THE BIG TECH BOOST

START-UPS OPERATE IN A COMPLEX ENVIRONMENT AND MUST PRIORITISE ENTERPRISE TECHNOLOGY IF THEY WANT TO SCALE. HERE'S HOW THEY ARE GETTING AFFORDABLE OPTIONS 

- NIDHI SINGAL

WE DON'T KNOW if the founders of OpenAI, the 2015 start-up behind ChatGPT, asked the artificial intelligence chatbot: Where will we get the money and the technology to make our dream come true? ChatGPT would have remained silent: it had not yet built its database. In 2019, Microsoft answered by investing $1 billion in OpenAI and recently added $10 billion. OpenAI launched ChatGPT in November 2022. Microsoft agreed to build a supercomputer by linking thousands of Nvidia GPUs or graphics processing units on its Azure cloud computing platform, allowing OpenAI to train increasingly powerful models.

But not every start-up can secure the endorsement or collaboration of the world’s largest tech company. What’s the alternative? Opt for software as a service (SaaS), infrastructure as a service, or platform as a service. All these are now being introduced by large, established companies that have been around for decades.

Yotta Data Services, a realty major Hiranandani Group company, is building India’s largest—and possibly the world’s 10th largest—supercomputer. It has ordered over 16,000 H100 Tensor Core NVIDIA GPUs, the latest GPUs used by the likes of Microsoft and Oracle. “The US, China and many other countries are far ahead of us in AI. They are working feverishly on AI models, especially generative AI models,” says Sunil Gupta, Co-founder, MD & CEO of Yotta Data Services. Gupta says if you make any large-scale AI model for training, you need a huge amount of computing and network capacity, which, with the Yotta Supercomputer going, will be easily accessible to Indians.