Servers no longer need to be purchased. CEO of Pro-Data on how businesses are transitioning to infrastructure as a service

Today, more and more businesses are moving to cloud services—to scale faster, avoid freezing capital in hardware, and reduce infrastructure risks. We spoke with the head of Pro-Data about how the cloud technology market in the region is changing, who needs local infrastructure, and why the next big demand will be driven by AI.

Evgeny Vishnevskiy, CEO of Pro-Data, LinkedIn

About Me

I have been managing Pro-Data for six years now, and together with the team, we are developing the cloud technology market in Uzbekistan and Kazakhstan.

When we started in 2021, cloud technologies were relatively unknown in Central Asia. Such services were already offered by individual players, but mostly these were large telecom operators for whom the cloud direction was just a part of their business. For the past six years, we have been fully focused on this market and have likely become one of the key drivers actively developing it in our region and helping to make cloud technologies a full-fledged service for businesses.

Before Pro-Data, I worked at TBC Bank, Uztelecom, and other telecommunications operators. Almost all my previous roles were related to business development and sales, but always in technological and IT fields.

When I left the bank, I decided to start my own project. It had to be related to the IT sphere, which I know well and have been engaged in for almost my entire life. At the same time, I wanted to create something new that did not yet exist in the Uzbek market.

I had two options. The first was a project in the Infrastructure as a Service (IaaS) sector, i.e., a cloud provider. The second was related to smart homes and automation—essentially, the development of the Internet of Things. However, thanks to my experience working with internet providers, the cloud technology direction was closer and clearer to me. Therefore, I ultimately chose it.

We saw that, like the rest of the world, Uzbekistan was experiencing a growing level of digitalization and increasing business demands on IT infrastructure. We understood that demand for such services would increase. We decided to try developing this direction—and our assumptions were confirmed.

What is the Sharing Economy

The sharing economy is a model where you give up ownership in favor of access to a resource as a service. For example, we sell infrastructure as a service and show clients that in many scenarios, it is much more profitable, secure, and reliable to use infrastructure as a service rather than owning it yourself. This is the foundation of our business.

In other sectors, the model works the same way. For example, with scooters: people use them every day but do not necessarily own them. The same applies to offices—you can buy a space, rent it, or use a ready-made space in a co-working center.

The word "economy" comes from "saving." It is about more rational use of resources. If you own a server but do not use it at full capacity, it still consumes electricity, requires cooling, and after becoming obsolete, it must be disposed of. This is not only uneconomical but also environmentally unfriendly. Therefore, in many cases, the sharing economy allows for more efficient use of resources and reduces environmental impact. Owning something makes sense only when it is difficult or impossible to use it as a service.

The same applies to infrastructure. Previously, in Kazakhstan and Uzbekistan, it was practically impossible to use infrastructure as a service: you had to buy your own servers, place them in a data center, and operate them yourself. Now, there is no such need. Cloud providers can offer dedicated physical servers, virtual machines, or platform services. Therefore, the level of dependence on the provider can be flexibly adjusted today.

How We Help Clients

One of the main problems when purchasing your own infrastructure is time. In a small company, purchasing equipment can take two to three weeks, while in large companies, banks, and government organizations, it can take 6, 12, or even 18 months. Modern equipment usually needs to be ordered through a chain: integrator-distributor-vendor.

At the same time, infrastructure is purchased for several years in advance with a buffer, because making a mistake with the configuration is expensive. As a result, the process turns out to be long.

Infrastructure as a service can be obtained much faster. A virtual server on our cloud platform can be launched in two to three minutes. A dedicated server can be obtained in five minutes if the required configuration is already mounted in the data center. A custom configuration takes three to five days, and if equipment needs to be brought from abroad, it takes a little longer.

The second problem is reliability and security. Organizing a reliable server room or data center on your own is possible, but it is expensive, difficult, and requires specific competencies. Not all companies have people who can handle this.

And thirdly, your own infrastructure needs to be operated and maintained. This takes time and focus from the team and management. We help the client not spend resources on this and focus on their own business.

Yes, sometimes this approach is more expensive than doing everything yourself. But ultimately, the company wins by directing its focus to its own products and services.

When the Cloud is More Profitable than Own Equipment

Renting is especially profitable when the business is constantly growing. For example, a new payment system or bank is launched: today they have 100 clients, in a month 10,000, and in six months—2 million. What infrastructure should they rely on? If they buy a small one, it will later need to be expanded, and sometimes equipment must be written off and new equipment purchased, especially if a powerful server is needed for a high-performance application or database.

In the cloud, you pay only for what you use: today you take one virtual server, add more as load increases, and after the peak period, you get rid of excess infrastructure. Therefore, for growing businesses, infrastructure as a service is often more profitable.

The second point is risk. There are startups that do not yet know if they will succeed. If the probability of success is conditionally one in ten, why buy infrastructure and then decide what to do with it if the project closes? It is better to take it as a service and avoid capital expenditures. The same applies to large companies launching new products. It is unknown in advance which of them will work, so infrastructure can be taken as a service without freezing money.

How Pro-Data Infrastructure Works

Currently, we have seven locations in Uzbekistan: five are built on our own equipment, and two are partner sites where we started. In Kazakhstan, we have one location—in Almaty. All of them are located in data centers built in accordance with the requirements of the international reliability standard Tier III. The client connects to the site closest to them, and our services can be used from anywhere in the world.

For us, the key issue is uninterrupted operation. In Central Asia, due to the specifics of the energy infrastructure, power grid failures occur, especially during periods of extreme heat or cold. Data centers must not stop: our users—banks, payment systems, retail—must work 24/7.

Therefore, all data centers have two power lines from different substations, backup generators, and uninterruptible power supply (UPS) units that ensure operation during switching. Air conditioning is also backed up.

All equipment—servers, network equipment, storage and security systems—also operates with redundancy. If a physical device fails, the backup node takes over the load, and we can calmly replace or reconfigure it. For example, for every five servers, we take one backup—that is about 20% additional cost. On our scale, this approach is more efficient than if the client created similar backup infrastructure on their own.

In Uzbekistan and Kazakhstan, we have local client support teams, as well as infrastructure support and monitoring teams that work 24/7. This allows us to quickly respond to incidents and provide on-site support to clients.

Who Needs the Cloud

First of all, the cloud is relevant for growing businesses—companies with a large number of clients and data, for whom it is important that systems work reliably and without interruption. Especially if it concerns personal data of citizens of Central Asian countries.

If we talk about industries, the first is the banking and financial sector: banks, payment systems, payment gateways and applications, microcredit and insurance organizations. That is, companies operating in a regulated environment.

The second large segment is trade and e-commerce: chain retail, supermarkets, electronics and clothing stores, online stores. Their services must work 24/7 because during outages, every minute of downtime can mean loss of money and clients. Therefore, they choose us to minimize such risks.

We have about 300 clients in two countries, with approximately 60% of revenue coming from the banking and financial sector.

There are also more specific clients. For example, in Uzbekistan, our services are used by airlines. Flight schedule systems, archival and technical documentation of aircraft, email, and office applications can be hosted on our infrastructure. An airline has dozens of different systems, and a significant part of standard IT infrastructure can also run in the cloud.

Why Businesses Need Local Cloud

In Central Asian countries, the legislation of some countries provides for the storage of personal data within the country. Therefore, for regulated industries, if it concerns personal data, the choice is practically one—local infrastructure.

But even without such a requirement, companies sometimes consciously choose to store data within the country. This is also related to geopolitical risks. We saw this in the example of Russia, when after the beginning of geopolitical confrontation, many Western vendors left the market, and some Western cloud providers stopped servicing Russian companies. Businesses had to quickly transfer their data and services.

Another factor is technical support. With international providers like Google or Amazon Web Services, resolving an issue can take time: first, you need to go through several levels of support, submit a ticket, and wait for a specialist. Moreover, they do not always have access or authority to resolve the issue promptly.

Therefore, for some companies, it is important that data is located within the country, support is local, payments are made in local currency, and services continue to work even in case of internet problems or border closures. We specifically address these needs.

The Future of Cloud Infrastructure

The main change we are already seeing is the penetration and industrial use of AI. For cloud providers, the most obvious direction is infrastructure for AI. We are already doing this: our portfolio includes dedicated servers with graphics accelerators. We have also completed the implementation of a platform that allows distributing graphics cards between projects, selecting models, and deploying AI solutions within your own perimeter.

There are companies that are comfortable using ChatGPT, Claude, and other external services. But there are also those who cannot transfer their data or client data to them. They need similar solutions within their own secure perimeter. We can provide such infrastructure and a platform for managing it—both in our data center and at the client's site.

I believe that by the end of this year, 5-10% of our revenue will come from GPU and AI infrastructure. In the next three to five years, this share could grow to 50%. Traditional infrastructure will also continue to grow, but AI infrastructure is developing exponentially.

The second trend is that products and services must become simpler. Clients want to get more while spending less effort and time. For example, previously 1C was deployed on own servers in the office. Now many host it in the cloud, paying separately for the 1C subscription and the cloud server. In the future, there will be fewer such scenarios—companies will receive 1C immediately as a service.

The same is happening with other systems: instead of their own infrastructure, companies are gradually moving to ready-made SaaS products. Therefore, companies like ours will provide less infrastructure to end users and more to technology companies that will use it to create their own ready-made solutions.

This is exactly why we are developing our own products and services. We already have a cloud platform with various PaaS services. Last year, we finished work on managed Kubernetes, and now we are developing managed databases. The user does not need to independently create, deploy, and maintain databases—all this is done automatically through our panel.

We have about 10 such services in total, which will gradually appear on the platform.