Voice AI infrastructure company Deepgram has secured a strategic investment from EDBI, extending its $130 million Series C financing round announced on August 18, 2026. The fresh capital will fund the launch of an Asia-Pacific headquarters in Singapore and expand the firm's regional sales, support, and low-latency infrastructure. EDBI, the venture arm of SG Growth Capital, is positioned as a strategic anchor for Deepgram's regional expansion in a market where demand for voice AI is accelerating.
According to the report, Deepgram's Asia-Pacific customer base generated a 96% year-on-year increase in API requests. The company now operates across more than 20 markets in the region, supporting the four official languages of Singapore—English, Mandarin, Malay, and Tamil—while expanding coverage to Japanese, Hindi, Thai, Cantonese, and additional regional dialects. Globally, Deepgram's platform powers over 200,000 developers and 1,400 enterprises, has processed more than 50,000 years of audio, and has transcribed over a trillion words. The firm has appointed former Salesforce executive Sriram Ved as Vice President and General Manager for the Asia-Pacific region.
The report finds that Deepgram's platform functions as AI infrastructure rather than a consumer-facing application, emphasizing its scale and the critical nature of low-latency inference for contact-center and real-time applications. By locating engineering and commercial teams in Singapore, the company can deliver tighter latency, localized language expertise, and compliance options that appeal to regulated sectors including banking, insurance, and healthcare across Southeast Asia. EDBI's participation signals broader investor appetite for AI infrastructure that supports enterprise workloads, not just consumer-oriented generative models.
The analysis in the report notes that voice AI is shifting from pilot projects to core enterprise workloads in Asia-Pacific, with analysts estimating the regional market will surpass $10 billion by 2028, driven by contact-center automation, digital banking, and e-commerce. For Deepgram, the Singapore headquarters should enable tighter latency service-level agreements and localized model training, which are key drivers for improving net revenue retention and expanding revenue in multilingual markets. The deal underscores growing appetite for AI-infrastructure investments that complement generative AI front-ends, suggesting future Series C and D rounds for similar companies may command higher annual recurring revenue multiples as enterprise adoption accelerates. The partnership also validates Singapore's positioning as an AI deployment hub, potentially attracting more venture capital to AI-infrastructure startups targeting regulated verticals where data residency and compliance are non-negotiable. The move pressures rivals—especially the large cloud speech services—to strengthen local latency and compliance offerings, which could narrow Deepgram's differentiation on speed and data residency. Regional infrastructure investments often force incumbents to rethink distribution strategy, while differentiation built on proximity alone can erode faster than differentiation built on model architecture.