Summary
Highlights
The video opens by questioning the most complex engineering process in human history and discusses investment opportunities and risks associated with it. The creator emphasizes using his own money for investments shown and warns that this episode covers aggressive and risky investments. He then outlines his learning process, which involves extensive research and questioning, and introduces the main topic: the semiconductor manufacturing chain.
The first step in semiconductor manufacturing is architecture, dominated by ARM. ARM designs and licenses instruction sets that define how a processor processes data. These instructions are optimized for energy efficiency and minimal heat generation. ARM's strong competitive moat stems from the fact that most other solutions are optimized for or based on their technology.
In the second step, companies like Nvidia, AMD, and Amazon purchase ARM's architectural plans. Engineers then program the placement of billions of transistors, computational cores, signal pathways, and memory controllers on these plans. This precise design process can take months, and at these nanoscale levels, quantum effects like quantum tunneling can occur.
The third step involves fabrication by monopolists like TSMC. TSMC receives the designs and uses lithographic machines from ASML to etch circuits onto silicon wafers. This multi-layered process, from raw wafer to finished processor, can take up to three months. The production environment must be thousands of times cleaner than an operating room to prevent contamination.
Step four involves advanced packaging where GPUs (from companies like Nvidia) are integrated with HBM memory from SK Hynix or Micron. Step five extends to assembling these components into data centers. Eight such GPU+memory units are connected via switches (Broadcom) on a motherboard. CPUs (Nvidia, AMD, Intel) and SSDs (Samsung, Micron) are added. These units are then combined into server racks, which require sophisticated cooling systems (Vertif) and high-speed networking cables (Lumentum, Coherent), network switches (Nvidia, Broadcom, Arista Networks). Data centers also require immense, uninterrupted power (Eton, Schneider Electric, Constellation Energy, small modular reactors), and specialized real estate (Equinix, Digital Realty).
The speaker notes that while many companies are sponsoring these developments, the long-term profitability is debatable. He explains that professional investors are already aware of these trends, leading to significant growth in the mentioned companies. He invests in this sector despite high valuations, betting that the high demand for data centers and AI will last longer than the two-year projections, expecting continued positive surprises in quarterly reports.
The speaker discusses using a price-to-forecasted-revenue ratio alongside revenue growth forecasts to identify undervalued companies. He highlights Broadcom, which has a historically low P/S ratio but incredibly high (48.6% CAGR) and consistently upgraded revenue growth forecasts. This suggests investors expect high returns, but potentially a future slowdown in growth. He also reviews other companies like Micron, SanDisk, Nvidia, Arista Networks, Vertif, Monolithic Power Systems, TSMC, ASML, and AMD, noting their valuations and growth prospects.
The creator discusses adjustments to his Global Equity Momentum (GEM) strategy after listening to an interview with its creator, Garm Antonaci. He clarifies that his previous method of checking index strength over 12 months, excluding the last month, is more suitable for individual stocks than entire indexes. He reverts to the original GEM principle of simply looking at the last 12 months for four ETFs: two equity (Nasdaq, emerging markets) and two bond (long-term, short-term). He monitors which ETF shows the strongest trend and invests accordingly.
Currently, emerging markets ETFs are outperforming, and he recently purchased one. He notes that Nasdaq is catching up and anticipates a potential shift. He hopes for a quicker end to the conflict in the Middle East, which would boost emerging markets. The strategy involves switching ETFs only when the leading ETF's advantage exceeds three percentage points. He also discusses the tax implications of frequent trades on a regular investment account versus tax-advantaged retirement accounts like IKE and Oki, where he plans to direct future GEM contributions.