Sridhar Vembu, founder and chief scientist of Zoho, has issued a stark warning regarding the current artificial intelligence investment boom. Vembu suggests the rapid influx of capital into AI could culminate in a market distortion similar to the dot-com and telecom bubble of the late 1990s, despite the underlying technology's genuine potential.
AI Boom's Market Distortion
Vembu highlighted that the anticipated sharp increase in prices for smartphones and laptops by 2026 serves as a clear indication of the AI investment boom's market impact. He elaborated on social media, stating, "That is the AI investment boom, funded (ultimately) by the mother of all credit bubbles unleashed post-pandemic."
According to Vembu, this boom has significantly inflated demand across numerous sectors beyond just AI chips. Industries experiencing heightened pressure include electric power generation, transformers, diesel backup generators, cooling systems, memory, CPUs, and GPUs.
Echoes of the Telecom Bubble
Drawing parallels to the telecom investment frenzy between 1995 and 2000, Vembu noted that a market bubble can form even when the core technology is transformative. He cited the widespread availability of high-speed optical fiber connections in remote areas today as a testament to the telecom technology's eventual success.
"Yet that did not prevent most telecom equipment companies from going bankrupt in 2001-3," Vembu wrote. "The AI credit boom will meet the same fate, and I actually do believe the technology will achieve its full potential."
Vembu emphasized the delicate balance required: staying optimistic about AI technology while rigorously avoiding overexposure to the speculative investment bubble it has created. "Staying positive on the tech while being careful to not lose one's shirt in the bubble unleashed by boosters of the same tech - that is the careful balancing act needed today," he advised.
Rising Memory Costs and Software Development
Further exacerbating market pressures, Vembu commented on reports indicating a 500% surge in memory prices over 12 months, reaching ten times their previous lowest levels. "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 stated.
These escalating memory costs, Vembu believes, will necessitate fundamental changes in software development paradigms. He argued, "For a long time, programming languages were designed with the assumption that memory is 'free'. That era has now ended." Future development, including for AI applications, will require more memory-efficient programming languages and smarter compilers, prioritizing code safety and productivity without the burden of memory bloat.