Investor Dan Niles views the recent weakness in AI chip stocks as a temporary setback driven by Apple's potential China memory exemption and Meta's possible cloud expansion, not a peak in AI demand. He expects long-term growth from Nvidia's Vera Rubin platform, which will require significantly more memory than current systems, and sees parallels to 1990s semiconductor cycles that recovered strongly after corrections.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Tech investor Dan Niles said the recent pullback in AI semiconductor stocks is temporary rather than a major reversal. He attributed the decline to concerns that Apple could receive approval to source memory chips from China, and to speculation that Meta might monetize excess AI computing capacity through a cloud offering.
Why it matters
Niles remains bullish on memory-chip makers over the longer term, noting that Nvidia's upcoming Vera Rubin AI platform is expected to require "at a minimum over 3x more memory than a Blackwell platform." He compares the current cycle to the late 1990s internet infrastructure boom, when semiconductor stocks experienced steep corrections before climbing to much higher levels.
What to watch
Niles flagged several risks that could pressure AI spending later in the year, including more efficient AI models that require fewer computing resources, rising competition from low-cost open-source models, and increasing semiconductor costs. He plans to remain bullish until large-cap earnings begin in late July.
The recent pullback in AI semiconductor stocks reflects near-term investor concerns rather than a fundamental shift in demand, according to Niles. Two specific catalysts weighed on sentiment: the prospect of Apple securing a partial exemption to purchase memory from China—a move that could ease supply constraints but also reduce demand for non-Chinese suppliers—and Meta's exploration of a cloud computing offering. While both developments initially caught market attention, Niles argues they represent cyclical headwinds, not an end to the AI spending cycle.
Niles draws a historical parallel to the late 1990s internet infrastructure boom, when semiconductor stocks endured steep corrections before climbing to much higher levels. The forthcoming Vera Rubin platform, requiring substantially more memory than Nvidia's current Blackwell generation, anchors his bullish case: the installed base of AI systems will need upgraded memory capacity, supporting both near-term capital spending and longer-term chip demand. However, he acknowledges material risks on the horizon—more efficient models that reduce compute requirements, competition from low-cost open-source alternatives, and rising semiconductor costs—which could dampen spending later in the year. His own positioning reflects this layered view: constructive through late July, when large-cap earnings reports may reset expectations.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
