Bank of America says an AI-memory driven chip rally could be more durable than investors expect

Bank of America is taking a more durable view of the semiconductor rally, according to a Yahoo Finance report published June 25. The note, attributed to the bank’s research team, frames the current cycle as supported not just by general compute demand, but by a specific bottleneck tied to artificial intelligence: memory used to run and store AI workloads.

The report’s core argument is that AI memory demand could make the cycle “more durable,” implying that investors may be underestimating how long the supply-demand balance could stay favorable for chip makers and parts of the AI supply chain. Memory is increasingly central to AI systems, which rely on fast storage and large working memory to handle training and inference workloads.

While the Yahoo Finance piece focuses on the durability thesis rather than a detailed list of companies or specific price targets, the implication is that the semiconductor trade may have more than one growth engine. In previous AI-driven market narratives, the spotlight has often centered on accelerators such as GPUs and specialized AI processors. This time, the emphasis shifts toward the memory side of the stack.

This matters for market participants because semiconductor cycles are typically linked to capital spending by customers and the speed at which supply can be matched to demand. If AI memory requirements continue to grow as systems scale, that could translate into longer stretches of order visibility for suppliers than a more purely compute-driven story might suggest.

The report comes at a moment when investors are trying to separate near-term earnings momentum from longer-run structural demand tied to AI adoption. A durability argument, if reflected in company guidance, can influence how traders think about whether semiconductor outperformance is likely to fade quickly or hold up through additional quarters.

For Bank of America, the thesis also fits its broader role as a market barometer for large-cap technology and cyclical sectors. The bank’s research frequently spotlights where it sees constraints in the AI supply chain, and whether those constraints are likely to persist as customer deployments expand.

One limitation is that the Yahoo Finance item, as presented in the available materials, does not provide granular details such as the identity of the exact semiconductor segments singled out, the names of specific stocks, or the quantitative estimates behind the durability claim. It also does not disclose any clear timeline for when the memory-driven support would be expected to show up in earnings.

Looking ahead, market watchers will likely look for follow-through from both memory and wider chip supply chain players, such as commentary on demand visibility, order backlogs, and guidance. They will also watch whether additional bank research or company updates echo the same “AI memory” durability framing, or whether it remains a one-off call in the latest market narrative.