By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panelists generally express skepticism about the long-term prospects of IREN, SNDK, and ZETA, with concerns around high valuations, execution risks, and potential margin compression.
Risk: Margin compression due to vertical integration by hyperscalers like Microsoft.
Opportunity: Potential high-margin oligopoly in the memory market driven by HBM scarcity.
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 (AI) stocks have been at the epicenter of the stock market's growth in recent years. Nvidia initially led the way thanks to the soaring demand for its powerful AI chips, but plenty of new opportunities have emerged since the trend kicked off. Makers of smaller components of chips and other players in AI infrastructure have presented compelling opportunities for investors, and some companies have already translated AI into meaningful growth for their businesses.
Even at this stage of the AI trend, there are some stocks tied to it that look poised to beat the S&P 500 by a wide margin. Indeed, I think these three could double by the end of 2027.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
Neocloud company Iren (NASDAQ: IREN) has been left behind by the stock market so far in 2026. Its 5% year-to-date return pales in comparison to the performance of peer Nebius, which has almost tripled year to date. They are similar companies in the neocloud industry, but Nebius has signed more long-term computing capacity deals with hyperscalers.
Iren has had less dealmaking activity, and its recent shareholder dilution hasn't helped matters. However, management is starting to turn its existing contracts into revenue while waiting patiently for opportunities to secure new deals with higher annual recurring revenues per megawatt.
For instance, this month, Iren delivered Horizon 1, a direct-to-chip liquid-cooled AI cloud deployment, to Microsoft. That lease will bring in roughly $500 million in annual recurring revenue for the neocloud. Once the remaining three Horizon sites are delivered, Iren will be looking at almost $2 billion in annual recurring revenue from that hyperscaler alone.
Then, the company will finally have the revenue growth profile of a neocloud play rather than that of a fading crypto miner. Nebius has already made the transition to realizing revenue from some of its contracts, which is where Iren has fallen behind.
Next year, the company will start to realize even more revenue from its customers. And the sponsorship deal it inked with the Golden State Warriors professional basketball team should put it on the map with smaller AI enterprises and developers, which is the customer base Iren would prefer to pursue. The same bullish catalysts that fueled Nebius' outperformance this year should show up for Iren in 2027 and beyond.
Sandisk (NASDAQ: SNDK) may seem like a strange stock to recommend since it is already up by more than 500% year to date. It's also up by roughly 4,000% over the past year, so predicting that the NAND flash memory maker will double yet again before the end of 2027 may sound excessive.
However, Sandisk truly has the fundamentals to back up another rally, especially as demand for memory chips remains massive. Elon Musk said that the shortage of memory chips is the biggest bottleneck constraining the pace of the AI build-out, which implies they are a bigger deal than AI processors right now.
Nvidia is still reporting parabolic demand growth for its chips, but Sandisk is simply doing better, by a lot. It delivered 372% year-over-year revenue growth in its fiscal 2026 fourth quarter, with sales up by 51% sequentially. Net income surged by 91% sequentially, producing a 77% net profit margin.
Momentum in the memory space doesn't look likely to slow down anytime soon. The midpoint of management's guidance is for $10.55 billion in fiscal 2027 first-quarter revenue, which would be an 18% sequential growth rate. Sandisk's financial numbers and guidance have been far more impressive than Nvidia's, and it's even outpacing memory peer Micron on revenue and net income growth.
Management's investor day presentation highlighted numerous multiyear sales deals that extend to its fiscal 2030, offering meaningful revenue visibility and high growth for an extended period of time. Its price-to-earnings ratio of 22 serves as the icing on the cake, since that's a lower valuation than most tech stocks.
While Iren and Sandisk provide key pieces of AI infrastructure, Zeta Global (NYSE: ZETA) is a software player that's actually translating its AI investments into revenue. The company operates an AI-powered marketing cloud platform that generates high annual recurring revenue from marketers and businesses that rely on its software.
The company has invested heavily into agentic AI, and those efforts have helped Zeta achieve an impressive track record of outperforming expectations: In Q2, it delivered its 20th consecutive beat-and-raise quarter. Revenue surged by 44% year over year in the second quarter, which prompted management to boost its full-year guidance.
CEO David A. Steinberg told investors that the company is "still in the early stages" of what its platform can do for businesses as AI continues to evolve. Increasing AI adoption was cited as a major reason for the beat-and-raise quarter.
Zeta Global already works with more than half of Fortune 500 companies, and as its platform adds more features, its annual recurring revenue from these customers should continue to grow. The company has a solid revenue foundation to build upon as it gains more market share in the agentic AI industry.
Before you buy stock in Iren, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Iren wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $432,621! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,335,314!
Now, it’s worth noting Stock Advisor’s total average return is 973% — a market-crushing outperformance compared to 213% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of August 21, 2026. *
Marc Guberti has positions in Iren. The Motley Fool has positions in and recommends Micron Technology, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.
Four leading AI models discuss this article
"The transition from crypto mining to AI cloud for IREN does not mitigate the long-term risk of share dilution and the high cost of maintaining hyperscale-grade infrastructure."
The article conflates infrastructure build-out with sustainable equity value, specifically ignoring the massive capital expenditure (CapEx) risk inherent in IREN and the cyclicality of memory for SNDK. While IREN’s pivot from crypto mining to hyperscale cloud is a narrative win, the dilution risk remains a structural headwind that could cap upside despite the Microsoft contract revenue. SNDK’s 4,000% trailing return is a red flag for a 'double by 2027' thesis; memory is notoriously commoditized, and a 77% net margin is likely unsustainable as supply catches up to demand. ZETA is the only play here with a defensible software moat, but it trades on high expectations that leave zero room for a single earnings miss.
If memory remains a structural bottleneck through 2027 as Musk suggests, SNDK could maintain pricing power far longer than historical commodity cycles suggest, potentially justifying its premium.
"Sandisk's 4,000% move already embeds the bull case; Iren and Zeta are binary on execution; none warrant a 'double by 2027' blanket call without valuation anchors or downside scenarios."
The article conflates three fundamentally different theses into one 'AI doubling' narrative, which obscures real risk. Sandisk's 4,000% YTD move already prices in extraordinary growth—a 22x P/E on a memory chip maker is cheap only if you believe sequential 51% growth sustains for 18+ months, which is historically rare. Iren's $2B ARR thesis depends entirely on three unbuilt sites materializing on schedule and Microsoft not renegotiating terms. Zeta's 20 consecutive beats are impressive but don't address whether agentic AI adoption actually drives incremental spend or cannibalizes existing marketing budgets. The article provides no valuation framework—just 'could double by 2027'—which is unfalsifiable marketing copy.
If Sandisk's multiyear contracts extend to fiscal 2030 with 18% sequential growth locked in, and memory truly is the constraint Musk claims, the stock could easily compound 15-20% annually without doubling. Iren's Horizon ramp is real and de-risks the thesis materially if Q1 2027 shows $500M ARR materializing.
"Doubling by 2027 requires unrealistically strong revenue visibility and valuation multiple expansion given execution, cycle risks, and macro uncertainty."
The piece leans on hype rather than solid fundamentals. Iren’s Horizon deployments and a single $500m ARR contract with Microsoft could be transformative, but full visibility hinges on delivering three more Horizon sites and additional contracts—uncertainty remains. Sandisk’s momentum depends on memory-cycle strength and multiyear deals, but a 22x P/E for a commodity-like manufacturer may imply stretched assumptions. Zeta Global shows growth in AI marketing, yet profitability and sustained ROI on agentic AI at scale are unproven. Across all three, the upside rests on long-duration deals and AI hype, not guaranteed earnings visibility.
The risk isn't just execution; AI hype can fade, regulators tighten, and these names may not sustain the multiple expansions needed to double by 2027, even if near-term catalysts hit. Without verifiable, durable ARR growth, the thesis remains highly speculative.
"N/A"
[Unavailable]
"The valuation of hardware players like Sandisk is justified if the industry structure has permanently shifted from commodity cycles to a high-margin oligopoly."
Claude, your skepticism on Sandisk's 22x P/E is misplaced because you are treating it like a legacy commodity cycle. If the memory market shifts to a high-margin oligopoly driven by HBM (High Bandwidth Memory) scarcity, that multiple is actually a discount. The real risk isn't valuation; it's the 'hyperscaler leverage' trap. If Microsoft or others decide to vertically integrate their own silicon, IREN and Sandisk become vulnerable to margin compression that no amount of ARR can offset.
"Vertical integration is a margin risk, not an extinction event—but the real wild card is whether AI capex sustains long enough to lock in the multiyear contracts these theses depend on."
Gemini's vertical integration risk is real, but underweights the countervailing moat: HBM production is capital-intensive and requires years to scale. Microsoft integrating silicon doesn't eliminate SNDK/IREN demand—it shifts it. The actual risk is margin compression during the transition, not obsolescence. Claude's 22x P/E critique holds only if memory reverts to commodity pricing; if HBM scarcity persists through 2027, that multiple is defensible. The unasked question: what happens to these valuations if AI capex cycles slow before hyperscalers lock in multiyear deals?
"SNDK’s 22x P/E can be defensible if HBM scarcity persists and multiyear contracts lock in pricing, but Horizon execution, MSFT terms, and regulatory headwinds are the bigger wildcards."
Claude, I think you underplay the moat SNDK could build via HBM leadership and long-term contracts; if scarcity persists, 22x P/E isn't crazy. The bigger blind spot is execution risk around Horizon ramp and possible MSFT renegotiations—those could collapse ARR power quickly. Also, don't ignore regulatory risk to hyperscalers which could blunt demand for third-party memory suppliers. Margin compression remains real if new capacity comes online, even with pricing power.
[Unavailable]
The panelists generally express skepticism about the long-term prospects of IREN, SNDK, and ZETA, with concerns around high valuations, execution risks, and potential margin compression.
Potential high-margin oligopoly in the memory market driven by HBM scarcity.
Margin compression due to vertical integration by hyperscalers like Microsoft.