AethexAI raised $3 million in pre-seed funding to build voice AI systems for enterprises in Africa and the Middle East, where customer support still depends heavily on phone calls.
The round was led by 4DX Ventures, with participation from Enza Capital, Dorm Room Fund, Mojo Ventures and Stanford GSB 26 Fund. Individual investors included Stanford faculty, telecom executives and AI researchers from Anthropic.
Founded by Mariama Diallo and Ayooluwa Odemuyiwa, AethexAI is building voice AI for customer service, debt collection, customer activation and KYC checks. The startup is also launching APIs and SDKs for developers and a platform for enterprises to test its tools.
The company built its own small models and orchestration layer instead of using existing voice AI tools. It said the decision was driven by latency, cost and language needs in Africa and the Middle East, where systems must handle local English, French and Arabic dialects, code-switching and informal speech patterns.
Follow us on WhatsApp | LinkedIn for the latest headlines
AethexAI trained its Kora model series using anonymized call-center recordings, audio from radio stations and data annotated by university students. The models range from 300 million to 1.7 billion parameters. The company said it now handles more than 17,000 calls a day and is building local partnerships with telecom providers and forward-deployed engineers to support enterprise clients.
Key Takeaways
AethexAI’s funding shows that voice AI in emerging markets is not just a copy of what works in the US or Europe. In Africa and the Middle East, voice remains a main channel for customer service, banking, telecoms and support, but the infrastructure is different. Calls may run on local telephony systems, networks can have latency issues and customers often mix languages or speak in local accents. Large models hosted far away may be too slow or too expensive for real-time calls. AethexAI is betting that smaller models trained on regional data can work better for these markets. That gives the company a different path from global voice AI firms such as ElevenLabs, Deepgram, Sierra and Cognigy. Its challenge will be quality and trust. Enterprises will need systems that respond quickly, understand customers, avoid errors and meet data protection rules. If AethexAI can prove that its models reduce costs without hurting customer experience, it could become part of the voice infrastructure layer for banks, telcos, insurers and fintechs across the region.