
That same year, Ravin started another company in Las Vegas: Rimini Street. Same playbook, same logic. Oracle saw it and thought, "Here we go again?" In 2010, they formally sued, with the same reasoning: copyright infringement.
This time, however, things were different.
For the next 15 years, Rimini Street fought a hard battle against Oracle. The first trial in 2015 saw Oracle bring 24 claims; it only won on one, and that was ruled as "innocent infringement."
What does that mean?
The court's message was clear: Rimini wasn't intentionally stealing your stuff; there were some process issues, just fix them.
In 2018, the U.S. Court of Appeals for the Ninth Circuit overturned more of the rulings, and Ravin was absolved of most personal liability. The court's exact words: Rimini Street's provision of third-party support constitutes lawful competition with Oracle's direct maintenance services.
Lawful competition. These four words took 15 years.
It wasn't over yet. In 2019, the case reached the U.S. Supreme Court. All nine justices unanimously ruled that Oracle must return $12.8 million to Rimini Street.
Oracle not only lost the case but had to pay back money.
There were some further twists. In 2023, a Nevada federal judge, on the PeopleSoft front, found Rimini liable for repeated infringement and issued a new injunction.
This issue was moderate. Significant? The core Oracle Database and SAP businesses were unaffected. Minor? Well, they did get caught on something.
By July 2025, the two sides finally sat down and signed a final settlement. Oracle returned $37.9 million in attorney fees, and Rimini Street voluntarily exited the PeopleSoft product line (about $20 million in annual revenue), but its core business remained completely unaffected.
Let's be honest, Oracle's motive for this lawsuit was clear.
It wasn't afraid of Rimini Street stealing its code. It was afraid of Rimini Street proving one thing: software and maintenance are two separate commodities.
Once this perception spreads, the annual multi-billion dollar maintenance fee cake would lose a huge slice.
After proving that, step two was simple: price and service.
Oracle charges you 22% of your license fee, increasing by 4-8% annually. Rimini Street charges about half, with prices largely fixed.
Oracle's maintenance doesn't cover your decades-old custom code; Rimini covers it all. For the highest priority failures, Oracle offers no guaranteed response time; Rimini pledges 10 minutes, averaging under two minutes in reality.
A large conglomerate in Oman called Khimji Ramdas had over 700 custom modules on its SAP system. Switching systems meant redoing everything. Their technical lead's exact words: "We should have switched to Rimini years earlier. Our total maintenance costs were slashed by 80%."
Japan's largest automotive aftermarket retailer, AUTOBACS, has been with Rimini for a full decade. Their system hasn't crashed once, and all the saved money has been invested into AI and IoT.
Moreover, beyond Oracle and SAP, this playbook can be replicated.
After Broadcom acquired VMware, it eliminated perpetual licenses and drastically raised prices, leaving many long-time VMware clients with nowhere to turn. Rimini Street launched VMware support services in 2024 and signed over a hundred contracts in less than a year and a half.
By this point, Rimini Street has fought three rounds with original vendors: Oracle, SAP, VMware. Three rounds, all victories.
What does the court say? Lawful competition. What do clients say? A quarter of net profit in savings. Can it be replicated? Yes, Broadcom's clients are lining up.
So, to sum it up in one sentence: The lock has indeed been broken.
03
But there is one problem it cannot circumvent. It took 15 years to break the lock, burned over a hundred million dollars in legal fees, employed over two thousand engineers, and finally managed to pry clients away from the original vendors.
And then what? Doesn't this company itself need to make money? Did the 50% in saved maintenance fees truly all go into the clients' pockets?
You've probably already guessed. Of course not.
The money Rimini Street saves you is finding its way back onto its own books in a more subtle way. How is it done? It's actually quite simple. Maintenance is just the entry point.
After you switch to Rimini's maintenance, you save money and your system stays stable. Then Rimini says to you:
What about your day-to-day operations, monitoring, security, and system integrations? Why not hand them over to us? We have experts and tools, and it's cheaper than maintaining your own team.
You agree. So you sign up for Rimini Manage (managed services).
A while later, they come again: Your old system runs on old browsers and needs to be compatible with new systems, right? We have Rimini Connect. Can't handle security patches? We have Rimini Protect, specifically designed to shield old systems from attacks, with 75 security experts monitoring for you 24/7.
You sign again.
Finally, they unveil the real big move: AI. You don't want to migrate your system? That's fine. We'll overlay an AI layer on top of your existing system without touching the underlying code, allowing you to use AI for processing tickets, reconciliation, and approvals.
Partnering with ServiceNow, with 20 templates, this is already running at 26 clients.
Think about it. At this stage, can you still leave them?
One case study is particularly telling. Apsen, a pharmaceutical company in Brazil, had a perfectly running SAP system and didn't want to migrate.
Rimini helped them implement ServiceNow's AI workflows on top of the old system, which went live in a few weeks.
Previously, a single material transfer process involved manually handling over 100 requests per month, moving more than 50,000 finished goods, all through endless back-and-forth emails and spreadsheets.
After implementing AI, 70% of the manual processes were automated, and development cycles shrank from months to weeks.
How much money did they save? I couldn't find a public figure. But one thing is clear: If they ever try to replace Rimini now, they'll lose an entire, fully operational AI automation system.
This is Rimini's new lock. You'll notice its model is quite interesting.
Layer 1: Saves you money from the original vendor. Layer 2: Uses the saved money to sell you new services. Layer 3: The more new services you buy, the higher the cost becomes to leave them.
Simply put: It's earning back the money it saved you.
Looking at the data, this new lock is indeed effective. Client retention is around 90%. The total value of contracts locked in before expiration is $653 million, a historical high. International business grew by 14%.
But if you look closely, this new lock isn't as solid as the original vendor's yet.
First, security patches are a hard limitation. Rimini's security solution is called a "virtual patch," essentially adding a layer of protection outside the system rather than fixing the code itself.
For heavily regulated industries like finance and healthcare, this isn't enough; it won't pass audit requirements.
In other words, the juiciest part of the market remains inaccessible to them.
Secondly, once you choose Rimini, there's basically no turning back. Returning to the original vendor means Oracle would charge you 150% in backdated maintenance fees; SAP is similar. You saved money for three years, only to pay it all back in one go upon returning. It's a one-way street.
Another issue: In the company's latest financial report, new order value dropped 8.8% year-over-year. Its liabilities exceed its assets, resulting in negative net worth.
The sales team was just split into two groups: one dedicated to acquiring new clients, the other focused on serving renewals for existing clients. This indicates growth pressure and internal adjustments.
The original vendors aren't idle either. Oracle and SAP are actively pushing their clients toward the cloud.
In the SaaS model, the software license resides with the vendor. You don't even have the *choice* of maintenance. Previously, you could buy software from A and get maintenance from B. On the cloud, you must buy everything from A.
The original vendors' new lock relies on architecture to lock you in.
By now, you can probably see an interesting picture. Rimini Street spent 15 years breaking an old lock, and then, it installed a new lock of its own.
This new lock isn't as strong yet, but it is getting stronger. Meanwhile, the original vendors are forging an even harsher new lock.
Two locks are racing.
Rimini Street's new lock relies on service stickiness. You *can* leave, but who will manage those hundreds of custom modules if you do?
The original vendors' new lock relies on architectural lock-in. You simply *cannot* leave; the software isn't in your hands. Between these two locks, which one offers you more freedom? I don't have an answer.
Perhaps in the future, the AI-native ERP players—those currently only doing financial modules—will grow up to become the new lock-breakers. Or perhaps not. But one thing is certain: As long as there are people building locks, there will be people breaking them. It's a cycle.
It's similar in China. Stories of clients of Yonyou, Kingdee being tied to "信创" initiatives or locked in by cloudification follow essentially the same logic as Oracle and SAP.
What's interesting is that in China, no Rimini Street has stood up yet. Is the timing not right, or is no one daring to do it?
This article is from the WeChat public account "王智远" (ID: Z201440), author: Wang Zhiyuan.
