Theta Labs Teams Up with Top Singaporean University NTU to Supercharge Research Initiatives

Theta Labs, a global leader in decentralized cloud infrastructure joins forces with Nanyang Technological University, Singapore (NTU Singapore), one of the most prestigious universities in Asia.

Theta Labs, a leading provider of decentralized cloud infrastructure for AI, Web3, media and entertainment, has announced a collaboration with Nanyang Technological University (NTU). As part of this partnership, NTU will become a customer of Theta Labs in the academia segment, joining a group of universities that are utilizing Theta EdgeCloud to advance their research initiatives.

Theta Labs joins forces with top Singaporean university NTU to fuel research

Theta Labs, a global leader in decentralized cloud infrastructure, has announced a collaboration with Nanyang Technological University, Singapore (NTU Singapore), one of the most prestigious universities in Asia.

As the collaboration kicks off, NTU becomes the latest Theta Labs customer in the academia segment. Ranked among the top two universities in Singapore, alongside the National University of Singapore (NUS), and 15th globally in the QS World University Rankings 2025, NTU is in the same league with MIT, Stanford and the University of Pennsylvania.

The Speech and Language Laboratory, developed in collaboration with Professor Li Haizhou (National University of Singapore) and Professor Chng Eng Siong from NTU, joins Theta Labs’ premium ecosystem of academic partners, including Seoul National University, KAIST, University of Oregon, Peking University and many others.

The laboratory is focused on developing advanced speech recognition models for languages such as English, Mandarin and Singlish, and innovating solutions for noise reduction, speaker extraction and overlapping speech recognition.

Also, its team is developing large language models (LLMs) for generative error correction, domain-specific applications and contextual understanding. NTU’s SLL is pioneering research in diarization, deep fake detection and audio event classification under complex conditions.

Professor Chng Eng Siong welcomes the contribution of Theta Labs as a reliable tech partner:

"We’ve seen Theta Labs' success and collaboration with many of the top South Korean and US universities, and we’re excited to use EdgeCloud to scale our research at NTU’s Speech Lab. The platform’s easy-to-use distributed GPU infrastructure allows us to push our research into speech, language processing, multilingual ASR and real-time AI applications. This was a key driver for us to adopt Theta EdgeCloud."

The laboratory has yielded over $13.3 million in research funding from a clutch of backers and endowments.

Theta EdgeCloud fuels new research initiatives

Mitch Liu, cofounder and CEO of Theta Labs, stresses the importance of the collaboration for the future of AI breakthroughs:

"In a short few months after the launch of the EdgeCloud platform, we now have over 20 customers across academia, e-learning enterprises, professional sports and esports industries. We’re thrilled to welcome NTU and Professor Chng to the Theta family, quickly driving towards 50+ worldwide customers, with accelerating adoption of EdgeCloud, THETA and TFUEL."

NTU’s Speech Lab will leverage Theta EdgeCloud’s decentralized GPU infrastructure to drive advancements in speech and language AI. The platform will enable NTU to train and fine-tune multilingual ASR models at scale, develop robust real-time transcription systems and drive innovation in noise reduction and speech identification.

It underscores Theta EdgeCloud as the hybrid cloud-edge GPU platform of choice for advanced AI research in speech and language processing, and in general AI.

Building on the success of Theta Labs' 2024 initiatives, this partnership aligns with the key milestones outlined in the 2025 Roadmap. These milestones include the beta release of Theta EdgeCloud’s fully hybrid cloud-edge computing platform in June 2025 and the ongoing expansion of its decentralized edge network, now encompassing over 30,000 nodes globally.