Zoho founder Sridhar Vembu: I have come to believe that in software code generation …

Sridhar Vembu has said that artificial intelligence (AI) can be developed in a more affordable and sustainable way, with far lower energy and computing needs than what is commonly used today. In a post shared on X (formerly known as Twitter), Vembu pointed to the work done by Sarvam AI as an example of how high-quality AI systems can be built without heavy reliance on large-scale infrastructure. For those unaware, Sarvam AI is an AI model founded by IIT alumni – Pratyush Kumar and Vivek Raghavan. The Bengaluru-based startup recently claimed that the AI model has outperformed some of the biggest names in artificial intelligence, including Google Gemini and OpenAI’s ChatGPT.In the post, Vembu said Sarvam AI’s work shows that “world-class AI” does not require excessive compute power or energy consumption. He added that this approach is important as concerns grow over the environmental impact of large AI models, which rely on energy-intensive data centres.The Zoho founder also said he believes software code generation, in particular, offers scope for more efficient AI techniques. According to Vembu, new approaches could significantly reduce the energy and compute footprint needed to build and run such systems.Sridhar Vembu further warned that the current trajectory of AI development may not be sustainable in the long run. Vembu said the planet cannot support the rising energy demands of today’s AI models, highlighting the need for alternatives that balance performance with environmental responsibility.His comments come amid broader global debate over the cost, scalability, and environmental impact of advanced AI systems, as companies continue to invest heavily in computing infrastructure to train and deploy large models.

Here’s what Sridhar Vembu wrote on X

What Sarvam AI has demonstrated – world class AI can be done much more affordably and sustainably with much lower energy and compute footprint – is important.I have come to believe that in software code generation, much more energy and compute efficient approaches to AI can be developed. That is important because the earth cannot afford today’s AI energy footprint.