dtelecom Launches Voice AI Infra
A new developer platform called dtelecom has launched its real-time infrastructure for voice AI. The stack, which handles edge compute, STT/TTS, and payments for human-to-agent communication, is now live with an API and SDK after processing over 35 million minutes of audio.
The founding team, CEO Petr Malyukov and CTO Vadim Filimonov, have a history in AI-enabled communication platforms. Before dtelecom, they built YOUS, a communication app with a built-in AI translator that was awarded #1 Product of the Day on Product Hunt. Malyukov has been building companies across telecom, Web3, and AI for over 17 years. dtelecom's infrastructure is built as a Decentralized Real-Time Communication (DRTC) network on the Solana blockchain. This DePIN (Decentralized Physical Infrastructure Network) architecture is designed to be censorship-resistant and allows developers and users to own their data, a departure from traditional centralized providers. A key focus is building infrastructure for autonomous AI agents, not just human users. dtelecom built the first Speech-to-Text (STT) service powered by the x402 protocol, which allows an AI agent with a USDC wallet to programmatically discover, pay for, and use the API without human setup or stored credentials. To attract its initial user base of developers, the company launched a "Startup Grants" program. The program offers selected startups six months of free access to its production-grade infrastructure, including unlimited voice and video minutes and direct engineering support, removing early-stage cost barriers. Founded in 2022 and based in Dubai, dtelecom has been supported by grants from Google, the Solana Foundation, and ElevenLabs. The company recently opened its seed funding round, with one report noting a $1.5 million raise on February 25, 2026. The platform enters a rapidly growing market for voice AI infrastructure, which is projected to expand from $5.4 billion in 2024 to over $133 billion by 2034. This growth is fueled by enterprise demand for automation and the increasing prevalence of voice-first products.
