As the AI revolution escalated, the demand for Graphics Processing Units (GPUs) surged, resulting in a significant shortage that shows no signs of abating in the near future. The current shortage witnessed by the industry for high-end GPUs could persist until 2025, according to reports. While the industry is gripping a shortage of GPUs, E2E Networks, India’s first AI hyperscaler, is making cutting-edge GPUs more accessible within India, offering competitive pricing in the process.

Founded in 2009, E2E Networks offers NVIDIA’s H100 GPU, based on Hopper architecture, and the NVIDIA A100 Tensor Core GPU at a price point of just INR 412/hour and INR 226/hour respectively. Listed on the National Stock Exchange (NSE), E2E was also the first in India to use the NVIDIA V100 GPUs, the most advanced data centre GPU in the market. Even today, other hyperscalers still do not have a massive presence of GPUs in the Indian subcontinent, according to Tarun Dua, chief executive officer at E2E Networks Limited. 

In an exclusive interaction, Dua told AIM that around 10-15% of the 100 startup unicorns in India were born in the E2E cloud. “We began writing our cloud software around 2013-14. Today, we are very focussed on AI/ML. We have been working on the computational side of AI, and heavy-duty processing based on cloud GPUs.”

Competing with other hyperscalers

Despite the presence of hyperscalers such as AWS, Microsoft Azure and Google Cloud in India, E2E has banked on the multi-cloud approach, where enterprises leverage the services of more than one hyperscaler or cloud services provider to meet their diverse needs. 

“Multi-cloud is a reality and it’s common practice for companies to run different portions of their workloads, typically between 50% to 90%, on their chosen primary cloud platform, while still utilising other hyperscalers and cloud platforms for the remaining 50% to 10% of their services.”

Dua also states that there is a massive learning curve when it comes to using a hyperscaler’s services. “However, India being a ‘you do it for me’ kind of a country, we are aligned to that thought process that customers need a much easier interface to work with. They also need a lot of hand-holding.”

But the significant distinction, as pointed out by Dua, lies in E2E’s dedicated efforts to tailor its product to meet the specific needs of its customer base, which primarily consists of clients with specific workload demands in areas such as AI/ML and SME enterprise workloads.“We are very focused on making it super easy for our customers to use our platform. We cater to our customer’s specific needs whereas the other hyperscalers are focussed on doing everything for everybody.”

Moreover, when it comes to pricing, E2E also holds a significant advantage. According to Dua, a customer can easily save 50-60% of their cost by choosing E2E Networks. “In contrast to hyperscalers, where a single service is often built upon five different parameters, our platform is significantly more straightforward. You pay for precisely what you see. Typically, a single service is billed based on a single parameter, never more than one.”

More GPUs coming India’s way 

Currently, E2E Networks has close to 600 GPUs across its two data centres. However, the plan is to add more GPU capacity to meet the growing demand, driven by generative AI. “The number would change very quickly,” Dua said.

However, the significant growth in demand for AI has led to a shortage of high-end GPUs in the market, which are mostly developed by NVIDIA. Dua believes in the short run, the GPU shortage definitely affects everyone, however, “It’s not like we have run out of GPU capacity.”

He believes the shortage will soon be rectified and that the top GPU vendors have done relatively well in meeting most of the demand. “Given India’s growing priority on the global tech landscape, I believe we’ll witness an influx of GPUs into the country. I don’t view the GPU shortage as a long-term issue for India.”

While all the GPUs currently owned by E2E are from NVIDIA, Dua does not write off the possibility of acquiring GPUs from other vendors such as Intel or AMD. “We are proud partners with both AMD and Intel, primarily due to our strong relationships with them concerning their CPUs. Additionally, we will gladly evaluate their GPUs when they introduce them to us. However, as of today, they haven’t established the complete plug-and-play ecosystem that NVIDIA has built over time.”

Catering to Indian clientèle

Currently, E2E serves a diverse clientele, with nearly 3,000 active customers. Some of our top clients include prestigious institutions like IIT Madras, IIIT Bangalore, ISB, and innovative companies like CamCom AI, which has been a long-time user of our E2E GPU servers. “Besides, there are a dozen other startups who are using our services for computer vision and Natural Language Processing (NLP). 

“They have found our services to be incredibly cost-effective, offering them the best price-performance ratio in the Indian market. With E2E Networks, they have enjoyed a speedy response from our support and sales teams, ensuring their operations run smoothly and efficiently. Our clients often praise the predictability of our pricing, which allows them to budget effectively, and our prepaid services, which offer flexibility and convenience,” he concluded.