By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panelists agree that while AI-driven demand is driving high-bandwidth memory and storage, the structural shortage narrative is overstated and likely to be resolved by 2026-2027 with new capacity from Samsung, SK Hynix, and Intel. They caution about the cyclical nature of the semiconductor industry and the risk of a supply glut once capacity comes online.
Risk: A supply glut and subsequent pricing power loss once capacity comes online in 2026-2027.
Opportunity: Potential short-term gains in memory and storage stocks due to AI-driven demand and current supply tightness.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Artificial intelligence is driving one of the most powerful investment cycles in modern history. The spending is staggering, the innovation is relentless, and the adoption curve keeps steepening. Every quarter brings fresh confirmation that this is not a trend, but rather a structural shift reshaping the global economy from the ground up.
The rally over the last month has shown us just that, as AI related names lifted the S&P and Nasdaq to fresh all-time highs.
But here is the part investors need to understand: the biggest gains in this cycle will not go to those who chase the headlines. They will go to those who understand the volatility, respect the pullbacks, and use the chaos to their advantage.**The Tale of The Tape: An Extraordinary Run **
To appreciate the opportunity, you first have to note what has already happened.
Nvidia has been the defining winner of this boom, transforming itself from a graphics pioneer into the backbone of modern AI data centers. The stock is up over 1,000% since the start of 2023. Fiscal year 2026 revenue came in at $216 billion, up 65% year over year, with earnings per share growing 60% to $4.77.
But NVDA is only part of the story. The AI trade has rotated, and memory stocks have taken the baton.
From early 2023 through mid-April 2025, Nvidia dominated, but the game flipped early in 2026. SanDisk, Western Digital, Seagate, and Micron have been outperforming Nvidia and have soared over the last couple months.
The numbers in 2026 have been jaw-dropping. SNDK has rocketed more than +400% year-to-date. Micron has moved more than +130% and continues to follow Sandisk’s parabolic lead.
These are not small-cap moonshots. These are established names in the semiconductor food chain getting repriced as the market wakes up to the memory shortage embedded in the AI buildout.
Image Source: Zacks Investment ResearchWhy Memory? The Structural Case
AI is creating a massive ripple effect across the entire tech supply chain, and it is not just about Nvidia GPUs anymore. Every new AI model and every new data center being built also needs huge amounts of memory and storage chips to function.
The problem is the industry cannot build enough of those chips fast enough, and many analysts do not expect supply to catch up until at least 2027. That shortage is already showing up in company results.
Micron Technology says its entire 2026 supply of high bandwidth memory chips, which are critical for AI systems, is already sold out.
Western Digital says its high-capacity drives are fully committed to large cloud and hyperscale customers.
Meanwhile, Nomura believes SanDisk could raise data center storage prices sharply as shortages worsen.
This is what real pricing power looks like. Demand is overwhelming supply, and the companies making the critical AI infrastructure are gaining leverage because customers simply cannot get enough product.The Volatility Is Real
The march higher has not been a smooth ride. NVDA was down over 10% in late March of this year alone, even as its fundamentals hit all-time highs. Memory stocks swung violently with every tariff headline, geopolitical development, and sentiment shift.
Bubble talk surfaced repeatedly, and weak hands got shaken out.
This is the pattern we have seen throughout the history of the stock market. It happened with the railroads in the 1800s and it happened with the internet in the late 1990s.
Every transformational technology cycle features brutal shakeouts that feel catastrophic in the moment and look like gifts in hindsight.
The investors who won big were not the ones who avoided the volatility. They were the ones who had a process for navigating it.**Turning Volatility into Opportunity **
The single biggest mistake investors make in a cycle like this is treating every pullback like the beginning of the end. The second biggest mistake is chasing every rip without a plan. Both destroy returns.
Here is what works instead.
More . . .
------------------------------------------------------------------------------------------------------See ALL Zacks’ Long-Term Picks for Only $1
Through good markets and bad, one unique stock-picking method has clobbered the market’s average gain with an incredible +23.9% per year.
To help you take advantage of opportunities in today’s market, we’re opening the vault to reveal all our long-term recommendations. You’ll see stocks priced under $10… income investments… hidden value stocks and more.
Our private long-term investing services have closed 14 double and triple-digit wins so far this year. Although we can't guarantee 100% success, recent gains have reached as high as +141.86%, +218.59%, and +230.3%.¹
All for just $1.
Special opportunity ends at midnight on Sunday, May 24th. **See Stocks Now >> **
------------------------------------------------------------------------------------------------------ Use the Zacks Rank to Anchor Your Conviction — Pullbacks feel dangerous, because the news headlines drive down stock prices and make investors hesitant to pull the trigger. The Zacks Rank will help investors cut through that noise.
When earnings estimates are rising and the rank is strong, the fundamentals are telling you something the price action may be temporarily ignoring. That divergence is where opportunity lives and returns will come.
— A strong fundamental story needs a positive chart to maximize reward-to-risk. Investors can turn to volume indicators, moving averages, and Fibonacci levels to give them entry points and define risk/reward opportunities.
Use Technicals to Time Your EntryUse Technicals to Time Your Entry
After months of choppy, volatile trading marked by sharp rallies and pullbacks, a decisive bullish technical turn is often the clearest signal that the next leg is beginning.
Investors will have more success if they avoid buying the story and focus on the setup.
— In volatile markets, managing risk matters more than picking the perfect stock. A good setup is one where the potential upside is much larger than the downside. For example, if you risk losing $1 to potentially make $5, you can still come out ahead even if most of your trades do not work.
Respect Reward-to-Risk Above Everything ElseRespect Reward-to-Risk Above Everything Else
That is where many retail investors struggle. They spend all their time chasing the next hot AI name but pay far less attention to entry points, stop losses, and position sizing. In a fast-moving sector like AI, discipline and risk management are often more important than being right all the time
— A high Zacks Rank plus strong technical indicators. In a sector with this much fundamental momentum behind it, that combination is a green light. SNDK is the cyclical kicker for investors who are comfortable with volatility. MU is the durable memory core. NVDA remains the anchor of the entire infrastructure buildout. Each of these names has given back meaningful ground at various points in 2025 and 2026, and each time, the setup reset, and the next move higher began.
.
When Signals Align, Act with ConvictionWhen Signals Align, Act with Conviction
Embrace Volatility for Long-Term Gains
The AI boom is no longer just a story about one company or one chip. It has become a full-scale infrastructure buildout touching every corner of the semiconductor supply chain, from GPUs to memory to storage.
But investors should not expect a straight line higher.
The biggest opportunities in transformational market cycles almost always come wrapped in volatility. Sharp pullbacks, scary headlines, and sudden sentiment swings are not signs the trend is broken. More often, they are the moments that reset the next move higher.
That is why success in AI investing will not come from chasing hype or reacting emotionally to every market swing. Instead, it will come from staying grounded in the fundamentals, waiting for high probability setups, and managing risk with discipline.
The trend is still intact. The spending is still accelerating. And for investors willing to navigate turbulence instead of fearing it, the AI cycle may still be in its early innings.
Here’s the good news for long-term investors…a lot of great stocks are on sale, and we may potentially see more fantastic deals in the near future. But how do you find stocks that have a real shot at significant outperformance?
The Zacks Rank is a great place to start. Stocks rated highly by the Zacks Rank have the strongest underlying earnings trends. Buying these stocks on dips can put the odds strongly in your favor.
Short-term market fluctuations happen at a much greater frequency than a reversal of an earnings trend. That means investors can use them to find stocks that stand to profit – and continue to profit – despite what’s happening with the global political landscape.
To help you find today’s top stocks, we've opened up our
, which gives you access to recommendations from all of our long-term investor portfolios. It's made for markets like this.
Zacks Investor CollectionOver the next 30 days, for the total cost of only $1, you're invited to access:
• Real-time buys, sells, and exclusive market commentary from
Stocks Under $10,
Home Run Investor,
Value Investor,
ETF Investor, and
Income Investor.
• Plus, you'll get full access to our powerful research, tools, and analysis, including the Zacks #1 Rank List, Equity Research Reports, Premium Screener, and more, as part of
Zacks Premium.
Our private long-term investing services have closed 14 double and triple-digit wins so far this year. Although we can't guarantee 100% success, recent gains have reached as high as
+141.86%,
+218.59%, and
+230.3%.¹
Bonus Report: If you act quickly, you'll also get our exclusive Special Report,
5 Stocks Set to Double. It reveals 5 companies poised for immediate breakout. Each stock was selected because of its superior short-term profit potential under current market conditions.
I encourage you to take advantage right away. The earlier you get in, the greater profits you stand to make. But don't delay. This restricted $1 opportunity ends at midnight Sunday, May 24th.
Click here for
Zacks Investor Collectionand5 Stocks Set to Double>>Good Investing,
Jeremy Mullin
Jeremy Mullin is a stock strategist who combines the fundamental power of the Zacks Rank, technical analysis, and computer driven trading to find the best trades. Discover all his current recommendations in the
Commodity Innovatorsand
Zacks Counterstrikenewsletters.
¹ The results listed above are not (or may not be) representative of the performance of all selections made by Zacks Investment Research's newsletter editors and may represent the partial close of a position. Access grants you a comprehensive list of all open and closed trades.
This article originally published on Zacks Investment Research (zacks.com).
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Memory pricing power is real but likely transitory once new supply ramps, making current multiples vulnerable to rapid mean reversion."
The article correctly flags AI-driven demand for high-bandwidth memory and storage but overstates structural permanence. Micron’s sold-out HBM position and Western Digital’s committed drives reflect 2025-2026 tightness, yet new capacity from Samsung, SK Hynix, and Intel’s Ohio fab plus potential demand digestion by late 2026 could flip pricing power quickly. NVDA’s 1,000% run and cited 2026 figures also sit well ahead of consensus, suggesting the piece blends realized results with optimistic projections. Volatility is acknowledged yet framed mainly as entry opportunity, downplaying the sector’s history of 50%+ drawdowns when supply and capex cycles realign.
Even if shortages persist into 2027, aggressive capex by memory makers could restore equilibrium faster than modeled, triggering the same margin collapse and multiple compression seen after prior cycles.
"The memory shortage is real but cyclical, not structural—and the article's 2027 supply constraint thesis lacks sufficient evidence to justify the 130-400% YTD gains already priced in."
The article conflates two separate narratives that deserve scrutiny. Yes, memory/storage stocks (MU, SNDK, WDC) have outperformed NVDA recently—that's verifiable. But the article attributes this to a structural shortage lasting until 2027, citing sold-out HBM inventory and committed capacity. This is partially true but incomplete: memory cycles are notoriously cyclical, not structural. The 2024-2025 memory shortage was real, but DRAM/NAND capacity additions are ramping aggressively (Samsung, SK Hynix, Kioxia all expanding). The article also buries a critical detail: it's written as promotional content for Zacks' paid service, which creates incentive to extend bullish narratives beyond what fundamentals support. The volatility discussion is sound, but the 'structural shortage until 2027' claim needs hard evidence, not analyst opinion.
Memory supply is already normalizing faster than consensus expected—spot prices for DRAM and NAND have rolled over since Q1 2026, and new fab capacity is coming online ahead of schedule, which could compress margins for MU and SNDK by 2H 2026 precisely when the article suggests they'll accelerate.
"The current memory shortage is a cyclical peak, and investors are underestimating the risk of a supply-side glut once massive CapEx investments hit the market in 2027."
The article correctly identifies the rotation from pure compute (Nvidia) to memory and storage (Micron, Western Digital) as the AI infrastructure buildout matures. However, the piece reads more like a marketing funnel than a rigorous financial analysis. While the 'memory shortage' thesis is currently supported by high-bandwidth memory (HBM) demand, it ignores the cyclical nature of the semiconductor industry. Memory is a commodity; once capital expenditure (CapEx) catches up to demand, pricing power evaporates overnight. Investors should view the current 130%+ gains in Micron as a reflection of peak cycle optimism rather than a permanent structural shift. I am cautious, as the 'sold out' narrative often precedes a supply glut when capacity finally comes online in 2027.
If AI adoption continues to grow at an exponential rate, the 'commodity' cycle for HBM may be fundamentally broken, leading to a permanent, higher-margin pricing regime for memory providers.
"The AI hardware cycle is highly cyclical and vulnerable to demand slowdown or supply normalization that could erase much of the recent outperformance in NVDA and memory names."
Despite the bullish narrative, the article glosses over how fragile this cycle can be. AI hardware demand hinges on a sustained hyperscaler capex impulse; any moderation in cloud spending or a macro slowdown could trigger rapid multiple compression. The memory shortage thesis depends on delayed supply: if fabs ramp by 2027, pricing power fades and MU/SNDK/WDC revert toward the midcycle mean. Nvidia’s dominance may face competition from more efficient AI models or export controls. With valuations already pricing a best‑case scenario, a weaker macro environment could curb enterprise AI spend and derail the offshore AI buildout.
The strongest countercall is that demand stays resilient as hyperscalers continue capex beyond expectations. Memory constraints persist and Nvidia’s moat could widen via software and ecosystem lock-in.
"HBM qualification delays could push any supply glut past 2027 despite announced capacity additions."
Gemini notes the commodity risk in memory but overlooks how HBM requires lengthy customer qualification cycles that Samsung and SK Hynix have historically missed by quarters. A single delay in 2026 ramp certification would keep MU and WDC sold out into 2027, preserving margins even if total bit supply rises. This technical friction, not just AI demand, is what separates this cycle from 2017-2018.
"Sold-out inventory preserves revenue but not margin if pricing power erodes faster than volume growth."
Grok's qualification-cycle friction argument is plausible but needs specificity. Samsung and SK Hynix have actually accelerated HBM certification since 2023—NVIDIA's HBM3e qualification happened faster than prior cycles. The real risk Grok sidesteps: even if MU stays sold out, gross margins compress if ASPs (average selling prices) fall 15-20% while volumes rise. Shortage ≠ profitability. That's the 2027 trap.
"Yield-constrained HBM production will keep average selling prices elevated, insulating memory margins from simple capacity-based supply gluts."
Claude, you’re missing the secondary impact of the HBM qualification bottleneck. It isn't just about volume; it’s about yield. If Samsung and SK Hynix struggle with HBM3e/4 yields, the effective supply is lower than nominal capacity suggests. This keeps ASPs elevated for Micron regardless of total bit supply growth. The '2027 trap' ignores that hyperscalers prioritize supply chain security over cost, likely sustaining premium pricing even if capacity technically hits the market by mid-2026.
"Yield dynamics may be the deciding variable between a soft landing and a sharper margin re-rating"
Gemini's yield-focused angle adds nuance, but it's still under-specified. If HBM3e/4 yields disappoint, effective supply tightness could linger, propping up ASPs and margin despite capacity ramp (the 2027 trap becomes a 2027+ trap). Conversely, strong yields would unleash the same pricing pressure everyone fears. My key claim: yield dynamics may be the deciding variable between a soft landing and a sharper margin re-rating.
The panelists agree that while AI-driven demand is driving high-bandwidth memory and storage, the structural shortage narrative is overstated and likely to be resolved by 2026-2027 with new capacity from Samsung, SK Hynix, and Intel. They caution about the cyclical nature of the semiconductor industry and the risk of a supply glut once capacity comes online.
Potential short-term gains in memory and storage stocks due to AI-driven demand and current supply tightness.
A supply glut and subsequent pricing power loss once capacity comes online in 2026-2027.