thomas.wieberneit@aheadcrm.co.nz

The New Enterprise Moat? Zoho’s AppOS and Stack Sovereignty Signal the End of Fragmented SaaS

ZohoDay 2026 is in the books, and it has again been an intense two days of information and discussions, starting off with some impressive statistics. In time for its 30th anniversary, Zoho crossed the milestones of one million paying customers and an eye watering 150 million users. All this while not having raised a single dollar of external capital or buying technology or users. The company stays fiercely independent and continues to grow very profitably since it crossed the threshold of an annual revenue of $1bn back in November 2022.

If I wanted to boil this event down to a few key messages, it would be value, independence, platform, and, of course, AI. 

The conference in a nutshell: Value is the result of the smart use of automation with AI that works on top of the corporate system of record, powered by a platform that is built on an independently owned stack. This is also the secret sauce of Zoho, a philosophy that the company follows since its inception. 

And here is how Zoho brings this to work. 

Zoho owns and continuously improves its stack

Coming from the angle of sovereignty, Zoho extends this thought of independence to its customers now in an answer to Raju Vegesna’s not so rhetoric question “what will happen if someone can pull the plug?” on any of your essential systems. All of the sudden, the thought of local deployments or hybrid deployments with cloud apps operating on local data becomes very interesting, valuable, again. It is mitigating risk. According to Vegesna, clients of different sizes are asking for this model. Another part of this equation is the ability to mitigate risk while keeping control as Zoho Corp’s Chief Strategy Officer Vijay Sundaram maintains. Apart from being a business and deployment model, he argues that SaaS brought an institutionalized risk reallocation from the customer to the vendor. 

Customer Risk under a SaaS Model; source Zoho

Part of the price of this reallocation for the customer is ceding control to the vendor, which in turn can become a risk in itself. 

On top of this, the move to more use of AI and agentic systems, along with the proliferation of no-code and low-code systems, which moves control back to customers, is increasing their risk exposure, with all its associated security, compliance, operational and customization risks, as more and more business logic is built by the customer again.

Using its own stack, from hardware to AI, with what Sundaram dubs “the magnificence of the mundane”, the corporate systems of record, at its heart, Zoho offers customers control while moving less risk back to the customer.

The value of being able to mitigate risk, that is hard to quantify comes on top of the value that Zoho offers through owning its stack, therefore being sovereign itself and in control of (most of) its own risk.

Platform play

Zoho regards itself as a technology vendor rather than a business app vendor for quite some time now. Quite rightfully so, as the company builds its own hard- and software stack and as part of it, its own application IaaS and PaaS, to stick to SaaS terms. The problem is that this led to a sprawl of business apps and corresponding fragmentation on the data level, that even a middleware cannot keep up with; not even talking about shadow IT and shadow AI.

The consequence is a chaos of apps, inconsistently connected by various middlewares, these days with an army of (unreliable) agents on top of them. As an example, businesses have deployed 275 according to the 2025 Zylo SaaS Management Index

Enterprise software, current state, with agents; source: Zoho

Zoho intends to fix this by using a page from Apple’s iOS platform book. This way, the company wants to be able to not only keep the infrastructure layer clean but also leverage its already existing application foundation by streamlining it. Zoho calls this AppOS. AppOS is intended to have applications built within it, on a shared data foundation, using common business and process models, built-in workflows and automations, etc. In simple terms, AppOS contains everything that is needed to build and run applications and that is not domain specific. It gets enriched by domain specific platforms on which the various apps are built. 

With AppOS, Zoho wants to bridge the gap between the increasing proliferation of apps that gets fueled by AI code generation and the business need to maintain structure. It shall do so by providing a common data structure, a consistent business context and, last but not least, by enforcing standards. According to Mani Vembu, CEO of the Zoho Division of Zoho Corp., transitioning to AppOS will be a seamless experience for customers.

And then there is AI

Let me start with that I found it refreshing that AI was not a centerpiece of the presentations. Actually, Zoho caught some analyst flak for not pushing an AI story aggressively. 

Which they responded to with a strong counter argument. Most vendors bolt AI on top of an unreliable data infrastructure. Zoho wants to do it right and first lay the data foundation and with that what Zoho calls the “unified business context”. This unified business context is built of the combination of designed and discovered information, structured and unstructured data. Based on this context, AI will become more reliable. In addition, it enables Zoho  to offer own foundational LLMs and MoE models that are both, more reliable, more performing, and cheaper to run. Why? Because context, private data and specialized models always outperform generic data that is fed into massive models. With that approach, Zoho puts intelligence in front of the models, thereby avoiding the need for what I’d call “commodity inferences”. Plus, own models enable Zoho to allow for custom training and custom deployment of specialized models. As also a recent HBR report stated, the competitive advantage that businesses have, lies in context, not in the model, in particular not if businesses use the same models. This ties back to the unified business context of AppOS. Plus, inference will become expensive again, when the major LLM vendors switch from an investment mode to a value extraction mode to turn profitable. Combined with an emerging own agent ecosystem, Zoho again can promise to offer customers a high value.

My take

Zoho shows a remarkably consistent strategy that centers around “maniacally focusing on value”, monetary as well as less tangible value like the ability to control or balance risk. 

Given all the economic and regulatory uncertainty that we see these days, this strategy of maniacally focusing on value by the smart use of automation with AI that works on top of the corporate system of record, powered by a platform that is built on an independently owned stack, puts Zoho in a sweet spot. The combination of data, context, and sovereignty ensures this.

Of course, there are companies that can compete with Zoho on the breadth, width and depth of functionality, but not very many. Looking at Zoho’s overall capabilities, there is only Microsoft. Looking at business applications only, we can add Oracle and SAP, maybe ServiceNow soon. In a world that increasingly swings back to – and needs to swing back to – suite thinking, Zoho punches well above its weight. This is not to say that specialists have an important place in the software industry, but consistent and contextual data will become ever more important. 

Are there challenges ahead? Sure as! Zoho has big ambitions, enhancing its software, building out its AI, not to mention getting a deeper foothold in the ERP market, which is a topic that I didn’t cover at all here. And then, there is the topic of implementation cost. The better the price to value ratio of Zoho software becomes, and I’d consider it market leading, the more implementation cost matter. They can outstrip the licensing cost by an order of magnitude. 

Zoho’s Chief Scientist Sridhar Vembu cited examples of how efficient software development with the help of AI code generation has become. One of them has been a compiler that two developers could create in 2 – 3 days instead of the 3 months that it would likely take human engineers. This kind of value needs to hit the customer implementation, and fast. 

On top, as also Jon Reed alluded to in his excellent take on ZohoDay 2026, it still needs to be proven that AppOS delivers on flexible customizing while avoiding technical debt.

In closing, as Raju Vegesna said in his speech, customers license not software but peace of mind and not the software, but customer satisfaction is an enterprise software vendor’s ultimate moat.

I tend to agree. And Zoho is well on the way down that route.