The volume of AI-related traffic on Russian brokerage websites saw a significant shift during the summer of 2026. According to Servicepipe, the share of scraper bots, which collect data for AI training, surged from 7% in June to 68.82% in August. The total number of AI-related requests grew from 6 million to 8 million over the same period.
Meta and Anthropic emerged as the primary drivers of this activity. Meta accounted for 55.14% of scraper requests in August, totaling over 3 million, compared to 7.32% in June. Anthropic’s share reached 27.78% with more than 1.5 million requests, up from 12.26% in June. Other bots, such as OpenAI’s GPTBot, accounted for 8.83% of the traffic.
Data collection efforts by these companies focus on extracting text, tables, and financial analytics to improve AI models' understanding of market terminology and logic. While other bot categories like crawlers and assistants saw their share of traffic decline, the aggressive scraping by major US tech firms continues to scale as they seek to incorporate global financial context into their systems.
Industry experts note that brokerage sites are particularly attractive due to their structured and frequently updated financial data. While Russian developers also engage in data scraping, US-based firms operate on a significantly larger scale, targeting global data sets to enhance the capabilities of their neural networks, including the Llama and Claude models.
