macro challenges hinder Zoho's overall growth

 macro challenges hinder Zoho's overall growth

MUMBAI: Zoho Corp. is seeing a halt in sales growth as a result of the macroeconomic difficulties in its important North American market. Since many of the issues are unlikely to be rectified quickly, the company has been more cautious about investing and hiring, according to chief executive Sridhar Vembu.

"There will be sporadic changes. However, you must also consider the long-term context. Because of this, we became extremely careful over a year ago, and we still are, when it comes to employment and other costs," Vembu added.

Zoho, which was established in 1996, announced $1 billion in operating revenue in November of the previous year. Although the company's revenue is still increasing more slowly than it was previously, it is nevertheless "very cautiously" employing.

However, Vembu is upbeat about long-term tech outsourcing investment and noted that firm is still confident about India's growing IT spending.

Additionally, Zoho is in discussions to invest in the production of sophisticated chemistry cells, which are a vital part of batteries.

"We are looking at this area and corresponding with several interested parties. We need to develop deep tech and manufacturing capabilities in our country, but it's too early to say that. Advanced cell chemistry production-linked incentive (PLI) systems offer opportunity here as well. We're currently in talks with several good people who have experience in cell R&D, he said.

This effort can, however, be managed by subsidiaries and integrated subsequently. "For instance, we have a subsidiary in medical equipment where battery technology can be integrated," Vembu continued.

In order to serve particular industries, Zoho is now creating "domain-specific" artificial intelligence (AI) models. This, according to Vembu, also aids the business in maximizing its long-term capital expenditures.

really expensive to train and run are really large AI models. On the other hand, simpler models with between 10 and 20 billion parameters can perform admirably in a single domain and are significantly less expensive to run. These models can frequently fit on one or two GPUs, as compared to the huge language models with hundreds of billions of parameters, which require hundreds of GPUs. We would concentrate on that area since it offers the most leverage. This aids us in resolving issues unique to a given domain. We've actually begun construction on them," he declared.

Not just Zoho is attempting to capitalize on the desire for AI implementations across businesses. Ziad Asghar, senior vice president for product management at Qualcomm, stated on September 15 that the fabless chipmaker was trying to develop more compact AI models for use cases that required offline processing. Compared to large language models (LLMs), which can include hundreds or even thousands of billions of parameters, these smaller AI models typically have tens of billions of data parameters.

Vembu said that building such models is also less expensive. Smaller AI models need $5 million to $10 million in resources, he noted, as opposed to larger models, which cost $100 million to $200 million to create. In order to build and incorporate open standards, Zoho will also collaborate with government-backed AI initiatives like the local-language LLM project "Bhashini".

Vembu was giving a speech at a business conference in Bengaluru when Cliq Rooms, a platform for remote collaboration, was introduced.