Contents
In 2023-2024 the integration of artificial intelligence technologies into business processes has become a common trend in global markets. Commercial companies started using high-performance computing more often to gain competitive advantage. According to McKinsey, a US consulting agency, the rate of AI adoption in professional activities has increased to 72% in 2024. For comparison, in 2023, this figure was 55%.
Among the first to actively embrace artificial intelligence technology were business entities operating in iGaming.
It is interesting to note, that by 2024 companies working in the gaming industry have already developed a positive experience of cooperation with European service providers in organizing infrastructures for artificial intelligence. In this article we will talk more about how machine learning technologies are used in gambling and how such cooperation can develop.
How artificial intelligence is used in gaming
The organization of business processes in gaming companies often involves a rather complex structure of interaction between different divisions, departments and services. The introduction of artificial intelligence helps not only to solve technical issues related to the operation of applications and services, but also makes it possible to optimize the exchange of information between different teams within the company itself: to structure large amounts of data and to form effective marketing strategies.
Supporting the legitimacy of financial transactions and security mechanisms
Security is one of the most important areas where the introduction of artificial intelligence has played a significant role.
Every day gaming companies are faced with the need to process thousands of transactions, each of which requires complex evaluation. It is simply impossible to do such work without specialized software. AI systems perform end-to-end analysis of each transaction and identify suspicious activity. Neural networks process deposit data, help check payments for compliance with international AML and FATF policies, and block transactions from users with signs of gambling addiction and from minors.
Artificial intelligence systems protect web services from hacker and phishing attacks, tampering attempts and program code replacement. AI analyzes users’ actions, helping to combat multi-accounting and abuse of bonuses, and identifies signs of bots and the use of illegal software.
Artificial intelligence for marketing in gaming
Another one of the most notable applications of artificial intelligence in iGaming is the evaluation of the gaming experience. Neural networks track user behavior, assess their preferences, and take into account sizes of their bets.
Based on the data obtained, the system makes recommendations to marketing experts and helps to create personalized advertising offers. In some cases, neural networks not only make recommendations, but also automatically select for users the products they like most to make their gaming experience more vivid.
All collected data undergoes additional analysis, the company’s managers plan promotions and develop bonus offers on its results. Based on the information obtained, software developers also create new products that better meet the expectations of players.
Chatbots and conversational AI in gambling
Chatbots with artificial intelligence have long been a typical solution not only for gaming applications, but also for any web service with online support. In the gambling industry high quality of support work is a prerequisite for effective competition. Conversational AI answers typical queries around the clock, helps solve technical issues, informs users about upcoming promotions and available bonuses.
Some gambling companies may have tens of thousands of users online every day. Transferring the routine work of processing requests to neural networks helps to significantly reduce personnel costs and reduces the load on support staff.
How gambling companies save up to 30% on running AI servers
Artificial intelligence tasks require high computing power with expensive graphics processing units (GPU). For this reason, cost optimization at the project planning stage is relevant even for large businesses.
Cooperation with service providers allows gaming companies to save budget even before the server is launched. This is due to the fact that most European providers have direct contracts with the world’s leading manufacturers of components and can offer their clients hardware at 20-30% cheaper.
At the same time, gambling companies quite rarely purchase GPU servers for AI as property. It is much more profitable for them to rent hardware with the buyout option (leasing).
The contract is concluded for 3-5 years and every month they pay the service provider an amount equal to 2-3% of the total cost. So, for example, if the cost of the server is €20,000, the monthly payment will be only €400. When the term of the agreement ends, the hardware becomes the client’s property.
What GPU servers are used for AI systems in iGaming
The architecture of any GPU server is very similar to the classical one with CPU-processor; only a block with GPU chips is additionally installed for parallel calculations.
- Graphics processing units (GPU) are selected based on the business requirements of the project. In the first half of 2024, NVIDIA L4, NVIDIA L40, and NVIDIA L40S chips are the most popular for neural networking. NVIDIA HGX™ H100 is recommended when maximum performance is required.
- The best CPUs for AI servers remain the 5th generation Intel® Xeon® lineup and AMD’s AMD EPYC™ 9004 chips.
- In all cases, the choice of Random Access Memory is DDR5 SDRAM – the fifth generation of RAM with double data transfer speed. The amount of RAM usually ranges from 512 to 1024 GB.
- NVME solid-state drives with the latest NVMe Gen5 interface are still the only option for high-performance computing. The number of drives and total storage capacity directly depends on the technical requirements of the project.
- The form factor of the server depends on the configuration of components selected by the client. Server size is measured in rack units called Units, abbreviated as U. For AI work, servers are most often designed with heights between 1 and 8 U.
When gaming companies approach a service provider, they usually already have an idea of what kind of server configuration is needed. However, high-performance hardware becomes obsolete very quickly. Before assembly, the provider’s technical specialists always check the configuration and it often turns out that it is possible to use more modern components without increasing the budget.
The agreed-on project is additionally sent to the graphics processor manufacturer for evaluation (because GPU is the most expensive and most important part of the server), and the latter makes recommendations on the use of software and hardware maintenance.
So, an actual GPU server architecture for AI might look like this:
- HPE DL380a Gen11 4DW.
- 2x Intel Xeon-Gold 6444Y 3.6GHz 16-core.
- 8x HPE (1x64GB) Dual Rank x4 DDR5-4800.
- 2x HPE 7.68TB NVMe.
- Intel E810-XXVDA2 Ethernet 10/25Gb 2-port SFP28.
- 4x NVIDIA L40 48GB PCIe.
- 4x HPE 1800W-2200W.
This is a real example of a server that was launched for the business needs of a company working in the iGaming industry. The total cost of all components for the project was about €100,000, but by working with a service provider, the client was able to save on the cost of the hardware and received free expert evaluations and recommendations, including from NVIDIA technical specialists. In addition, the server was purchased under a leasing contract, which allowed the gaming company to optimize costs.
Additional infrastructure scaling options for gambling companies
In the process, almost every gaming company faces the need to scale their infrastructure. This is not always related to the change of GPU server performance. New applications and services are emerging, and new resources are required to organize internal business processes.
In such cases, service providers can offer a variety of opportunities, from launching new dedicated servers to attracting additional computing power through the use of cloud services.
Launching a new dedicated server often becomes cheaper due to already existing contracts for hardware maintenance or renting space in the data center. The time for preparation and launch is significantly reduced due to the fact that the service provider is already familiar with the specifics of the client’s infrastructure and may not spend resources on audits and additional checks.
Setting up and running a virtual machine or connecting to cloud services can take a few days, and sometimes even hours, after the project is approved. The advantage of using cloud technologies is that they allow you to quickly increase computing power and storage volumes on an almost unlimited scale.
At the same time, you have to remember that cloud resources are available only as part of a rental service. In the long run they are more expensive than working on your own hardware. For this reason, cloud services are often used as a temporary solution when, for example, tests need to be carried out or reserves are required for a repair period.
Technical and maintenance services of GPU servers for gaming
Designing, configuring and running an AI server is a rather complex process that requires the attention of experienced engineers not only at the beginning, but also at the subsequent stages of work. Often the provider’s agreement with gaming companies provides for full server maintenance. Such a service has several advantages at once:
- The client gets the opportunity to engage in their own business tasks without having to worry about maintaining infrastructure availability.
- The provider’s technical staff monitors hardware performance and makes repairs, handles licensing issues and updates software.
- The provider itself purchases new hardware directly from the manufacturer, and stores it in its warehouse, if necessary.
- The technical staff of the client company receives administrative access to server resources via a secure VPN.
Some service providers offer post-warranty maintenance service on a subscription basis. The idea is that the provider replaces any failed components (even the most expensive and discontinued ones) free of charge, provided that the client pays a monthly subscription fee. This is especially beneficial in cases where a leasing contract for hardware rental with a term of 3 to 5 years has been concluded.
There are also other convenient services that are often used by gaming companies entering into an agreement with European providers.
- Backup and disaster recovery. As part of this service, the client can receive both: additional infrastructure for organizing and storing backups, as well as a comprehensive hardware and software solution for creating and storing backups on the provider’s hardware.
- Services for server migration or relocation to a new data center. This service can sometimes be obtained free of charge as a bonus to other agreements if the hardware is relocated within the geographical boundaries of a particular country or region.
- Physical connection of infrastructure to Internet exchange point. In Germany, service provider will organize the connection to the world’s largest Internet exchange point DE-CIX in Frankfurt am Main free of charge for its clients if they have valid contracts.
- Free Internet with a channel width of up to 100 Mb for a specific project is another bonus you can get from some European providers.
- Organization of a private cloud and connection to public cloud services. This service is often used by fast-growing gaming companies and startups that need to scale their infrastructure.
Consultant on hosting solutions, network and cloud technologies A 30-minute valuable consultation where you’ll find flexible improvement options for your infrastructure. 
Practical experience shows that for companies operating in the iGaming industry, a wide range of services and facilities is not the only important criterion for evaluating the provider’s performance. The quality of service and responsiveness of the support service is of great importance. If requests are received only by e-mail and are processed within several days, it often becomes a decisive argument in favor of refusing to work together.
Some European providers are taking market requirements into account and offer support services in multiple languages with the ability to communicate via messengers.
If you need infrastructure to launch an AI project but don’t know where to start, sign up for a free 30-minute consultation with our portal experts. You’ll learn which hardware is best suited for your needs and what bonuses you can get when you enter into an agreement with a service provider to run and maintain a GPU server for AI.
Article author Olga Boujanova Consultant on hosting solutions, network and cloud technologies
