By 2026, the construction of AI infrastructure will extend beyond just GPUs and networking. They are urgently utilizing all available resources within the data center, particularly focusing on flash storage media. If your data platform was constructed under the belief that flash media would be consistently affordable, abundant, and reliable, the market is now presenting a severe reminder of the current situation.
Gartner classifies our present condition as an AI-powered memory supercycle. Due to the strong correlation between pricing and supply limitations, they anticipate a drastic increase of over 200 percent in the average selling prices per gigabyte for NAND this year. Shortly after the significant price increases of 40 to 2029 percent observed up until late 2025, Gartner predicts that this market trend will continue until 2027, before easing off in the years following.
It’s not only demand that causes volatility. In addition, we are examining the influence of ongoing and new geopolitical unrest on significant energy resources, which in turn is affecting the manufacturing and supply chains of the tech industry. Storage market fluctuations typically follow a cyclic pattern, but the rapid pace and magnitude of AI deployments have resulted in an unusually narrow period of demand.
Lead times for clients have extended to as long as six months.
