India's AI firms hit by severe GPU shortages despite ample funding
AI startups in India are struggling to obtain GPUs, memory and power for model training, with delivery delays of three to four months.
Funding streams for Indian artificial-intelligence firms are robust, yet the sector is bottlenecked by a shortage of GPUs, high-bandwidth memory and reliable power. Start-ups often get free cloud credits, but once they order physical GPUs, deliveries can take three to four months, causing entire engineering teams to lose a quarter of productivity. Companies such as Juspay and robotics venture CynLr cite the same delays, with Juspay’s COO Sheetal Lalwani noting a net loss after heavy AI-infrastructure spending.
Venture investors like Rahul Agarwalla and Ashish Taneja warn that securing capacity in advance is difficult for early-stage firms. Memory supply from SK hynix, Samsung and Micron is already fully committed for 2026, and data-centre power is increasingly locked by large players. The government’s IndiaAI Mission has made over 38,000 GPUs available, but specialized workloads and security-sensitive firms need dedicated rack space that the pool does not provide.
Why it matters
GPU shortages limit India's AI growth, affecting startups, investors and the country's tech competitiveness.
In this story
