कृत्रिम launches India’s first frontier research AI Lab to democratise AI innovation; commits investment of ₹10,000 crore by next year

Announces investment of ₹2,000 crore (equity and debt) today
Releases कृत्रिम-2 12B,- best-in-class large language model for Indic.
Open Sources 6 new state-of-the-art Indic AI models across text, voice and speech to enhance AI adoption in India
Develops “BharatBench” as the global benchmark for evaluation performance of AI models on Indic tasks
Announces deployment of India’s first and most advanced GB 200 supercluster in partnership with NVIDIA

Bengaluru, February 4, 2025: कृत्रिम, India’s first AI unicorn, today announced the launch of कृत्रिम AI Lab, the country’s first AI frontier research lab. This pioneering initiative is set to make AI research accessible, attract the best talent, and establish India as a global leader in open-source AI. As AI transforms industries worldwide, कृत्रिम AI Lab takes a step towards creating state-of-the-art AI infrastructure, supporting the research and developer community, and building cutting-edge AI models tailored for Indic languages and real-world applications.

In line with building its growth roadmap, कृत्रिम announced an investment of ₹2,000 Cr today (equity and debt) with a commitment of ₹10,000 Cr by next year. कृत्रिम also strengthens its global collaborations, by announcing India’s first GB200 cluster deployment along with NVIDIA scheduled to go live by March 2025. कृत्रिम aims to build India’s largest supercomputer in India by the end of the year.

Bhavish Aggarwal, Founder, कृत्रिम said, “With कृत्रिम, our vision is to develop AI for India and make it better for Indian languages, data scarcity and cultural context. While we are behind global benchmarks, we have made good progress in the last one year. With कृत्रिम AI Lab, we hope that the entire Indian AI community collaborates to create a world-class Indian AI ecosystem.”

कृत्रिम AI Lab champions open source AI and democratise AI innovation by fostering collaboration with academia, startups, and developers. It deploys state-of-the-art computing resources to accelerate AI capabilities. With its multilingual AI models that represent India’s diverse linguistic landscape, कृत्रिम AI Lab ensures equitable access to AI across all Indian languages. By creating AI solutions optimised for scale and affordability, the research lab enables cost-efficient AI innovation and attracts top talent that pushes the boundaries of AI knowledge.

कृत्रिम has open sourced new state-of-the art Indic AI models to enhance AI adoption in India:

कृत्रिम-2 and कृत्रिम-1 LLMs: The कृत्रिम 1 7B model (India’s first LLM) was launched on January 2024. कृत्रिम 2, launched today, is an improved model. More information here: https://tech.olakrutrim.com/krutrim-2-a-best-in-class-large-language-model-for-indic-languages/
Chitrarth 1: India’s first Vision Language Model built on top of कृत्रिम-1, is capable of understanding images and documents. Additional information here: https://ai-labs.olakrutrim.com/models/Chitrarth-1
Dhwani 1: India’s first Speech Language Model, built on top of कृत्रिम-1, is capable of undertaking tasks like speech translations. More information here: https://ai-labs.olakrutrim.com/models/Dhwani-1
Vyakhyarth 1: State-of-the-art (SoTA) Indic Embedding model for use cases like Search and RAG. More information here: https://ai-labs.olakrutrim.com/models/Vyakyarth-1-Indic-Embedding
कृत्रिम Translate 1: State-of-the-art (SoTA) Text-to-text translation model. More information here: https://ai-labs.olakrutrim.com/models/Krutrim-Translate
Additionally, कृत्रिम develops ‘BharatBench’ as the global benchmark for Indic performance. More information in the technical report is here: https://tech.olakrutrim.com/bharat-bench/

कृत्रिम AI Lab focuses on critical AI research domains essential for digital transformation and development of India’s first large-scale AI models, capable of competing globally. The multimodal Lab is aimed at building AI that understands and processes text, speech, and visuals across multiple languages. It also addresses India’s data scarcity challenge by digitising knowledge artefacts and creating high-quality datasets. Additionally, by incorporating AI for real-world scale, कृत्रिम AI Lab ensures AI systems that are efficient and adaptable to India’s high-volume, resource-constrained environments.

To supercharge innovation in India, कृत्रिम previously announced over 50 new services on कृत्रिम cloud making it ready for most applications needed by Indian developers. The infrastructure services include Virtual Machines (VMs), cloud storage, state-of-the-art security measures for data protection and cutting-edge observability features for precise data monitoring. Additionally, the AI cloud services announced were AI Pods to enable low-cost access to best-in-class GPUs, AI studio for building compound AI applications, model catalogue with latest LLMs and vision models, and now code/low code training, finetuning, inference and model evaluation capabilities.

To support the company’s ambitious AI and cloud initiatives, कृत्रिम also announced its plans to scale up its data centre capacity to a massive 1 GW by 2028. The data centre will be a computing and data storage powerhouse, enabling कृत्रिम to offer world-class AI services to India and the world.