The generative AI platform Yandexgpt from Yandex: Progress in artificial intelligence and robotics
Xpert pre-release
Language selection 📢
Published on: May 31, 2025 / update from: May 31, 2025 - Author: Konrad Wolfenstein
The generative AI platform Yandexgpt from Yandex: Progress in artificial intelligence and robotics-Image: Xpert.digital
From Russia to the world: Yandex 'way to the leading technology brand (reading time: 39 min / no advertising / no paywall)
How Yandex conquers global markets with international spin-offs
Yandex has established itself as a leading company in the field of artificial intelligence (AI) and robotics and continuously invests in innovative technologies. Remarkable developments include the generative AI platform Yandexgpt, pioneering systems for autonomous vehicles as well as versatile robotic solutions for logistics and delivery. This progress is evidence of Yandex 'long -term engagement in technological innovation.
A decisive change in the corporate strategy is the extensive restructuring, which led to the establishment of new internationally operating spin-offs such as Nebius, Avride and Yango. These independent companies are now driving the global marketing and further development of the technologies developed by Yandex. This enables a dual orientation: While the Yandex LLC continues to take a strong position in the Russian market and drives research and development, the international reach is also growing through the new spin-offs.
This article illuminates the technological core areas of Yandex, including groundbreaking progress in AI and robotics, as well as the strategic partnerships and comprehensive research and development initiatives. The company acts in a dynamic market environment and positions itself both locally and globally as a key figure in modern technology development. The restructuring and the innovative move into new markets illustrate that Yandex consistently aims to consolidate its leadership role in these future -oriented areas.
Suitable for:
- The strategic importance of artificial intelligence and robotics for Russia (reading time: 72 min / no advertising / no paywall)
Yandex 'strategic focus on AI and robotics
Yandex has established itself as a leading technology company, the strategic core of which is deeply rooted in the areas of artificial intelligence and robotics. This orientation is not just a secondary aspect of the corporate strategy, but a fundamental pillar for innovation, market differentiation and future growth. The latest corporate restructuring and the resulting international spin-offs have further refined this strategy and gave it a new global dimension.
Confession for innovation and market positioning in advanced technologies
Yandex is known for its advanced technologies and innovations in the areas of AI, machine learning (ML) and processing of natural language (NLP). This technological expertise enables the company to deliver high -relevant search results and personalized recommendations. The company continuously invests in research and development (F&E) in order to improve its technologies and services, to stand out from the competition and to meet the changing needs of the users. This basic obligation to KI/ML as a core element of the business is crucial for understanding the long -term vision and the competitive position of Yandex.
Yandex's growth strategy includes the diversification of services beyond the core search engine, international expansion, strategic partnerships and a focus on mobile and emerging technologies such as AI and language search. AI and robotics are therefore an integral part of Yandex's wider growth and diversification strategy, which aims to create a comprehensive digital ecosystem.
In addition, Yandex's AI development is in line with national strategic goals. The national strategy of Russia for AI in 2019 is entering large technology companies such as Yandex and aims to significantly improve the position of the country in the ACI area by 2024 and to achieve a global leadership role in certain areas by 2030. This indicates potential state support and a defined role in the technological further development of Russia. An important aspect is the development of localized AI models that are tailored to the Russian language and the Russian market.
Effects of corporate structure and spin-offs (Nebius, Avride, Yango, Toloka)
The corporate landscape around Yandex has recently changed significantly, which has a direct impact on the strategy in the field of AI and robotics. In July 2024, Yandex NV (now Nebius Group) sold the Yandex LLC (the Russian unit) and retained several business areas operating outside of Russia. The Nebius Group, headquartered in Amsterdam, now has Nebius.ai, Toloka, Avride, triple and minority stake in other companies focused on AI. This development is of central importance, since the original technology and talent of Yandex in the field of AI and robotics are now largely promoted internationally by these spin-offs. It is therefore essential to clearly distinguish between the ongoing activities of the Yandex LLC in Russia and the strategies of these international units.
Nebius, which emerged from the splitting of Yandex, originally planned the structure of sovereign clouds, but then turned over to drive the KI “gold rush” by rebuilding its cloud platform for AI workloads. The company builds on one of the largest commercially available AI infrastructure transactions based in Europe. This represents a significant international advance into the Ki-Cloud infrastructure that uses the basic cloud expertise from Yandex. Partnerships, such as those with Nvidia, and investment plans ($ 1 billion in European AI infrastructure until mid-2025) underline this ambition.
In the area of autonomous vehicles, the fact that Yandex AV (Autonomous Vehicles) ceased the company in the west due to sanctions to found Avride, an external company that acquired the assets of the Yandex unit for autonomous driving. Avride and Yandex had an agreement for the joint use of technology until the beginning of 2025, after which their development channels diverge. This illustrates how geopolitical factors directly influenced the corporate structure and strategy and led to the internationalization of Yandex's autonomous driving technology under a new brand.
Yango Tech, part of the global Yango Group (which seems to be connected to the wider Yandex ecosystem after the restructuring), offers B2B technology solutions, including Yango Tech Robotics (warehouse automation) and Yango Tech Autonomy (delivery robot). Yango seems to be the international commercialization low for various robotics and AI-based B2B solutions that were developed within the Yandex ecosystem.
Toloka, a company for AI data solutions, received a strategic investment led by BezoS expeditions, whereby the Nebius Group retained an economic majority stake, but made the majority voting rights to enable Toloka greater independence. This underlines a strategy of gaining external investments and specialist knowledge for specialized AI service units that enables them to scale independently, while Nebius focuses on the core AI infrastructure.
These restructuring indicate a conscious strategy of decoupling valuable AI and robotics assets from the geopolitical and operational restrictions that the Russian-centered Yandex is exposed. Yandex developed significant AI/robotics technologies, but was faced with sanctions and geopolitical pressure that impaired international operations. As a result, important international assets such as Nebius and Avride have been spun off and maintain international investments (e.g. Amsterdam for Nebius, Austin for Avride) and actively international investments (NVIDIA at Nebius, Bezos at Toloka, Hyundai near Avride). This is not just a sale of assets, but a strategic maneuver to survive, the continuous innovation and global market access of the most promising KI- and robotics technologies in Yandex, shielded from the restrictions of Russian unity.
The creation of separate units also enables the risk and specialization of the focus. Yandex had a variety of AI/robotics projects. After the restructuring, independent companies were created: Nebius (Ki-Cloud), Avride (Robotaxis), Yango Tech Robotics/Autonomy (B2B robot), Toloka (AI data). Each of these units has its own CEO and strategic partnerships (e.g. Avride CEO Dmitry Polishchuk, Hyundai partnership; Toloka CEO Olga Megorskaya, Bezos-Investment). This structure enables each unit to pursue their specific market opportunities more aggressive and attract investors who are interested in this special niche instead of watering down the focus within a larger conglomerate. It also isolates financial and operational risks to individual units.
Although legally separated, these spin-offs carry the technological DNA and often also key personnel from Yandex, which gives them a lead in their respective international markets. Avride acquired assets from the Yandex unit for autonomous driving and initially shared code. The NEBIUS team helped build the Yandex cloud stack. Toloka was created within Yandex. Yango Tech solutions such as the picking robot and delivery robots are based on technologies that were probably developed or initiated within the Yandex F&E department. This implies that the innovation culture built by Yandex and the technological foundations are now used and adapted by these new units for a global audience, which creates a “distributed Yandex” effect in the AI and robotics area.
The following table is used to clarify this complex structure:
Important Yandex-connected companies in the field of AI and robotics after restructuring
Important Yandex-connected companies in the field of AI and robotics after restructuring-Image: Xpert.digital
After the restructuring of Yandex, several companies were created that work in the field of artificial intelligence (AI) and robotics. Yandex LLC, based in Moscow, remains the central actor for AI development in Russia, including Yandexgpt, Alice and Robotic Research. Important partnerships exist with HSE, MIPT and Skoltech as well as Russian institutions such as the post office. The NEBIUS Group NV, under the direction of Arkady Volozh, acts as international holding and developed from Yandex NV after the Russian assets were sold. NEBIUS.AI with headquarters in Amsterdam focuses on AI-centered cloud infrastructures, GPU cluster and investments in European AI technologies. The company uses Yandex experience in cloud and data centers. Avride, based in Austin, develops autonomous driving solutions and has partnerships with Hyundai and Uber, whereby it emerged from Yandex technologies for autonomous driving. Yango Tech, as part of the Yango Group in Dubai, offers international robotics and AI solutions for warehouse, autonomy and retail, whereby its roots are probably based on Yandex technologies. Toloka Ai is now an independent unit under the NEBIUS group with a focus on AI data solutions and serves customers such as Amazon and Microsoft, whereby it was originally a Yandex platform.
Pioneering artificial intelligence
Yandex has established itself as a pioneer in the field of artificial intelligence, with significant developments in generative AI, extensive research and the integration of AI solutions into a wide range of products and services. These efforts extend from basic voice models to specialized applications and open source contributions.
Yandexgpt and generative AI: development, skills and ecosystem
Yandexgpt, a neural network from the GPT family developed by Yandex LLC, is at the center of the company's generative AI strategy. It is able to create and revise texts, generate new ideas and to record the context of user discussions. The training is based on an extensive data set of books, magazines, newspapers and other open sources on the Internet.
1. Nuclear technology and development (Yandexgpt versions)
The development of Yandexgpt testifies to fast iterations and a high strategic priority. First announced as Yalm 2.0 in February 2023, the model was officially presented on May 17, 2023. As early as September 7, 2023, Yandexgpt 2 was presented, which, according to the developer, provided better results in 67% of cases than the previous version. The stable version Yandexgpt 5.0 was released on February 25, 2025. This rapid sequence of versions indicates continuous and significant improvements.
2. Integration into service (Alice, Shedevrum) and business offers (Yandex Cloud)
Yandex pursues a strategy of deep integration of its advanced AI. Yandexgpt is integrated in the virtual assistant Alice -a counterpart to Siri or Alexa -and available in various Yandex services and applications. Alice recorded an estimated 66 million monthly users in Russia, which creates a wide user base for the AI functions.
In June 2023, Yandexgpt was integrated into the Shedevrum image generation app, which enables users to create complete contributions with titles, text and suitable illustration.
For the B2B sector, Yandex offers companies via its public cloud platform Yandex Cloud access to Yandexgpt's API. In July 2023, 800 companies already took part in closed tests. You have two modes (API and Playground-an interface in the Yandex Cloud console for testing models and hypotheses) and two model versions: an asynchronous version for complex tasks and one for quick real-time answers. This demonstrates a clear strategy to use advanced AI to improve the user experience in end customer products as well as for monetization via B2B offers.
3. Applications in education and other sectors
The versatility of Yandexgpt can be seen in practical and effective applications beyond the general chat. Yandex Education has introduced Yandexgpt-based tools for teachers who help to create tests, to review tasks, to recognize the plagiarism and to create teaching plans. A pilot project with the Higher School of Economics (HSE) made it possible to use Yandexgpt for their theses.
In a remarkable application, Yandex and the Moscow Institute of Physics and Technology (MIPT) trained a AI to evaluate the work of the student Olympiad “Lomonossow tournament”, which reduced the examination time of three months to one day.
Since October 2023, Yandexgpt has been able to create short summaries of Russian-speaking online videos. In the medical field, Yandexgpt 5 Pro is used by the NN Petrov National Medical Research Center for Oncology to process documents for clinical studies, which shortens the verification time from weeks to minutes. These applications underline the ability of Yandexgpt to solve complex problems in specialized specialist areas.
AI research and development landscape
Yandex maintains a robust internal F&-E-Department, Yandex Research, which focuses on fundamental problems of machine learning and contributes to both its own products and the broader scientific community.
1. Yandex Research: core areas (NLP, computer vision, ML) and key projects
Yandex Research focuses on computer vision, autonomous vehicles, NLP, language technologies as well as search and recommendation systems and publishes results at leading conferences.
In the NLP area, research includes representation and understanding of language. The projects include Specexec (massive parallel speculative decoding for LLMs on consumer devices), Petals (decentralized inference and fine-tuning of LLMS), sequoia (scalable speculative decoding) and PV-tuning (extreme compression of llms).
Computer vision research focuses on image calls and generative modeling. Projects include invertible Consistency Distillation (ICD) for text-controlled image editing, adaptive teacher students Collaboration for text-conditional diffusion models and CAP (Correlation-Aware Pruning for Sparse Vision models).
Yandex Research also provides data records, such as the Shift's Dataset (for distribution shifts in the event of weather forecast, translation and movement forecast) and a text-to-image data set for the search for similarity in the billion dollar area. These specific projects indicate a deep commitment to optimizing LLMS and advanced image models.
2. Academic cooperation and talent development (e.g. HSE University)
Yandex works closely with universities to promote innovations and develop talents. There is an important partnership with the Faculty of Computer Science (FCS) of the HSE University, where a joint laboratory for data science has been set up, which focuses on research and training.
This cooperation is to be expanded in the next ten years, with new educational programs, advanced AI research and research into generative AI in the educational process. A joint doctoral program in the AI area is also planned. The Yandex Research ML Residency Program in cooperation with HSE offers another opportunity to promote talent. These collaborations are crucial for the development of a talent pipeline and ensure that Yandex remains at the forefront of AI research, especially in Russia.
Applied AI: Transformation of products and services
AI is the heart of many products and services from Yandex and drives both efficiency and relevance.
1. AI in advertising, search and personalization
Yarnit, an offer tailored to Yandex, uses AI for advertising solutions, including keyword research, text position, localization and creative design for the Yandex market. Yandex itself uses KI/ML to improve search algorithms, increase the accuracy of the results and to provide personalized content. This is central to the advertising business and the core product, the search engine.
2. Yandex Alice: The conversational AI assistant
Alice, the virtual assistant from Yandex, integrates Yandexgpt and has an estimated 66 million monthly users in Russia. Alice works in Russian and English. However, questions about their autonomy and the potential state influence on answers to political questions were raised. Alice is an important point of contact for the AI of Yandex with consumers. Her large user base makes her an important platform for the use of AI functions, while the concerns regarding content control illustrate the complex operating environment for AI in Russia.
Open source contributions and community engagement
By publishing open source software, Yandex makes significant contributions to the developer community.
1. Catboost, perforator and their effects
Catboost is an open source gradient boosting library developed by Yandex, which is characterized by its ability to effectively handle categorical characteristics. It is used by companies such as Jetbrain, Cloudflare and Careem and is enjoying a broad acceptance (approx. 100,000 Pypi installations per day in April 2022). Areas of application include real-time recommendation systems, fraud recognition and predictive maintenance.
Perforator is another open source tool from Yandex for continuous profiling for monitoring and optimizing servers and applications. It can help to reduce the infrastructure costs by up to 20% and support languages such as C ++, C, GO, Rust, Python and Java.
These tools demonstrate the significant contributions from Yandex to the open source community, strengthen the company's reputation and promote broader acceptance of its technologies by solving real problems for developers and companies worldwide.
The development and rapid use of Yandexgpt in such different sectors as education and healthcare position it as a key element of Russia's technological sovereignty in the AI area. Global AI models are predominantly in English, while Russia pursues a national AI strategy that emphasizes independence and leadership claim. Yandexgpt is specifically trained on Russian data and is characterized by its Russian -speaking skills, which is crucial for the domestic market. Integration into critical Russian sectors such as education (Yandex Education, HSE) and healthcare (petrow-oncology center) and API access via Yandex Cloud for Russian companies underline this strategic importance. Yandexgpt is therefore not just another LLM, but a strategic tool for Russia to reduce the dependence on foreign AI, to adapt the AI to the specific cultural and linguistic context and to drive the AI adaptation into business and public services.
Yandex follows a two-track AI strategy: On the one hand, large all-purpose models such as Yandexgpt are developed, on the other hand, highly specialized, powerful tools such as Catboost and perforator are created and provided as open source. Yandex invests strongly in Yandexgpt, while Yandex Research operates top research in NLP and computer vision. At the same time, Catboost (Gradient Boosting) and Perforator (profiling) were developed and source open. These tools address specific, critical needs in ML workflows and system optimization and have found considerable distribution. This dual approach makes it possible to offer Yandex to be a leader in both basic AI and to offer practical, widespread solutions for specific ML/developer problems. The open source provision of these tools probably also serves talent acquisition and brand formation.
While Yandex LLC focuses on Russia, Nebius and his subsidiary Toloka are driven by significant international investments to bring Yandex's expertise from Yandex in the areas of Ki-Cloud infrastructure and data annotation/curation. Yandex developed robust cloud capabilities (Yandex Cloud) and had internal data identification capacities (Toloka emerged from Yandex). Nebius emerged from Yandex's cloud and data center activities and is now concentrating internationally on AI cloud infrastructure. Toloka, now under the Nebius Group with external investments (Bezos Expeditions), offers AI Date Services worldwide. NEBIUS invests strongly in data centers and works with Nvidia for his AI cloud. This signals a strategic step to make the collected knowledge of Yandex in the field of cloud services and data management (essential for AI) accessible to a global audience, under a new, internationally oriented corporate structure (Nebius Group) that attracts considerable capital and partnerships.
The following table serves for a better overview:
Overview of Yandex 'Kern-Ki technologies and applications (Yandex LLC & connected companies)
Overview of Yandex 'Kern-Ki technologies and applications (Yandex llc & connected company)-Image: Xpert.digital
Yandex and its affiliated companies offer a wide range of core AI technologies and applications. Yandexgpt (Yalm), a generative language model of the GPT family, is characterized by text position, revision, idegeneration and comprehension understanding. It is used in areas such as virtual assistants (Alice), content creation (Shedevrum), education, medicine, video summary and via Yandex Cloud in the B2B area. The latest version, Yandexgpt 5.0, was published in February 2025. The NLP research of Yandex includes language representation, LLM optimization (Specexec, Sequoia, PV-Tuning) and decentralized LLMS (Petals), which are primarily used for basic research and the improvement of Yandex products. In the computer vision area, Yandex Research deals with image call, generative modeling and model compression (CAP), which is used in image processing and in text-to-image models.
With Yandex Alice, the company offers a conversational AI assistant who integrates Yandexgpt and supports both Russian and English. Alice has around 66 million monthly users in Russia alone. Catboost, an open source gradient boosting library, is known for your handling of categorical data and your GPU-based training. It is used in recommendation systems, fraud detection and predictive maintenance and is used by companies such as Jetbrain and Cloudflare. Perforator, an EBPF-based open source tool for continuous performance profile, is used to optimize servers and applications and support several programming languages.
Yandex is also active in the area of advertising: Yarnit supports AI-based keyword research, text position, localization and creative design, which optimizes advertising campaigns within the Yandex network. The NEBIUS.AI Cloud platform offers a specialized infrastructure for AI workloads, supported by GPU clusters and partnerships such as Nvidia, and is particularly aimed at companies in the areas of technology, media and life sciences. Toloka AI, a member of the Nebius Group, offers solutions for data annotation, training and evaluation of AI models and employs over 200,000 annotators in more than 40 languages. This enables the high quality creation of training data sets, supported by strategic investments by Bezos Expeditions and Mikhail Parakhin.
🎯📊 Integration of an independent and cross-data source-wide AI platform 🤖🌐 for all company matters
Integration of an independent and cross-data source-wide AI platform for all company matters-Image: Xpert.digital
Ki-Gamechanger: The most flexible AI platform-tailor-made solutions that reduce costs, improve their decisions and increase efficiency
Independent AI platform: Integrates all relevant company data sources
- This AI platform interacts with all specific data sources
- From SAP, Microsoft, Jira, Confluence, Salesforce, Zoom, Dropbox and many other data management systems
- Fast AI integration: tailor-made AI solutions for companies in hours or days instead of months
- Flexible infrastructure: cloud-based or hosting in your own data center (Germany, Europe, free choice of location)
- Highest data security: Use in law firms is the safe evidence
- Use across a wide variety of company data sources
- Choice of your own or various AI models (DE, EU, USA, CN)
Challenges that our AI platform solves
- A lack of accuracy of conventional AI solutions
- Data protection and secure management of sensitive data
- High costs and complexity of individual AI development
- Lack of qualified AI
- Integration of AI into existing IT systems
More about it here:
Delivery robots and warehouse automation: How autonomous systems shape the future
Innovations in robotics and autonomous systems
Yandex and the specialized units emerged from the company have made significant progress in robotics and autonomous systems. These range from self -driving vehicles and robotaxis to delivery robots and complex warehouse automation solutions. The development in this area reflects a similar strategic division as in the AI sector: a strong technological basis that was laid in Russia, and increasingly international commercialization by spin-offs.
Autonomous mobility: from self -driving cars to robotaxis
Yandex's ambitions in the field of autonomous mobility have been many years and have led to the development of a robust technology platform, which is now being further developed internationally.
1. Yandex Group for Autonomous Driving (SDG): Technological milestones and inserts (primarily historical/Russian context)
The project for self -driving cars from Yandex, which aimed at robotaxis, started its first prototype in May 2017. The technology was intensively tested in Russia (Innopolis, Moscow), Israel and the USA (Nevada, CES in Las Vegas, Ann Arbor, Michigan). By the end of 2021, the fleet included around 170 autonomous vehicles that had covered over 14 million kilometers. An important technological progress was the development of its own lidar systems in November 2021. The technology tap was originally comprised by Velodyne Lidar, Radare, Cameras, GNSS sensors, Intel CPUS and NVIDIA GPUS that ran on a Linux operating system. This phase established Yandex as a serious player in research and development of autonomous vehicles and laid the technological basis for later outputs.
2. Avride: The spin-off, Hyundai partnership and international operations
After Yandex AV had to stop his western operations due to sanctions, Avride was founded as a new company based outside of Russia and acquired the assets of the Yandex SDG. Avride has its headquarters in Austin, Texas, and maintains offices in Tel Aviv, Belgrade and Seoul.
Avride continues the cooperation with the Hyundai Motor Group, which has been Yandex SDG since 2019 (then still Yandex SDG) to develop robotaxis. Hyundai Ioniq 5 vehicles are used, which are assembled in the US state of Georgia and then equipped with Avride's autonomous technology. It is planned to expand the fleet to 100 autonomous Ioniq 5 by 2025. The first vehicles should be used as part of an exclusive contract with Uber in Dallas (the articles were probably 2024 or 2025).
After the technology agreement has expired in early 2025, Yandex and Avride go technologically separate ways. Avride relies on rotating, dome-assembled lidar systems, while Yandex (in Russia) uses four semi-solid-state eyelidars and a cleaning system. Avride is therefore the international successor to Yandex ambitions in the field of autonomous driving. The partnerships with Hyundai and Uber are crucial for scaling and market access in the USA. Technological divergence underlines adaptation to different operational and market -related requirements.
Robotics in logistics and delivery
Yandex recognized the potential of the robotics for logistics and delivery applications early on and developed corresponding solutions that are now also marketed internationally.
1. Yandex Rover, Grubhub partnership and Last Mile solutions (Yandex and Yango Tech Autonomy)
Yandex presented autonomous lovers (Yandex Rover) in 2019, which are based on the technology tack of self -driving cars. These luggage robots navigate on sidewalks at 5-8 km/h and have been used in Russia and the USA for the delivery of food, purchases and packages since 2020.
An important early international initiative was the Yandex SDG partnership with Grubhub (July 2021) for robot-based delivery to US college campuses, including Ohio State University and the University of Arizona (end of 2021). The £ 150, six -wheeled Yandex Rover were used, with the aim of operating over 250 campuses.
Yango Tech Autonomy, part of the Yango Group, currently presents autonomous delivery robots for urban environments. In partnership with the food technology and retail company Roots, these robots are already in use in Dubai (Sobha Hartland, Expo City) for the load-mile delivery of food. The robots provide orders within a radius of 2 km in less than 30 minutes, have a loading volume of 60 liters, can avoid obstacles and overcome curbs. This shows a multi-track approach for the last-mile delivery car, from the first Yandex inserts to current international commercialization by Yango Tech Autonomy in new markets such as Dubai.
2. Warehouse automation: inventory, warehouseist and picking robot (Yandex and Yango Tech Robotics)
Yandex (Russia) developed two prototypes of warehouse robots by August 2022:
- The inventory robot scans QR codes in the logistics complex of Yandex Market in Sofyino. It has a 12 meter high mast, lidar and sensors. It currently serves a storage area of 1,000 m², and 20,000 m² is planned.
- The warehouse operator robot (prototype) is designed for the transport of goods into Darkstores. It is compact, agile and resistant to low temperatures and is to take over 75% of the goods movements in the future.
Yango Tech Robotics (part of the Yango Group) now goes one step further internationally and presents the picking robot, a warehouse automation solution based on imitation learning (behavioral cloning from real warehouse scenarios). This robot achieves a picking accuracy of 95% in 97% of the product types and was presented on Logimat 2025. Yango Tech Robotics also offers gripping pliers (400 items/hour) and solutions for industries such as electronics, retail, e-commerce, 3pl, production and pharmaceutical. A case study describes a landfill robot that increased the productivity of 250 to over 400 boxes per hour.
Yandex initiated the F&E in the area of warehouse automation for its own logistics needs. Yango Tech Robotic is now driving this with more advanced AI (imitation learning) and a wider range of solutions for various industries internationally, which indicates a significant B2B market focus.
Suitable for:
- Autonomous mobile robots (AMR) and artificial intelligence (AI): cost reduction and efficiency boost in intralogistics
Robotics F&E, intellectual property and future development
The innovations in the robotics area are underpinned by targeted research, the protection of intellectual property and the training of specialists.
1. Important patents in the area of robotics
The Yandex Group for autonomous driving LLC considers patents for key technologies such as the calibration of lidar systems (US patent 11754690), the control of the track posture of vehicles (US patent 11753037), the object recognition using lidar (US patent 11740358, 11574483), the training of ML algorithms for object recognition on a distance (US patent 11676393) and the design of a transfer robot (US design patent D1023097).
Yandex has a total of 195 patents worldwide (186 unique patent families, including 27 active), with most of the registrations in Russia. The focus is on digitization, ML, AI, autonomous driving and speech recognition. These patents demonstrate specific technological innovations in core areas of autonomous navigation and robotic perception and underpin the company's robotics products.
2. Educational initiatives for robotics talents (e.g. Skoltech)
Similar to the AI area, Yandex is also investing in the training of robotics specialists in Russia. In partnership with the Skolkowo Institute for Science and Technology (Skoltech), a master's degree in “AI Robotik” was launched. This trains robotics engineers, with a focus on autonomous systems and the integration of KI/ml in hardware. Students work real challenges from Yandex Robotics, complete internships and use the infrastructure of Yandex. This ensures a pipeline of qualified engineers for domestic needs and potentially also for international spin-offs.
The robotics strategy of Yandex and its outlets reflects the procedure in the AI area: a strong F&-base and initial implementations in Russia (SDG, warehouse robot for Yandex Market) form the basis for international commercialization and further development through specialized spin-offs (Avride for robotaxis, Yango Tech for delivery and warehouse robot). This pattern indicates a strategic plan: to incubate and test advanced technologies within the Russian Yandex ecosystem in order to then outdo them into internationally oriented units. These can open up global markets, capital and talents and at the same time reduce the risks associated with the Russian parent company.
While the Yandex SDG was involved in the modification of vehicles such as the Toyota Prius and Hyundai Sonata, Avride's partnership with Hyundai for the Ioniq 5, where Hyundai provides the EV platform and Avride integrates its autonomous technology, indicates a strategic change. Avride is increasingly positioning itself as a provider of autonomous driving technology and not as a complete vehicle manufacturer. Early works by the Yandex SDG included the modification of existing car models. Avrid's main partnership with Hyundai provides that Hyundai installs the IONIQ 5 vehicles especially for the integration of Avride technology. Hyundai is even described as a company that drives its “business with autonomous vehicle foundries” to support companies like Avride. This “Foundry” model enables Avride to concentrate on its core competence -autonomous driving software and systems -and at the same time use an established automobile manufacturer for the vehicle platform. This is a more capital-intensive and scalable approach for entering the Robotaxi market.
The emphasis on “imitation learning” and “behavioral cloning” by Yango Tech Robotics for his picking robot indicates a focus on more adaptable and human -like robot skills. Traditional warehouse robots are often dependent on structured environments and preprogrammed tasks. Yango Tech Robotic emphasizes that his picking robot is trained by “behavioral cloning from real warehouse scenarios” and has the ability to “adapt like a person” with the aim of replicated “human -like skill and adaptability - on a large scale”. If it is successful, this approach could lead to robots that are easier to use in diverse and less structured environments, require less explicit programming for new tasks and can manage a greater variety of articles and situations. This addresses essential restrictions on the current warehouse automation and could offer a competitive advantage over robots that are based exclusively on preprogrammed rules, especially in complex, dynamic storage environments.
The following table summarizes the most important robotics solutions:
Yandex 'most important robotic solutions and initiatives (Yandex llc & connected companies)
Yandex 'most important robotic solutions and initiatives (Yandex llc & connected company) -Image: Xpert.digital
Yandex and its connected companies develop a wide variety of robotic solutions with high strategic importance in different areas of application. In the area of self-driving cars, they work with technologies such as lidar, radar and cameras as well as AI software to achieve level 4 autonomy. They originally used vehicles such as the Toyota Prius and Hyundai Sonata, while the internationally oriented brand Avride uses the Hyundai Ioniq 5. Tests took place in Russia, Israel and the United States, with over 14 million kilometers driven so far. There is also a partnership with Hyundai and Uber to provide autonomous vehicles, with a fleet of up to 100 Hyundai Ioniq 5 planned by 2025. These developments strengthen Yandex 'position in the field of autonomous robotaxis.
Delivery robots such as the Yandex Rover and the internationally oriented solutions from Yango Tech Autonomy rely on sidewalk navigation, lidar and cameras. The Yandex Rover has a loading capacity of £ 150 and is used on US college via Grubhub and the Russian Post. Yango Tech focuses on international markets with a 60-liter capacity and delivery times under 30 minutes and works with roots in Dubai. The aim is to optimize the last delivery mile for food and packages.
In the logistics centers, Yandex drives automation with warehouse robots. Your inventory robots scan QR codes up to 12 meters, while compact warehouse operator robots for DarkStores, including refrigeration environments, are designed. These are already used in Yandex Market logistics centers in Russia and improve efficiency and process speed.
Yango Tech also offers solutions for commercial, international B2B storage automation. These include picking robots with an accuracy of 95 %, gripping tongs for up to 400 items per hour and landfillers who can master more than 400 boxes per hour. The technologies were presented on LogiMat 2025, for example, and target broader industry applications.
At the same time, the range of skilled workers in the area of robotics and artificial intelligence is secured in research and development through a partnership with Skoltech. With the master program "Ai Robotik", Yandex laid the foundation for future innovations in autonomous systems, AI-based hardware and frameworks in the robotics area.
Our recommendation: 🌍 Limitless reach 🔗 Networked 🌐 Multilingual 💪 Strong sales: 💡 Authentic with strategy 🚀 Innovation meets 🧠 Intuition
At a time when a company's digital presence determines its success, the challenge is how to make this presence authentic, individual and far-reaching. Xpert.Digital offers an innovative solution that positions itself as an intersection between an industry hub, a blog and a brand ambassador. It combines the advantages of communication and sales channels in a single platform and enables publication in 18 different languages. The cooperation with partner portals and the possibility of publishing articles on Google News and a press distribution list with around 8,000 journalists and readers maximize the reach and visibility of the content. This represents an essential factor in external sales & marketing (SMarketing).
More about it here:
Yandex and his spin-offs: How strategic partnerships drive global AI growth
Ecosystem, strategic alliances and investments
The commitment of Yandex and the connected companies in AI and robotics is significantly supported and promoted by a network of strategic partnerships, targeted investments and the use of the Yandex Cloud platform. These elements are crucial for research, talent development, market access and the scaling of the operations, both in Russia and internationally.
Important partnerships in AI and robotics (national and international)
The collaborations from Yandex and its spin-offs extend to academic institutions to global technology and service companies.
National partnerships (Yandex LLC/Russia):
Yandex maintains close relationships with leading Russian educational institutions. This includes collaboration with the Higher School of Economics (HSE) in the field of AI formation and research, including a pilot project on Yandexgpt. With the Moscow Institute of Physics and Technology (MIPT), a AI was developed to evaluate Olympiad work. Another important cooperation with the Skolkowo Institute for Science and Technology (Skoltech) exists as part of a master's degree in AI robotics. In the area of autonomous logistics, Yandex entered into a partnership with the Russian Post in October 2021, in which 36 robots were used for autonomous deliveries in Moscow.
International partnerships (spin-offs):
- Avride: The collaboration with the Hyundai Motor Group, which started in 2019 with the Yandex SDG, is central to the development of robotaxis based on the Hyundai Ioniq 5. An exclusive contract with Uber provides for the use of autonomous vehicles in Dallas.
- Yandex SDG (before Avride): An early international partnership was the one with Grubhub for the use of the Yandex Rover on US college campuses.
- Yango Tech Autonomy: Cooperates with Roots, a company for food technology and retail trade, for the use of delivery robots in Dubai.
- Yango Tech Retail: has a strategic partnership with Grand Hypermarkets (VAE, Kuwait) for AI-based retail solutions.
- Nebius.ai: Has an important partnership with Nvidia, is a launch partner for the Blackwell Ultra Ki-Fabrik and supports the DGX Cloud Lepton Marketplace.
- Toloka: Companies like Amazon, Anthropic, Microsoft and Shopify counts. A strategic investment was led by Bezos Expeditions and Mikhail Parakhin (CTO of Shopify).
These partnerships are of crucial importance. The distinction between national (Yandex LLC) and international (spin-offs) cooperations illustrates their different strategic focus and target markets.
Investment strategies and focus of the connected companies (Nebius, Toloka, Yango)
The investment strategies of the international units resulting from Yandex testify to a clear focus on aggressive growth.
- Nebius Group/Nebius.ai: Nebius, which emerged from Yandex's cloud and data center activities, plans to invest $ 1 billion in AI and Cloud infrastructure in order to gain market shares, of which over $ 1 billion in the AI infrastructure in Europe until mid-2025. the USA, Israel and Iceland and is aiming for a capacity of 100 MW until the end of the reporting year (Q1 2025). Before returning to Nasdaq, Nebius received a private investment round of $ 700 million in which Nvidia took part. Investment expenses (Capex) in the first quarter of 2025 amounted to $ 544 million, with a forecast of $ 2 billion for the year as a whole. The liquid funds were $ 1.44 billion. The core business with AI infrastructure generated sales of $ 55.3 million annually in the first quarter of 2025, with an annualized run rate (ARR) of $ 249 million (end q1) or $ 310 million (April). By the end of 2025, an ARR will be sought of $ 750 million to $ 1 billion.
- Toloka: In May 2025, Toloka received a strategic investment of $ 72 million, led by Bezos Expeditions and Mikhail Parakhin to scale quickly. The NEBIUS Group keeps the economic majority stake, but releases the voting right control.
- Yandex Group for Autonomous Driving (before Avride): In September 2020, the Yandex and Uber joint venture was outsourced, with Yandex investing $ 150 million.
- Yango Tech (part of the Yango Group): The Yango Group is a global technology company, and Yango Tech is her B2B Ki ecosystem. Specific investment figures for Yango Tech are not detailed in this information, but its activities (robotics, retail ACI, autonomy) indicate significant ongoing investments.
These investment patterns show that the international units are strongly geared towards growth, with Nebius massively investing in capital-intensive AI cloud infrastructure and Toloka attracts strategic external capital for the scaling of its specialized AI date services.
The role of Yandex Cloud in the AI ecosystem (primarily Russia)
Yandex Cloud is the primary vehicle for Yandex LLC to offer AI services like Yandexgpt the Russian B2B market. In 2024, the turnover of Yandex Cloud (part of the Yandex B2B Tech Group) was 19.80 billion rubles (+50% annually). The market share of Yandex B2B Tech was around 13% of a market of 240 billion rubles in 2024, and it is expected to grow faster than the market.
The number of customers from Yandex Cloud exceeded 44,000 (+50% in 2024), with corporate customers made up of 51% of consumption. The focus is on external users because the internal consumption of the Yandex Group was below 5%. Yandex Cloud invests heavily in the platform infrastructure (scalability, security, performance), with investments in this area increasing 200% in 2024. The consumption of security services increased by 110%. The Yandexgpt 5 Pro solution for the processing of documents for clinical studies was developed in collaboration with Yandex Cloud. The strong growth and customer acquisition of Yandex Cloud indicate a successful monetization of its cloud and AI skills within Russia.
The development of a global-local ecosystem is emerging. While Yandex LLC focuses on Russia and target its spin-offs international markets, there is potential for a symbiotic relationship. The spin-offs benefit from the basic F&E and the talent pool from Yandex. Yandex LLC, in turn, could potentially benefit from international advances and best practices of his former units, provided that relationships and IP agreements allow this. The common “Yandex-DNA” could promote informal exchange of knowledge. Yandex LLC has strong F&- and academic connections in Russia. Spin-offs such as Avride, Nebius, Toloka and Yango build on the technology and often also the staff of Yandex. These now act in competitive international markets and drive innovations (e.g. Avride with Hyundai/Uber, Nebius with Nvidia, Toloka with global AI laboratories), while Yandex LLC continues to develop AI/Robotics for the Russian market (Yandex Cloud, domestic robot). Although legally separated, the common technological heritage and personnel fluctuation could create an informal ecosystem in which innovations and knowledge of the international units (including indirectly) could influence the Russian operations of Yandex or benefit them, especially in basic research areas. The success of the international arms could also validate the nuclear technologies developed by Yandex.
A shift in the profile of strategic investors can be observed: from internal investments to specialized technology and growth capital. The original major investments by Yandex LLC were internal in the company (e.g. its own $ 150 million in the SDG). The international spin-offs are now attracting specialized technology investors (Nvidia with Nebius) and top-class growth capital (Bezos expeditions at Toloka). This change in the investor profile signals that the outsourced units are assessed and financed on the global AI/robotics market based on their specific advantages and growth prospects, instead of as part of a diversified Russian technology group. It also brings strategic partners into play that can accelerate their growth.
Geopolitical constraints have worked as a catalyst for the formation of this international ecosystem. The sanctions and the geopolitical climate were direct triggers for the formal separation and internationalization of the most important KI and robotics assets from Yandex. Without this pressure, Yandex might continue to have operated this as internal departments or subsidiaries with a more centralized control, based in Russia. Yandex developed and implemented Before 2022 International KI/Robotics (e.g. SDG tests in USA/Israel, Grubhub partnership). Sanctions impaired the western operations of Yandex AV and led to the suspension of the Nasdaq nodes of Yandex NV. As a result, Yandex's Russian assets were sold and the emergence of the Nebius Group (NL), which holds international business. Avride was explicitly founded because Yandex AV hired the company in the west due to sanctions. The current structure with independent international units (Nebius, Avride, Yango, Toloka) is therefore a direct consequence of geopolitical events that forced a strategic reorganization in order to maintain and expand these valuable technology assets on the global stage.
Yandex: Technological realignment and global AI ambitions
The activities of Yandex and the international units in the field of AI and robotics resulting from the company are characterized by dynamic development, strategic realignment and considerable market potential. The future development will depend on the ability to successfully act both on the Russian home market and in global competition and to master the respective challenges and opportunities.
Future roadmap and expected developments in the KI- and robotics ventures from Yandex (Yandex LLC and spin-offs)
The road maps of the various entities indicate ambitious growth plans and continuous technological development.
Yandex LLC/Russia
Russian unity is expected to continue the development of Yandexgpt, Yandexart and other generative AI models for the Russian market. Another focus is on the expansion of the Yandex Cloud services and the increase in the market share in Russia. The promotion of AI talents by university partnerships (HSE, Skoltech) remains an important pillar. The activities will be based on the goals of the national AI strategy of Russia by 2030, with increased cooperation with China appearing possible in the AI area. Yandex Research in Russia will continue to concentrate on fundamental problems in ML, NLP and computer vision and publish at top conferences.
International spin-offs
- NEBIUS.AI: Plans a rapid scaling of his Ki cloud infrastructure with the aim of reaching an ARR from USD 750 million to 1 billion by the end of 2025 and over 1 billion USD by 2026. The expansion of data centers in the USA, Europe and the Middle East as well as the introduction of the latest Nvidia Blackwell hardware are planned. The profitability is sought for the second half of 2025.
- Avride: intends to enlarge its robotaxi fleet (up to 100 Ioniq 5 in 2025) and to expand the over-the-end partnership beyond Dallas. Another technological divergence of Yandex in the field of autonomous driving is expected.
- Toloka: With the new investments, it will quickly expand its business with AI data solutions, whereby the focus on agent safety, red teaming and high-quality data for LLMS is. The expansion of the customer base from leading AI laboratories and companies is another goal.
- Yango Tech (Robotics & Autonomy): Plans the expansion of his B2B robotic solutions (warehouse, delivery) into new international markets and the further development of AI-based skills such as imitation learning.
These plans indicate aggressive growth goals for the international units and the persistent strategic importance of AI and robotics for Yandex in Russia.
Competition positioning in global and Russian technology landscape
The competition positioning varies greatly between Yandex LLC in Russia and the internationally operating spin-offs.
Yandex LLC/Russia
In the Russian market for AI training records, Yandex is a key player alongside Sber Ai and MTS AI. Yandexgpt and Gigachat from Sber are leading Russian LLMs, although they manage behind the top models in a global comparison. A strength is the excellent understanding of the Russian language.
International spin-offs
- Nebius.ai: aims to become an important player in the field of AI infrastructure in Europe and worldwide and compete with other AI cloud providers. The partnership with Nvidia is an important lever.
- AVRIDE: In the Robotaxi area, appears against established actors such as Waymo and Cruise and uses the partnerships with Hyundai and the Uber platform.
- Toloka: positions itself among the world's leading AI data companies and competes with other providers of data advertising and AI training data.
- Yango Tech Robotics: In competition with numerous providers of warehouse and logistics automation.
Yandex is dominant in Russia for localized AI. The spin-offs face intensive global competition, but have strong technological foundations and strategic partnerships.
Identified challenges, opportunities and strategic implications
The future prospects are associated with specific challenges and significant opportunities.
challenges
- Yandex LLC (Russia): Restricts of resources and limited computing power for AI scensorship/development compared to global giants. Possible effects of geopolitical isolation on access to global talents and technologies. The maintenance of the global competitiveness of Russian LLMs. Data protection concerns and regulatory compliance.
- International spin-offs: intensive competition in the areas of Ki-Cloud (Nebius), robotaxis (Avride), AI data (Toloka) and robotics (Yango). Execution risks for quick scaling. Management of the technological divergence of the original Yandex technology tap with at the same time maintaining innovative strength. Navigation through complex international regulations.
opportunities
- Yandex LLC (Russia): Strong demand for localized AI solutions in Russia. Leadership role in Russian -speaking AI. State support from the national AI strategy. Growing Russian AI market.
- International spin-offs: enormous global growth of the AI market. Strong technological heritage from Yandex. Ability to attract international capital and talents. Strategic partnerships with important technology files (Nvidia, Hyundai, Uber, Bezos Expeditions). Focus on specialized niches.
The overarching strategic challenge lies in coping with the perception and reality of two interconnected, but increasingly different “Yandex” ecosystems. Yandex LLC in Russia operates as part of national strategic goals and local market needs. The international spin-offs have to assert themselves on a global stage, driven by other market dynamics and investor expectations. These spin-offs develop technologically and surgically from Russian unity (e.g. Avrides Hardware). The “Yandex” brand is still strongly connoted by Russian, which could be a challenge for the international units despite the legal separation. This creates a fundamental tension: How can the international units use their Yandex heritage for credibility and at the same time establish independent global identities? How does Yandex LLC keep his pace of innovation with potentially greater insulation? The long -term success depends on effectively manage this duality.
The Yandex saga (original company and its spin-offs) serves as a significant real test case for how strategically important technology assets can be restructured and internationalized in response to geopolitical shifts and sanctions. Yandex was a globally recognized technology company with significant international operations and stock markets. Geopolitical events led to serious surgical restrictions and the need for radical restructuring. The solution was to separate Russian and international assets and the establishment of new legal persons for the latter. These new units are now trying to be independent on the global market. The career of Nebius, Avride, Toloka and Yango will be observed closely as an indicator of whether and how advanced technology ecosystems can adapt, survive and even flourish after such a profound geopolitical and entrepreneur. It tests the resistance of technology and talent regardless of national origin in the face of global market forces. Your success or failure will offer teachings for other multinational technology companies that are exposed to similar pressure.
The development of units such as Nebius (Ki-Cloud), Toloka (AI data) and Avride (robotaxis) as independent, well-financed companies reflects a broader trend towards specialized “Pure Play” AI companies. AI development is complex and requires a deep specialization (e.g. LLM infrastructure, data advertising, autonomous driving). NEBIUS focuses exclusively on AI cloud infrastructure, toloka on AI data solutions and avride on autonomous vehicle technology. These units attract significant, targeted investments (Nvidia at Nebius, Bezos near Toloka, Hyundai near Avride). Investors support increasingly focused companies instead of conglomerate, especially in capital-intensive and rapidly developing AI sectors. This specialization enables deeper expertise, faster innovation and clearer promises for values for customers and investors-a trend that is expected to continue in the AI industry if different niche tires.
Based on the analysis, the following strategic recommendations can be derived:
- Yandex LLC (Russia): Should the Russian language and localization further strengthen as a central distinguishing feature. The promotion of local AI talents must be continued. A deeper collaboration with Brics states and other non-western partners in the AI area should be explored.
- International spin-offs: The focus must be on the implementation and scaling in the respective global markets. The structure of strong, independent brand identities is crucial. Partnerships should be used to speed up the market penetration. The technological lead must be kept by focused F&E. Synergies should be explored without endangering independence.
In summary, it can be said that Yandex and his affiliated companies drive an impressive range of AI and robotics initiatives. The strategic realignment through the creation of international units is a courageous answer to complex geopolitical and market -related challenges. The future success will depend on how effectively this decentralized structure promotes innovations, gains market shares and masters the inherent complexities of the global and Russian technology market.
We are there for you - advice - planning - implementation - project management
☑️ SME support in strategy, consulting, planning and implementation
☑️ Creation or realignment of the AI strategy
☑️ Pioneer Business Development
I would be happy to serve as your personal advisor.
You can contact me by filling out the contact form below or simply call me on +49 89 89 674 804 (Munich) .
I'm looking forward to our joint project.
Xpert.Digital - Konrad Wolfenstein
Xpert.Digital is a hub for industry with a focus on digitalization, mechanical engineering, logistics/intralogistics and photovoltaics.
With our 360° business development solution, we support well-known companies from new business to after sales.
Market intelligence, smarketing, marketing automation, content development, PR, mail campaigns, personalized social media and lead nurturing are part of our digital tools.
You can find out more at: www.xpert.digital - www.xpert.solar - www.xpert.plus