In brief
• Basis AIR allocates graphics processing units (GPUs) across teams so that hardware does not sit idle.
• Basis BOSS gathers IT performance data in one place and uses AI to help pinpoint the root cause of outages.
• Basis Repository verifies third-party software components before they reach the application build.
• Internally, AI assists engineers with documentation and troubleshooting, and helps investors review financial reporting. Final decisions remain with humans.
Basis AIR: keeping GPUs from idling
Graphics accelerators are the most expensive component of AI infrastructure. Typically, they are dedicated to specific teams: one department's cards run around the clock with queued tasks, while another's remain idle half the time. Basis faced this problem internally and developed an in-house tool that subsequently evolved into a commercial product.
Basis AIR aggregates accelerators into a shared pool and allocates resources according to task priorities and departmental quotas. The company estimates this increases usable GPU utilization from 30% to 80%—allowing more projects to run on the same hardware.
The product functions autonomously: other Basis solutions are not mandatory, as infrastructure on compatible Kubernetes platforms is sufficient. The company describes it as the first Russian standalone platform in this class.
Demand for such tools is growing along with the rise in enterprise AI projects. iKS-Consulting estimates the Russian market for AI and GPU infrastructure management software at 430 million rubles in 2026, projecting an average annual growth rate of 36% through 2031.
Basis BOSS: understanding the root cause of failures
The second direction is observability. This marks the next evolution beyond standard monitoring: rather than merely indicating an overloaded server, the system correlates infrastructure, application, and user service data to reveal exactly where an issue originated and what it affected.
Basis BOSS includes an integrated AI analyst: users can query system health and incidents in conversational natural language via chat. The underlying technology is based on a platform by Russian developer Proto, in which Basis acquired a 70% stake in September 2026.
Basis Repository: a filter for third-party code
Modern software relies heavily on third-party libraries, including open-source packages. Basis Repository serves as a centralized gateway through which such components reach developers. Every component is inspected according to security policies configured by the client. If a vulnerability is subsequently discovered in an already approved version, its distribution can be restricted.
Internal AI: an assistant, not a replacement
Within the company, engineers are already adopting AI across their workflows, such as preparing technical documentation. The core principle remains consistent: the model suggests, the specialist reviews and decides.
• “Digital Tech Lead” is an AI assistant for maintaining internal infrastructure. It collects and correlates logs, metrics, events, and documentation, formulating hypotheses regarding the causes of malfunctions. An engineer validates them.
• Basis Investor Data is a service designed for investors. It integrates with compatible AI assistants via the MCP protocol to locate metrics, compare periods, and reference source documents. The company notes this is the first such service provided by a Russian issuer.
To what extent can AI accelerate infrastructure software development overall? iKS-Consulting provides these benchmarks: a 10–20% boost in R&D productivity, reaching up to 20–30% in teams with mature testing and quality control processes. Analysts attribute the primary gains not to writing raw code, but to accelerating the entire lifecycle: documentation, code review, and refinement.
Financial results
During the first nine months of 2026, Basis’s revenue rose by 29% year-on-year to 4.46 billion rubles. OIBDA—operating income before depreciation and amortization—grew by 32% to reach 2.46 billion rubles, representing an OIBDA margin of 55%. Profit grew faster than revenue, as expenses remained at the previous year's level.
“We see that the adoption of AI has already helped increase labor productivity and bring back in-house tasks that were previously outsourced to third-party contractors. Largely thanks to AI, we are able to keep expenses flat compared to last year while increasing profits,” — David Martirosov, CEO of Basis
Key points to keep in mind
Increasing GPU utilization “from 30% to 80%” represents the company’s internal estimate rather than an independent benchmark test.
Market forecasts and productivity impact assessments are provided by iKS-Consulting. They serve as industry benchmarks rather than guaranteed outcomes for individual teams.
About the company
Basis is a Russian software developer specializing in dynamic IT infrastructure management, virtual desktops, and cloud services. Established in 2021 through the combination of assets from Tionix, Digital Energy, and Skala Software, it is controlled by RTK-DPC Group. Since December 2025, its shares have traded on the Moscow Exchange under the ticker BAZA. Its products are listed in the unified register of Russian software and certified by FSTEC of Russia.
Learn more about the products on the company website: basis.ru.
erid: 2VtzqumieA2
Advertising. OOO “Basis” INN 7731316059