Memory chip makers are riding a fresh wave of investor optimism. Shares of Micron and SanDisk jumped Monday after Melius Research published a bullish report projecting that demand for memory chips will stay robust through the end of the decade, driven largely by AI infrastructure buildouts and data center expansion. The analyst call arrives as the semiconductor industry debates whether the current AI-driven boom has staying power or represents another cyclical peak.

Micron Technology and SanDisk are suddenly back in Wall Street’s good graces. Both memory chip manufacturers saw their stocks climb Monday following a research note from Melius Research that projects memory demand will remain elevated through 2030, a timeframe that extends well beyond typical semiconductor forecasting windows.

The timing couldn’t be more critical for an industry that’s grown accustomed to brutal boom-bust cycles. Memory chips, the workhorses that store data in everything from smartphones to massive AI training clusters, have historically suffered from chronic oversupply as manufacturers race to build capacity during good times, only to watch prices crater when demand softens. This time, Melius Research argues, things look different.

What’s driving the sustained optimism? AI infrastructure, plain and simple. The explosive growth in large language models and AI training workloads has created what appears to be insatiable demand for high-bandwidth memory (HBM) and advanced NAND flash storage. Every new AI model requires exponentially more memory capacity than its predecessor, and the buildout of data centers optimized for AI inference is just getting started. Microsoft, Google, and Amazon are all racing to expand their AI-capable infrastructure, creating a structural tailwind for memory suppliers.

Micron has been particularly aggressive in positioning itself for this shift. The company has poured billions into developing next-generation HBM products specifically designed for AI accelerators, the specialized chips that power everything from OpenAI’s ChatGPT to enterprise AI applications. SanDisk, meanwhile, has focused on the storage side of the equation, developing high-capacity solid-state drives that can handle the massive datasets AI systems require.

But the Melius Research forecast isn’t just about AI hype. The report points to broader enterprise digitization trends, edge computing growth, and the ongoing automotive industry’s shift toward software-defined vehicles, all of which require substantially more memory and storage than previous generations of technology. The convergence of these trends creates what analysts describe as a “structural demand environment” rather than a cyclical uptick.

Investors have learned to be skeptical of perpetual-growth narratives in semiconductors. The industry has seen this movie before, most recently in 2018 when memory chip prices collapsed after a frenzied buildout led to massive oversupply. Memory manufacturers watched their margins evaporate as prices for DRAM and NAND chips fell by double digits quarter after quarter. The question now is whether the current supply-demand dynamics are fundamentally different.

The bear case remains straightforward – if memory demand is genuinely this strong, competitors will inevitably add capacity, eventually bringing supply back into balance and pressuring prices. Samsung, the world’s largest memory chip manufacturer, has already announced plans to increase production. Chinese manufacturers are also investing heavily despite ongoing trade restrictions.

Yet the bull case has teeth this time around. AI memory requirements aren’t just incrementally higher, they’re orders of magnitude larger than traditional computing workloads. A single AI training cluster can consume as much HBM as thousands of conventional servers. And unlike previous demand drivers like smartphones or PCs, where units sold eventually plateau, AI model complexity and deployment scale show no signs of slowing.

The stock market’s reaction suggests investors are willing to bet on the optimistic scenario, at least for now. Both Micron and SanDisk have seen their valuations expand as semiconductor investors rotate toward memory plays that offer direct exposure to AI infrastructure spending. The sector had been trading at depressed multiples for months as concerns about Chinese competition and cyclical headwinds weighed on sentiment.

What happens next depends largely on pricing. If memory chip manufacturers can maintain pricing discipline even as they add capacity, the extended demand runway that Melius Research envisions could translate into sustained profitability. But if competitive pressures force prices down faster than costs decline, even strong unit demand won’t save margins. The industry’s track record on the pricing discipline front is, charitably, mixed.

For now, though, memory chip investors are enjoying the rally. The Melius Research report provides a fundamental justification for bullish positioning at a time when AI infrastructure spending shows no signs of slowing. Whether that demand actually persists through 2030 remains to be seen, but the market is clearly betting that this semiconductor cycle has legs.

The memory chip rally sparked by Melius Research’s bullish forecast captures the tension at the heart of the AI boom – is this time truly different, or are we watching another semiconductor cycle play out with a new narrative? For Micron and SanDisk, the answer will depend on whether AI infrastructure demand can actually sustain the kind of multi-year growth trajectory that justifies current valuations. The structural drivers look compelling, but semiconductor investors have been burned before by stories of limitless demand. What’s clear is that the memory chip sector has suddenly become one of the more interesting battlegrounds for investors trying to figure out which parts of the AI infrastructure stack offer durable growth versus cyclical opportunities.