Yandex Cloud results for the first six months of 2025

Yandex Cloud’s results are not just some financial figures. As one of the leading providers, we traditionally share the key market trends we spotted in the first six months of 2025.

Yandex Cloud’s revenue grows

Yandex Cloud is part of the Yandex B2B Tech business group. In the first six months of 2025, the Yandex B2B Tech’s overall revenue by its two key areas, Yandex Cloud and Yandex 360, amounted to around ₽21B. As compared to the same period last year, this is a 53% rise. This growth rate is twice as high as that of the Russian enterprise solution market where Yandex B2B Tech operates.

These first six months of 2025 marked three important achievements of ours: YCDR for SOC and SIEM cloud-based solutions, public access to Yandex BareMetal, and YTsaurus for working with any amount of data. These solutions are especially in-demand for those clients who have strict requirements to security and infrastructure (first and foremost, for banking and large businesses). We regard these products as having large potential and will continue developing them in a consistent way.

The number of Yandex Cloud clients in the first six months of 2025 exceeded 47,000. Large businesses steadily increase its cloud service consumption: over the period in question, their share amounted to 53% of total revenue. Meanwhile, the share of large and mid-sized business in overall cloud consumption reached 85%.

Most industries saw a significant increase in demand for information security services. Banking, insurance, production, oil and gas, retail, and ecommerce companies use such services to protect their websites and apps against DDoS attacks, as well as to monitor incidents and respond to those. Such companies also started combining two different modes for using our services: cloud-based and on-premise. Fintech companies turned more to the data platform, for example, when it comes to quickly processing information in high-load apps. Insurance companies are more focusing on automating routine tasks and raising the customer service quality. Meanwhile, retail companies see comprehensive analytics as their top priority, since the businesses need to process data from various systems in real time, as well as maintain its quality and provide access to it in a fast way.

The number of services available on the Yandex Cloud platform amounted to 75. Over the first six months of 2025, the share of IaaS reached 52%, while the share of the platform-as-a-service (PaaS) tools (including data platform, container-based development, ML, and security tools) is 42%. Meanwhile, the share of ML services increased.

Businesses massively use ML

Businesses get more and more interested in machine learning, as companies are actively embedding AI into their workflows. Over the first six months of 2025, the revenue for Yandex Cloud ML services increased by 160%. This is due to both the high demand for AI tools and the new on-premise mode, when you can deploy ML services in your own infrastructure.

The Foundation Models consumption growth rate increased by 130%, which caused us to contribute more to developing this product. As such, we:

• Added new YandexGPT versions. With YandexGPT 5 Pro, our clients can create AI assistants on their own; this includes assistants that run smart search through company documents or databases.
• Opened access to open-source visual generative models (VLMs): Deepseek VL2 Tiny, Gemma 3 27B, and others. These models allow companies to create product descriptions based on images or quickly find info in documents. We also opened access to Qwen3 235B, the largest LLM in the Russian market.
• Launched LoRA, a method for quick neural network tine-tuning. This allows businesses to leverage multiple fine-tuned neural networks, easily switching from one to another or using them together for various business scenarios.

We are actively developing both Foundation Models and AI Studio to expand the coverage for business scenarios of using generative models. Currently, we are heading towards building agents and agent scenarios; with this in mind, we launched Assistants API with built-in RAG and added generative responses in SearchAPI.

We are also working on services for automating hands-on tasks. In particular, we:

• Added meaningful tags, search, and other new features to Yandex SpeechSense. Now, apart from analyzing conversation by words and sentences, it can also do so based on the context and meaning. This significantly simplifies hypothesis check and analytics collection. Apart from that, the search screen now has the summarization option.

• Launched Yandex Neurosupport, an assistant for call center agents. It acts as a copilot by analyzing customers’ text queries and generating smart suggestions.

In the first six months of 2025, 25,000 clients used YandexGPT, and 7,000, YandexART. As for Yandex SpeechKit, a speech recognition and synthesis tool, it was used to voice over more than 77,000,000 minutes of text.

Enabling more protection: Developing the security service ecosystem

The security services show sustainable growth. In the first six months of 2025, every fifth Yandex Cloud paid version user leveraged them. The revenue for security services increased by 90% YoY. The fastest-growing one is Yandex Smart Web Security that helps businesses protect web apps from external threats.

We partnered with SolidSoft, one of Russia’s largest developers of firewalls for web apps, to create business solutions in the cyber-security domain. Yandex Cloud and SolidSoft will join the features of SolidWall WAF and Yandex Smart Web Security to provide the clients with all-round protection.

To respond to the growing demand for information security tools, we launched a few new products:

• YCDR: Service for monitoring and responding to cyber incidents. With Yandex Cloud Detection and Response, the security team can promptly detect cyber-attacks and other threats. This service is based on our security operation center (SOC), and its algorithms enable detecting anomalies in information systems in a faster way.

• Yandex Identity Hub: Service for centralized user and access management. It protects businesses against compromising employee credentials and accounts. Delivered as SaaS, it can be used in organizations with various infrastructure types: on-premise, cloud-based, or hybrid.

We are actively using AI technologies in our security services. In Yandex Security Deck, ML models help reduce the risk of sensitive data leaks and enable transparency through analyzing employee logs. In YCDR, ML assists analysts in ticket categorization and summarization. Meanwhile, Yandex Smart Web Security leverages ML to protect the resources against L7 DDoS attacks and bots.

Efficient data management

The data services are steadily growing, with the revenue increasing by 40% in the first six months of 2025. Yandex Data Processing, Yandex Managed Service for Apache Airflow^®, Yandex Data Transfer and Yandex Managed Service for ClickHouse^® are those that demonstrated the largest growth.

We launched YTsaurus, an all-round platform to work with any amount of data. YTsaurus is available both on-premises and as a managed cloud service and can be used to analyze exabytes of enterprise data and train sophisticated ML models with billions of parameters.

We also launched two new data services that are currently at the Preview stage:

• Yandex Managed Service for Trino enables building data lakes. It provides high-performance and scalable access to heterogeneous data sources and allows you to quickly run analytical queries in real time without any need for prior joining or conversion.
• Yandex Managed Service for Apache Spark^™ enables udating Apache Spark^™ versions with minimum impact on the customer environment. It can be flexibly adjusted to various requirements through integrating external libraries.

Additionally, we are working on Yandex’s in-house database management system, YDB, that is also available both in the cloud and on-premises. In particular, we added vector search that enables working with AI models. This kind of search allows YDB users to work with unstructured data (80% of all data), such as video, audio, image, and sensor data. When it comes to AI and ML-based solutions, vector search is a very important feature (e.g., for RAG).

The demand for data analytics and visualization tools is steadily growing, especially among businesses. In response to that, we released a few updates to Yandex DataLens:

• DataLens Editor: Tool that allows you to customize charts and tables using JavaScript. We are also working on a native LLM-based assistant that will help create charts using text descriptions.

• DataLens Gallery: Public library of out-of-the-box dashboards and visualizations for various domains and tasks. It enables users to view the best examples and share their own solutions with the community.

• DataLens Enterprise: On-premise version of DataLens that is delivered based on a subscription or a perpetual license. It targets large businesses with distributed infrastructure and dozens of thousands of users.

Flexible and reliable: New infrastructure management opportunities in Yandex Cloud

We are giving a special focus to developing our infrastructure tools and increasing reliability of our products. In the fist six months of 2025, we released a few updates that foster sustainability, flexibility, and easy management. They include:

• Switching traffic across availability zones: Now you can temporarily disable traffic coming to one or more zones at the Yandex Network Load Balancer и Yandex Application Load Balancer level. This enables continuous app operation during testing, maintenance, or production issues.
• Increased reliability: We published best practices for designing a fault-tolerant infrastructure and a tutorial to simulate peak loads and availability zone failure for testing purposes.
• Cloud Shell: Russia’s first browser-based command line shell in a cloud provider console. This tool allows you to use a full-featured command line without installing or configuring anything locally.
• Public access to Yandex BareMetal that enables renting dedicated physical servers, thus getting additional capacity for IT infrastructure development and designing new digital products. With BareMetal, you can flexibly implement hybrid IT scenarios by running high-load apps, testing new products, or deploying systems within completely closed environments. By integrating Yandex BareMetal with the Yandex Cloud environment, businesses can combine various types of infrastructure and work with those as a single tech stack.

SourceCraft: Platform for shared secure development

We are actively working with developer tools by creating an environment that would cover the entire development life cycle, from writing code to its deployment and maintenance. With this in mind, we introduced:

• SourceCraft: Platform for shared development projects with CI/CD tools, smart project navigation, native AI assistant, and advanced features for teamwork management. SourceCraft also received security tools to protect software products, GitHub repository mirroring, and integration with other apps through API.

• SourceCraft Code Assistant: Native to SourceCraft, this tool helps you generate, complete, and understand code. It is already used by thousands of developers working in various domains, from retail and finance to manufacturing. In 2025, Code Assistant received the chat mode where you can ask technical questions about the project and get code suggestions.

Partner program growth

We are expanding our partner ecosystem and continue fostering our collaboration with integration and developer companies, resellers, and other businesses that have certified experts and proven track record of implementing Yandex Cloud projects.

In the first six months of 2025, the number of active partners increased by 30% as opposed to the same period last year, amounting to 659 companies; while the partner channel revenue rose by 50%.

When working with our technology partners, we focus on AI. In particular, we launched a joint GPT lab with Jet Infosystems, singed an agreement with Lansoft on embedding our AI services into their products, and started working on our joint AI agenda with SoL Lab.

Education initiatives: Expanding courses, workshops, and certifications

In the first six months of 2025, over 13,000 people mastered cloud technologies through Yandex Cloud’s education programs that are available both on our own software and on our partner websites, such as Practicum, Netology, OTUS, SkillFactory, Slurm, and others. Since we started our programs, around 73,000 people completed the courses.

In these six months of 2025, we launched five new public online courses on DevOps, cyber security, machine learning, and data. Currently, there are 30 courses available.

2025 saw a new Yandex Cloud education format: hands-on workshops. Now the students can learn new topics through a workshop, while also using the tutor’s feedback. These workshops are available for the following topics: ClickHouse^® for data engineers, Greenplum^® for data engineers, and Kubernetes^® for developers.

In the first six months of 2025, we launched three new certification tests: YDB Administrator, YDB Developer, and DataLens Analyst. Now there are five certification tests available. Over the period in question, 216 people took tests and 66 got certified.

We also ran research on how Yandex Cloud certification impacts professional growth and business. The result showed that certification is not only about confirming one’s knowledge; it also helps structure one’s experience, raise confidence, expand the area of responsibility, and get a better position on the job market.

In the first six months of 2025, over 7,000 university students used Yandex Cloud technologies for studying. We are still supporting around 95 universities, including HSE, ITMO, Saint Petersburg State University, and Moscow Institute of Physics and Technology, with our cloud techs. Meanwhile, our Cloud Academy has had 210 students starting this year.

Developing Yandex Cloud Boost

Yandex Cloud Boost is a program that helps tech teams to launch and scale up their products. In the first six months of 2025, 160 startups received grants totaling more than ₽77M.

Kazakhstan’s attention to Yandex Cloud Boost has been especially high, with Astana Hub becoming the most active partner in this country. Meanwhile, startups are mostly interested in AI projects with generative models.

Yandex Cloud Social Tech for healthcare and sustainability

Yandex Cloud Social Tech continues launching initiatives for solving hands-on tasks in healthcare and environment:

1. The NMRC of Oncology named after N.N. Petrov started using a YandexGPT-based solution to process applications on clinical trial ethical approvals. Now the initial document verification takes just a few minutes as opposed to a few weeks.

2. Along with the Far Eastern Federal University’s researches and Data Analysis School’s students, we created and trained a neural network that detects the volume, weight, and type of waste. It helps both the environmental monitoring services and volunteers to simplify waste collection on coastal areas.