저자: TomLee · 외국인
요약 보기
10년물 금리 5.2% 상승 환경에서 재무가 탄탄한 대형 기업이 경쟁 우위를 얻는다고 주장. 9월 30일 PCE 방법론 변경으로 연간 PCE가 3.4%→3.0%로 낮아지고 금리 인하 여지가 생긴다는 전망. 이를 근거로 마벨(MRVL), AMD, 아마존(AMZN)을 AI 인프라 수혜주로 제시하며 장기 보유 관점을 제시함.
다듬은 글
Yields have risen to a level where they compete with stocks. We're still trying , I think the market is still trying to figure out why yields are higher. I mean, there are several reasons. But I think two things are getting lost in this conversation . The first is where the rates will be in 6 months, because where will inflation be?
And second, I think that increasing profitability improves competition for stronger companies. This means that there will be companies that become even more attractive as profitability increases. And I think, as you point out, Scott, that's why the Magic Seven is growing. You know, when yields are rising, their ability to finance remains easy, but it's harder for competitors.
Fundstrat founder Tom Lee just presented a groundbreaking macroeconomic analysis on CNBC that completely changes the way institutional investors navigate the current interest rate environment. As the yield on the 10-year U.S. Treasury bond topped 5.2% and the Federal Reserve raised its benchmark rate to 4.0% after last raising it from 3.75%, retail traders are panicking over high borrowing costs.
However, Lee emphasizes that the 5.2% yield environment acts as a deadly competitive filter for corporate America. High interest rates punish debt-dependent competitors while widening the competitive moat for financially strong companies that generate organic free cash flow. In addition, Li emphasized that the next 6 months will be a crucial turning point for structural inflation.
Importantly, on September 30, the Bureau of Economic Analysis will release an updated methodology for calculating personal consumption expenditures, which is projected to immediately lower annual figures by 20–40 basis points. Lowering the overall PCE index from 3.4% to 3.0%. Combined with a reduction in flash memory price spikes and a stabilization of crude oil prices near $100 per barrel, inflation will fall sharply, paving the way for the Federal Reserve to ease its tight policy.
Rather than waiting for the Central Bank to cut rates , institutional “smart money” is actively accumulating capital-intensive infrastructure leaders that control the physical bottlenecks of the AI supercycle . Today we analyze three high-confidence market strategies. A custom chip titan with a $135 billion infrastructure order book, an AI accelerator pioneer with 107% growth in data centers, and a cloud giant generating $16.6 billion in quarterly AWS operating income.
Moving on to our first market strategy, Marvell Technology Inc., trading under the ticker MRVL, perfectly demonstrates how capital-intensive semiconductor titans are turning a 5.2% Treasury yield environment into an unstoppable competitive moat. While debt-laden chipmakers struggle to service their obligations at a base rate of 4.0%, Marvell acts as the necessary nervous system of the AI ecosystem , generating enormous operating leverage.
Data center operations now account for a staggering 74% of Marvell's total revenue, insulating the company from broader macroeconomic fluctuations. In its latest quarterly financial report, Marvell reported $2.739 billion in total revenue. But the real highlight for investors is the explosive growth in profitability: EPS ( non-GAAP) grew 50% year-over- year.
As compute clusters scale from thousands of GPUs to massive megawatt AI factories, standard network architectures face severe packet collisions and latency. Marvell is addressing this connectivity crisis through multi-year development partnerships with Amazon Web Services and a landmark $2 billion systems integration deal with Nvidia .
These strategic alliances bring Marvell's proprietary compute accelerators and high-speed silicon photonics directly into forecasts revenue growth to $11 billion in fiscal 2027 and $15 billion by 2028, driven by a massive $135 billion infrastructure order backlog across the broader technology landscape.
As enterprise AI adoption moves from static model training to high-volume agent-based inference, token processing requirements will grow exponentially, increasing demand for Marvell electro-optics. Importantly, the market for specialized computing and optical components is projected to grow by almost 90 % annually through 2028.
With high gross margins (non- GAAP) and organic cash flow growth, Marvell's strong balance sheet fully offsets debt service risks in a prolonged high interest rate environment, allowing management to aggressively capture market share while competitors with reduced budgets reduce research spending. With an attractive valuation relative to long-term earnings acceleration, Marvell represents an elite institutional asset for forward-looking technology portfolios .
Combining specialized silicon manufacturing capabilities with next-generation AI accelerators brings us to our second solid investment —MD. While smaller chipmakers are squeezed by 5.2% Treasury yields to fund capital-intensive research, AMD has the impeccable balance sheet and cash flow scale needed to challenge entrenched hardware monopolies.
The final confirmation of AMD's architectural path was a landmark agreement with OpenAI to jointly develop specialized silicon accelerators optimized for advanced models. The foundation of AMD's hardware expansion is the MI300 and MI400 GPU series, specifically designed for enterprise workloads with high inference throughput.
To capture the costs of end-to-end data center architecture, AMD is implementing its Helios rack scale AI infrastructure, integrating 72 high-performance processors into unified liquid- cooled server racks. These integrated Helios racks deliver premium gross margins of over 75%, allowing AMD to offer turnkey solutions for hyperscalers rushing to deploy megawatt compute clusters.
In its latest financial report, AMD's data center division showed explosive revenue growth of 107% year- over-year, reaching $6.7 billion. This proves that global cloud providers are aggressively diversifying, moving away from ecosystems with limited supply from a single vendor. This massive shift in demand confirms AMD's role as an indispensable co-architect of the next-generation AI infrastructure alongside leading-edge developers.
As the broader technology market moves toward agent workloads that require continuous token processing, ROCm's open software ecosystem and AMD's high-density memory architectures deliver higher compute efficiency per watt. Importantly, AMD operates with virtually no net debt and growing organic cash flow, allowing management to aggressively finance the release of next- generation chips without the costly debt service at the central bank's 4.0 % interest rate.
While short-sighted retail investors focus on temporary fluctuations in the PC market, AMD's corporate order book continues to grow rapidly, supported by hyperscaler commitments to cloud services. relative to a projected compound annual earnings growth rate of 45%, AMD offers an exceptional asymmetric entry point for long-term institutional investors seeking multi-year capital appreciation.
The move from silicon chip design to global cloud infrastructure brings us to the third dominant player—Amazon.com Inc., trading under the ticker AMZN. Amazon perfectly demonstrates Tom Lee's core thesis of a "capital fortress" that uses the 5.2% Treasury yield environment to crowd out weaker competitors.
While debt service costs punish cash-strapped technology vendors, Amazon operates a financial titan capable of financing multibillion-dollar infrastructure construction solely from internal operating cash flow. According to the results of the last quarter, Amazon received $200.6 billion in total corporate revenue.
But the real institutional profit driver is Amazon Web Services, which grew 36.7% year-over-year to $42.2 billion. AWS generated a staggering $16.6 billion in quarterly operating income with an operating margin of 39% , operating as an unparalleled high-margin cash engine. Amazon has solidified its dominant position at the heart of enterprise AI workloads by investing $13 billion in direct strategic capital in Anthropic with a structured commercial trajectory to $33 billion in total commitments.
Instead, Anthropic has contractually committed to spending over $100 billion over the next decade on AWS cloud infrastructure and its own hardware. Importantly, Anthropic’s large-scale implementation of Amazon’s Trainium specialized silicon validates AWS as a lower-cost vertically integrated alternative to traditional off-the-shelf GPU architectures.
As corporate spending shifts from static model training to agent-based inference of large volumes of data. AWS's proprietary chip architecture reduces the total cost of ownership for enterprise customers by 30%, while significantly increasing Amazon's gross margin. Additionally, Amazon's retail logistics network continues to generate growing operating leverage through the introduction of automated robotics, adding another highly profitable pillar to corporate revenues.
With a spotless balance sheet with tens of billions in liquid cash reserves, Amazon remains completely immune to the central bank's 4.0% interest rates , capturing the lion's share of the global expansion of cloud technology. Trading at an attractive enterprise value multiple relative to growing operating cash flows, Amazon is an indispensable anchor asset for long-term investment portfolios.
Tom Lee's macroeconomic thesis is quite understandable. Treasury yields of 5.2% are isolating elite capital strongholds, while a decline in PCE inflation from 3.4% to 3.0% is setting the stage for a massive stock rally. Marvell Technology, Advanced Micro Devices, and Amazon are unstoppable monopolies with powerful balance sheets that control the markets for silicon, networking technology, and cloud nodes. Which stocks will you buy first? Share your thoughts below.
Comments 0