Nicolas Frapard, senior manager at Western Digital, states that beyond a certain scale, AI's primary challenge shifts from compute resources to data storage, driven by exponential data volume growth from sustained, large-scale AI deployment. He emphasizes that HDDs remain the most viable and economical solution for storing vast AI datasets at exabyte scale, with flash being up to 20 times more expensive per terabyte.
AI infrastructure TCO increasingly depends on tiered storage strategies, making high-capacity HDDs a critical, cost-effective component for exabyte-scale AI data estates.