원문 보기
I know it's been a very volatile week. If people have been holding long positions in these companies for the past few weeks, it's been a tough road. But, uh, these companies are at the center of one of the most important strategic initiatives for the United States, you know, which is AI and AI infrastructure, and that runway is years away.
Uh, and they're going to report next week. So I think this pullback of how leveraged funds are definitely creating more volatility . It always, uh, creates opportunities for a long-term investor, because if someone missed out on, uh, a DRAM trade or a semiconductor, and now they can buy it, or even Korean, they can buy it for 20% less.
You know, this is the entry point. But yes, it is very difficult. I mean, my institutional clients , many of them are going through very difficult years because changes in themes, winds, or even visibility cause an instant reaction, and that's something people are not used to. The recent semiconductor sell-off hit Wall Street hard, with the PHLX semiconductor index falling more than 20% from its summer highs and entering a tech market without assets.
But if you're sweating over your technology portfolio, you're misinterpreting the entry. Funstrat's Tom Lee shows that this sharp drop is not a structural collapse. This is classic high-speed leverage compression. US margin debt has now risen by an impressive 54% year-on- year, the sixth highest surge in the past 60 years.
Statistically, five of these six historical spikes triggered a sharp six-month market consolidation. This is a healthy sliding correction shaking up long-term traders, not a fundamental failure of the age-old trend of artificial intelligence growth. Tom Lee's rationale points to a massive mandatory clearing of logs.
When highly leveraged funds face margin requirements , the liquidity that arises at the push of a button creates an aggressive discount. Given that the multi-year runway of AI infrastructure is completely intact , Lee clearly views this 20% pullback as a gift of a generation for activist investors who missed out on the initial chip rallies.
However, the biggest profits will not come from chasing the most highly touted crowded mega-caps. While the mainstream media has focused on a few popular giants, a select group of ignored players in the semiconductor market have suffered even greater defeats. Today, we use Tom Lee's forecast to analyze five overlooked chip stocks with deep discounts that are trading well past the 20% mark, giving you maximum asymmetric upside potential in the next phase of this tech supercycle.
First on the list is SanDisk, ticker SNDK, a company that has been severely punished during this leverage rally , down about 42% from its recent highs. While short-term traders were panicking, SanDisk's fundamentals tell a very different story. The company just showed a stunning sequential jump in data center revenue of over 230%.
Driven by the industry's massive transition from AI training to non- memory-intensive AI workloads. SanDisk directly addresses infrastructure issues with its next-generation Bix 10 technology , a 332-layer NAND architecture that increases bit density by 59%. Production has already begun at their facility in Kitakami, significantly reducing execution risk.
Even more compelling is SanDisk's shift from unpredictable commodity cycles to a contract infrastructure model. They have already signed five multi-year agreements with customers representing a minimum contractual revenue of $42 billion, backed by financial guarantees of $11 billion. With high-speed flash memory sampling later this year and a projected industry-wide supply shortage through 2027 , SanDisk is uniquely positioned to capture market share and unlock huge long-term growth potential for patient, forward-thinking investors.
Moving on to our second asset, Marll Technology, trading under the ticker MRVL, represents another premium asset in the AI infrastructure space that has fallen more than 30% in this sectoral capital rotation. The fundamental growth story here is just beginning, thanks to a massive $2 billion investment and a strategic partnership with Nvidia.
This integrates Marll's custom accelerators and optical solutions directly into the NVLink ecosystem, providing an incredible validation of their technical potential. To double down on this hyper-growth niche, management aggressively optimized its portfolio, selling non-core automotive assets for $2.5 billion in cash to spur the development of highly profitable chips.
The underlying indicators confirm this explosive trajectory. Data center sales now account for 76% of total revenue, and management has raised its revenue forecast for fiscal 2028 to $16.5 billion. A huge catalyst is their specialized ASIC business, which is projected to grow from $1.5 billion to over $4 billion by 2028 thanks to partnerships with tech giants like Amazon, Google, and Meta.
As autonomous agent-based AI requires continuous distributed cluster connectivity, Marll's advanced network interface cards and optical components are ideally prepared to scale structurally across the enterprise cloud matrix to ensure a dominant position in global technology networks. Looking ahead, our third contender for hypergrowth is Astera Labs, ticker symbol ALAB, an explosive AI infrastructure pioneer that has pulled back more than 37% from its recent highs.
Despite short-term setbacks from retail sentiment, this specialized equipment player has just gained huge institutional visibility through its inclusion in the NASDAQ 100 index. Astera Labs uses a unique business model aimed at eliminating critical bottlenecks in data center connectivity. With an industry-leading gross margin of 76.3%, which outperforms major semiconductor manufacturers, their specialized products, including Scorpio intelligent switches, Aries PCIe retimers, and Taurus Ethernet connectivity modules, are critical for hyperscalers deploying advanced GPU clusters with optimal latency performance.
The company's recent figures highlight this massive surge in demand, delivering first-quarter revenue of $38 million, an impressive 93% increase over the same period last year. Looking ahead, the accelerated deployment of next-generation PCIe Gen 6 hardware , which already accounts for over a third of their revenue, is serving as a major catalyst.
With massive hyperscaler capex projections reaching $5.5 trillion worldwide by 2030, Astera Labs provides exceptional operational leverage to unlock enormous long-term growth potential for specialized technology portfolios seeking to capitalize on structural changes in infrastructure across the macro technology landscape.
Shifting our focus to the automated equipment sector, Pterodine, trading as the ticker te , is a key player in the AI infrastructure chain that has faced a significant downturn, down more than 33% from its recent highs. Pterodine's value proposition is growing rapidly as AI chips become larger, more power-intensive, and significantly more difficult to test.
In addition to traditional automated test equipment , they are aggressively moving into system , optical, and high-speed interconnect testing. With a portfolio spanning the entire AI lifecycle, from wafer-level testing to final deployment in the data center, the move to advanced AI compute and memory creates a structural tailwind.
More complex chips simply require longer and more thorough testing phases to protect profitability. With profit growth poised to outpace revenue growth thanks to operating leverage, Pterodine is rebuilding its entire profit model around artificial intelligence. By capturing new, fast-growing niches such as Photonix testing, with products like Photon 100 and Omnix, as well as high-speed connections, they are solving critical bottlenecks for hyperscale companies.
While traditional metrics may suggest a premium valuation, the company's aggressive earnings growth and deep-rooted role in the AI supply chain suggest that the market may be underestimating its long-term structural value in the global chip manufacturing market during this period of massive macroeconomic expansion.
Finally, rounding out our list is Coherent, ticker CO HR, a strategic optical networking giant whose shares have fallen significantly, by about 35%, during this recent market boom. Investors here don't see the forest for the trees. Optical connectivity is no longer just a side market for AI. It has become the defining physical bottleneck that dictates how fast hyperscalers can actually run their massive compute clusters.
Coherent is uniquely diversified across the entire optical stack, monetizing everything from transceivers and thermal technologies to next- generation optical channel switching and co- packet optics. Despite the short-term mood reset, the company boasts a record order backlog that extends all the way to 2028.
To meet this huge demand, management is actively expanding its 6- inch indium phosphide production capacity . Financially, the growth trajectory is undeniable. Revenue is forecast to grow from $7.1 billion in fiscal 2026 to $13 billion by fiscal 2028, and earnings per share are expected to grow more than 50% annually.
However, trading at a forward price- to-sales multiple of just 9.5 times, Coherent is worth less than half the value of key competitors such as Lmentum, representing huge long-term growth potential for investors around the world seeking significant asymmetric value . This market-wide correction presented disciplined investors with a rare window of opportunity.
By focusing on these five significantly reduced- impact infrastructure elements, rather than chasing crowded tech giants, you can position your portfolio for maximum asymmetric growth as the multi-year AI supercycle prepares for its next explosive growth phase.
Comments 0