Krutrim, an AI company which is based in Bengaluru and is a leading AI company recently launched a new approach of aid the developers and startups. Their solutions include a model of GPU as a service which tackles the issue of insufficient accessibility and affordability of processing units around the globe.
This scheme from Krutrim Cloud gets further chance for organizations and developers by bringing in to light the access to GPU resources within its brain computing system. Mr. Bhavish Aggarwal, co-founder of the startup Keurtrim made an announcement today, stressing the importance of this breakthrough for training AI models and optimizing them
According to Aggarwal, being the most economical cloud offering globally is the goal that they have set. Ravi Jain, EWA’s strategy leader, responded to that by disclosing that pricing was still under consideration. Moreover, the announcement shows that the first 1,000 developers would get special passes equipped with GPUs, free of charge and worth up to Rs. 10 000. Still, the company manager failed to reveal a given number of the deployed GPUs just for this service.
On the contrary, Jain pointed out that a series of cloud services will be launched “within a couple of weeks”. The future developer projects such as H100, H200, and Blackwell will be their most advantageous one. Jain added that discussions were ongoing with all three major processor manufacturers: Nvidia, AMD (Advance Micro Devices), and Intel.
The Jain mentioned that the company has its data center in Bengaluru and their plan to develop the infrastructure extends to additional centers in many locations. The choice of the place to come up with these centers will be determined by competent factors like the availability of resources and number of consumers.
Yotta Data Services has been adopting a new scheme for the GPU usage by charging $1.8 per hour for long term contracts and $5 per GPU hour for a short-term usage. The recent announcement by the company states the importation of 24,000 H100 GPUs and L40S in a spread-out manner. This first group of GPUs was delivered in March. Along with that, government revealed that it will be importing 10,000 GPUs and make them accessible to startups and researchers in the same month.
Besides offering its cloud platform to business users, developers, and researchers for computing resources, Krutrim will also offer its models including language models and other open-source models that are available on the cloud platform. The company highlights the fact that its models use SDKs, thus the production and improvement of the apps becomes easy.
On the other hand, Krutrim introduced mobile applications as standalone Krutrim AI assistant products to meet the needs of customers directly. These apps in many ways emulate the tools that OpenAI and Microsoft provide. Along with that the company introduced a range of application programming interfaces (APIs) aimed for the developers with the focus on the location-based services: routing, navigation and real-time traffic situations as well as other functions.
Krutrim has already started channel-wide conversations with foundries from Taiwan and the United States that would be responsible for the manufacturing of its own designed chips. The company announced SIP as its AI-focused system at the last month of the previous year, and the prototypes are scheduled to be finished by mid-2024. In contrast to an SIP, an SIP collects individual chips into one package for every function. Most of the time, it is meant to reduce cost and installation charges by saving space. This dream of Aggarwal about the future Indian supercomputer for AI matches the company’s goal to launch it by the next few years.
In addition, the company is exploring diverse hardware solutions for AI data centers, such as implementing cooling systems that involve submerging servers in insulating and cooling liquids.
Krutrim achieved a significant milestone earlier this year by becoming India’s inaugural unicorn of 2024. This achievement followed a successful fundraising round of $50 million, valuing the company at $1 billion. The investment was secured from a consortium of investors led by Matrix Partners India, a prominent venture capital fund. Unicorns refer to privately-held startups with a valuation of $1 billion or more.
