Anuar Zhaksylyk and Temirlan Bulanbay launched QazTWIN, a startup that creates 3D digital models of factories and collects data from thousands of industrial sensors.
For the joint Digital Business and Astana Hub project “100 Startup Stories from Central Eurasia,” Anuar and Temirlan shared how digital twins can help factories, how they landed their first contract, and why they’re now looking to enter the US market.
“Management often doesn’t have a complete picture of what’s happening across the factory”
– Tell us a little about yourselves. What were you doing before you started the company?
Anuar: – I studied at NIS and then went on to Nazarbayev University, where I made it to the semifinals of the ACM ICPC student programming championship. I launched my first business while I was still at university and was already earning good money from it. Later, I worked on several different projects before building a product in the renewable energy sector.
Temirlan: – I studied finance and started working in banking when I was 19. I later moved into corporate finance, where I worked closely with industrial companies and got to see how their businesses operate from the inside: how production is organized, where they make money, and where they lose it.
Anuar Zhaksylyk
– What inspired you to start QazTWIN?
Temirlan: – At first, Anuar and I were working on a project to digitally visualize industrial facilities. Soon after, we joined Astana Hub and began developing the idea into a startup. It grew out of a very clear problem: at large industrial sites, it can be difficult for management to keep track of the whole picture, from where workshops and equipment are located to how other facilities are laid out. A visual digital twin brings all of that together in a single 3D model.
We soon realized that visuals alone weren’t enough. We needed to add equipment data as well. That’s when it becomes a live model of the facility, showing exactly where problems are happening.
Anuar: – Every industrial facility has its basic data sources, but they’re not connected to each other: a SCADA system, which collects real-time equipment data from across the entire facility; an ERP system, which stores data on finances, procurement, inventory, and other business processes; and 1C. In reality, there’s more than enough data, but hardly anyone is actually doing anything with it.
Temirlan Bulanbay
And we thought: what if we brought the visualization to life by creating a huge data lake (a centralized repository where raw data is collected – Digital Business) and connected it to the visual model? Instead of linking the data to tables and diagrams, you tie it to an exact 3D replica of the facility, with every object in its real-world location. This gives you a much more complete picture of how the entire operation is working.
And that’s how QazTWIN evolved into the product it is today.
“You can chat with... your own factory”
– How do employees use QazTWIN?
Anuar: – For users, QazTWIN is a 3D digital model of the facility that also shows real-time data on how the equipment is performing. Since every employee has their own area of responsibility, the system can be configured with different roles and access levels.
For example, management is more interested in the financials and the big picture, while the chief process engineer needs to know that a bearing has failed on a particular pump in workshop No. 5. The visualization is fully customizable, so each user can adjust the interface to monitor equipment in whatever way works best for them.
– So if something breaks down at the factory, you’ll see it straight away?
Anuar: – Yes. And if there’s enough data, the system can even predict a breakdown several hours, or sometimes even days, in advance.
But the most important part is figuring out what caused the problem. How do factories handle this today? If an operator spots something unusual, they start calling the process engineer or messaging people who are near the equipment. Before long, three or four people are going back and forth, trying to work out what’s wrong. On average, this takes 8 hours, and in the worst cases, 2–3 days. Every hour of production downtime costs a huge amount of money. With QazTWIN, the time it takes to identify the problem is reduced dramatically. In typical cases, it takes just a few minutes.
Temirlan: – We rolled out the platform for our first client in February, and in just six months, they managed to cut downtime and save several hundred thousand dollars.
– Why does it take so long to figure out what’s wrong? The operators already have all the sensor data…
Anuar: – Imagine a factory with 10 workshops spread across 40 hectares and around 150 pieces of equipment. If a specialist notices a problem with production output, they first need to identify the problem area, then find the specific machine, and only then figure out what exactly has gone wrong.
The SCADA system does show real-time data from all the controllers installed on the equipment, such as pressure, temperature, and so on. But someone still has to analyze all that information manually. And how can you do that quickly when the system may contain several thousand parameters, including 400 dynamic ones that change every second?
QazTWIN sees the full picture and understands the entire production process. We pull in all the data from SCADA and also upload all the equipment documentation into the system, including manufacturers’ operating manuals, maintenance reports, and error logs.
Temirlan: – Creating a single “brain” and knowledge base allows the system to predict what might break and where, and quickly identify the cause of a problem. The final decision on what action to take still rests with a person.
– What else can QazTWIN do?
Anuar: – QazTWIN combines three things: a 3D model of the factory, data from its systems, and analytics. This gives the platform a wide range of features.
For example, there’s an asset manager. It isn’t directly involved in production, but it gives you a complete overview of every piece of equipment at the facility: where it is located, its dimensions, when it was last serviced or repaired, who is responsible for it, what the operating manual says, and so on.
We also have a task tracker. Instead of messaging someone on WhatsApp saying, “Replace the pump on the far right in workshop X,” you can select any piece of equipment in the system and assign a task to another person. They’ll get a notification and immediately see where the pump is, what it looks like, and what needs to be repaired.
Temirlan: – QazTWIN also has an LLM-powered chat trained for a specific production facility. And you can actually chat with… your own factory. Before, a director might check SCADA readings, analyze data in Excel, and go through work chats on WhatsApp. With our platform, they can simply open the chat and ask: “Hi, factory! What’s critical right now? What should I be paying attention to? Which tasks assigned to my team are still unfinished?” In response, they get a full report on what is happening at the facility at that very moment.
Anuar: – There are plenty of ways to use the chat. For example, new employees can ask it how the production process works. In other words, you can talk to the entire factory and ask anything, from identifying equipment problems to calculating financial metrics.
“If a factory hasn’t been modernized since the 1980s, digitalization won’t help”
– How many digital twins have you implemented so far?
Temirlan: – Our first project is now fully operational at a sulfuric acid plant. We’ve also completed two smaller implementations: digital twins for a medium-sized warehouse and a small oil refinery.
Strategically, we’re most interested in large corporate groups that own dozens of industrial facilities. A successful rollout with one such partner can open the door to many more factories. So right now, most of our efforts are focused on three projects within one of these groups.
Another seven companies in the energy and oil and gas sectors are currently on our waiting list. We’ve already visited some of their sites and carried out technical audits. We’re now preparing commercial proposals, while some have already moved on to the pilot stage.
– Can QazTWIN be used at any factory?
Anuar: – We have an ideal customer profile, and based on that, we’ve identified the types of facilities we definitely won’t work with. These are plants where workshops and equipment operate independently and aren’t connected to each other. We also won’t be able to help facilities that don’t have systems like SCADA or ERP.
Temirlan: – If a factory hasn’t been modernized since the 1980s, our platform alone isn’t going to close a 40-year technology gap. We visited four sites, looked at their operations, and told them, “Guys, you don’t need our system.” And that was despite the fact that the clients were ready to pay.
– How long does implementation take?
Anuar: – About six months. Digitizing a factory still involves a lot of manual work. First, you need to get a huge number of approvals just to access the site. Then you have to create the visualization, which also takes time, although by now we’ve become very good at it and improved the process as much as possible. We even developed our own engine so that complex 3D models can run smoothly even on older smartphones.
Temirlan: – The biggest drain on resources is integrating with the factory’s existing systems. This stage takes anywhere from one to three months. We’re now developing our own hardware that will sit in the server room, pull all the data from the SCADA system, and automatically match it to the right parameters and equipment. This should help us cut the integration process down to just a few days.
Anuar: – Overall, we want to bring the platform deployment time down to one month, and eventually to just one week. If we don’t, we won’t be able to scale quickly.
“The platform pays for itself with just a couple of hours of downtime saved”
– So how does the business make money?
Temirlan: – First, the client pays for the QazTWIN rollout, including site visits, visualization, system integration, and installation of additional equipment. After that, we switch to annual fees, which depend on the size of the facility.
Our first contract, with a sulfuric acid plant, was worth more than $500,000. We also earned revenue from two smaller projects, digital twins for a warehouse and a small oil refinery, and we now have recurring annual payments as well. Altogether, we’ve generated more than $1 million.
Anuar: – For a smaller factory with fewer workshops and less equipment, and where integration is relatively straightforward, rolling out our platform can cost from $50,000. The annual subscription is 20–25% of the implementation cost.
At a large factory, our services can cost more than $800,000. The final price depends on the size of the facility, the amount of equipment, and how complex the integration with its systems is.
– How quickly does the investment pay off?
Temirlan: – The most direct financial benefit for a factory comes from being able to warn about a potential breakdown in advance. Say replacing a part proactively takes just two hours, compared with 10 hours for an emergency repair. If one hour of downtime costs $100,000, then, as incredible as it may sound, our platform can pay for itself in just a few hours.
We also help reduce the time needed for scheduled preventive maintenance, which typically takes around 30 days a year at industrial facilities. If that can be cut by even two or three days, the savings can be significant.
There are also indirect benefits, such as saving staff time when diagnosing problems. With QazTWIN, specialists can see the entire operation at a glance, get the information they need faster, and make decisions more quickly.
– How much have you invested in the startup?
Temirlan: – We initially invested around $400,000–450,000, which came from an angel investor. The money went toward equipment, product development, salaries, and outsourced specialists.
That first contract allowed us to become self-sustaining. We’re now funding further growth with revenue from our projects in Kazakhstan.
– Are you currently looking to raise funding?
Temirlan: – Yes, but we haven’t opened the round yet. We want to increase our valuation first using our own funds by growing the client base and building up our expertise. We plan to raise up to $5 million. It’s difficult to find that kind of venture funding in Kazakhstan, so we’re looking toward the US.
Anuar: – We’re also currently in talks to raise a smaller $300,000–500,000 investment for our US company. We need the funding to establish ourselves in the US, build a sales team, secure a pilot project, and land our first customer there.
“It’s hard to break into the US market without the right introductions”
– Why did you choose the US market for expansion?
Temirlan: – The US market offers a completely different level of earning potential. For example, even a small cereal factory in the US can lose $100,000 for every hour of downtime. In Kazakhstan, you’d only see figures like that at a large industrial plant. But implementation there would take much longer. So from a business perspective, working with US clients makes more sense for us.
– What’s your plan for cracking the US market?
Anuar: – We’ve already registered a legal entity in the US. We’re now working on visas, raising investment, and plan to move there by the end of the year.
We’ll start with small and mid-sized manufacturers. They’re not a priority for the big automation companies, but they still make enough money to afford our platform.
In the US, our advisor is Sean Harley, who spent 25 years at Autodesk. He knows the digital twin space inside out and has deep experience in the industrial sector. His introductions can help open doors for us in the US, because breaking into the market without them is difficult. Once we have our first case study with proven financial results, selling should become much easier.
– What will happen to your projects in Kazakhstan?
Anuar: – They’ll continue to grow in parallel. All of our development will remain in Kazakhstan. In the US, during the pilot stage, we’ll only need one specialist on site to create the visualization and connect the platform to the factory’s existing systems.
Temirlan: – Kazakhstan will remain a priority for us in any case because we’ve already built a reputation here, and it’s starting to work in our favor. We also don’t want to let our clients down. The local ecosystem helps us a lot in that respect. For example, we got into the German Constructor Start accelerator after Astana Hub recommended the program to us. The cluster really keeps us on our toes: they invite us to events and introduce us to potential partners. We’ve even made sales through some of those meetings. The office space and tax incentives are also important for us. Given the size of our contracts, that support makes a real difference and allows us to invest more in growing the team.
– What other markets are you looking at besides Kazakhstan and the US?
Temirlan: – We also considered the EU. But during an accelerator program in Germany, we realized that the European market is highly bureaucratic. Everything takes a long time, and planning can take longer than the actual implementation. There are also strict standards and compliance requirements. So for now, we’ve put our plans to enter the European market on hold.
We’re also looking at the MENA region. Perhaps once we have a few successful US case studies, we’ll return to expanding into new markets.
“We want to reach a $100 million valuation within the next two years”
– How do you plan to grow the business?
Anuar: – Because our visualizations are created using geospatial data and engineering-level accuracy, we can train robots in a 3D simulation of the factory. They can then be deployed in the real facility with only minimal additional fine-tuning.
Another area we want to develop is workplace safety. A digital model of the factory can be used to train employees on how to respond to emergencies.
We’re also working on integrating QazTWIN with warehouse systems to make the long and expensive process of sourcing and finding spare parts for equipment more transparent. Based on predicted breakdowns, the system will be able to flag in advance what needs to be purchased so the right parts are already in stock.
– What are your plans for the next few years, and what’s the bigger vision for QazTWIN?
Temirlan: – In a few years, we see QazTWIN as an international company with clients in Kazakhstan, the US, Europe, and the Middle East.
We have big plans for R&D. We’ll reinvest available funds into researching new technologies for industry, with the goal of eventually rolling them out around the world.
Our product should become the next generation of industrial management systems. We want it to do more than just tell you what happened and where. It should also be able to answer two key questions: “What happens next?” and “What’s the best action to take right now?”
Anuar: – Over the next two years, we’d like to reach a valuation of at least $100 million.
Long term, we see QazTWIN becoming a unicorn. The industrial market is growing fast, and there still aren’t that many ready-made solutions out there.
Those ambitious goals feel achievable because we’ve built an exceptional team of people who put a huge amount of energy and talent into the product every single day.