Sridhar Vembu says it is becoming difficult to do business after memory prices climb 500% in 12 months

The memory crisis is showing no signs of slowing down. Thanks to demand for memory from AI data centres, memory chips have become more expensive. Now, Zoho founder Sridhar Vembu has admitted that the situation has reached a point where even he is finding it difficult to do business.

On X, Vembu shared a post citing a recent report that said, “Memory prices up 500 per cent in 12 months.” The Zoho founder claimed that this hike in prices, coupled with rising AI token costs was becoming a major problem. “Memory prices, along with AI token prices, have made business very difficult. We have held back from raising prices but it is becoming hard,” he wrote.

To give you some context, as per a report by Tom’s Hardware, there has been a steep rise in memory costs. Average prices for DDR5 memory between August 2025 and August 2026 rose from $90 (roughly Rs 8,600) to $425 (roughly Rs 40,000) for certain kits.

Higher-capacity kits have seen even larger jumps. A 128GB DDR5-6400 kit is now priced at $3,399 (roughly Rs 3,25,000), 10 times its lowest-ever tracked price of $329 (roughly Rs 31,490).

Srihdar Vembu’s comments come at a time when the entire tech industry is facing pressure from rising prices. Phonemakers, including Oppo, OnePlus, Vivo, and Nothing, have all raised smartphone prices in the past few months. There are rumours that Apple may also hike iPhone prices sometime soon. It appears that apart from device manufacturers, even software companies like Zoho are taking a hit.

Sridhar Vembu says programming languages need to evolve

The Zoho chief believes that this memory crisis may mark the end of how we design programming languages today. “For a long time, programming languages were designed with the assumption that memory is ‘free.’ That era has now ended,” Vembu added.

Instead, Sridhar Vembu claimed that it was time for programmers to develop more efficient compilers, even for AI. He wrote, “We need highly memory efficient languages and smarter compilers, even for AI to use.” Going forward, Vembu insisted that he was going to look into how programming languages can take up less memory. “Safety of code and productivity in writing code cannot come at the expense of memory bloat. That is my area of research,” he explained.

At the same time, AI companies are doubling down on setting more data centres globally to meet rising demand. OpenAI has just signed a deal with SoftBank’s SB Energy for a 10-gigawatt data centre in the US city of Ohio – making it one of the largest data centres to date. This deal is partly backed by a commitment from Nvidia.

Indian companies are also setting up data centres. HCLTech has announced plans to invest Rs 3,500 crore in data centres, while TCS plans to spend as much as $7 billion to set up a 1-gigawatt data centre unit.

– Ends