원문 보기
I believe that the markets are generally dominated by optimists. And, as you know, the S&P has more than doubled during this period. It was a very powerful period. And I think you should understand: when optimists rule, prices tend to be overvalued relative to intrinsic value , which means it's time for some caution.
This isn't the worst I've seen, but I think it's important to note that control is in the hands of the optimists. Warren Buffett always said, "Oh, we'll put that in the 'too hard' pile." This may be too difficult for the value investor. Definitely for someone with a capital "B". And it's difficult for someone who is from a small town.
Howard Marks's main warning is not a cynical bet against technology. This is an urgent reality check based on solid market data. The S &P 500 index has risen over 70% since the end of 2022, and stock valuations consistently trade at a large premium to internal cash flow. Meanwhile, global AI capital spending is projected to cross a staggering $1 trillion, largely due to large hyperscalers spending more than $500 billion annually on data center expansion, specialized equipment, and underlying energy infrastructure.
The market has entered a phase of unbridled optimism, where bad news is ignored and uncertainty about the implementation of plans is not factored into the price at all. For AI startups without proven profitability, calculating a five- year revenue model requires more blind faith than financial analysis.
As Marx wisely observes , such speculative bets belong squarely in Warren Buffett's "too complicated" pile. However, avoiding speculative mania does not mean staying away from the historic technological shift . The real winners in this environment aren't the high-profile, cash-burning programmatic stories.
These are battle-tested capital-generating pillars that Marx's model actively favors. In this video, we'll break down three deep-rooted market leaders that fit perfectly into this value and security model. The first is a software monopolist, backed by billions in recurring corporate revenue, that funds its AI capital expenditures without risking balance sheet stability .
The second is an infrastructure giant that provides multi-year fixed contracts for the supply of essential baseload energy that data centers so desperately need . And a hardware architect who oversees the specialized chips and network bottlenecks that hold each AI cluster together. These three companies give you direct access to AI growth, backed by a real margin of safety and unwavering intrinsic value .
First on our list is the software monopolist that embodies Howard Marx's ideal margin of safety—Microsoft Corporation, ticker MSFT. As macroeconomic uncertainty looms, Microsoft's intrinsic value remains anchored by an unmatched distribution network and deeply rooted commercial contracts. Total annual revenue rose to $331.8 billion , generating extraordinary operating cash flow of $182.9 billion.
While Wall Street is focused on increasing capital spending, Microsoft is demonstrating real operating leverage . The newly created Agents and Infrastructure division generated $268.1 billion in revenue, up 22.6% year-over-year, with the segment's operating margin increasing to 50.9%. Azure revenue continues to grow rapidly at 40%, with forecasts pointing to a level of around 45%.
Microsoft differs from purely AI startups in that its commercial success does not depend on winning speculative model races. Thanks to its huge corporate base, Microsoft is monetizing AI at the implementation and management levels. Comprehensive offerings, such as the E7 tier, integrate Copilot, security, identity, and agent management directly into existing workflows, making implementation easy during contract renewals.
Additionally , proprietary developments, such as the Maia 200 inference accelerator, reduce computing costs, allowing Microsoft to increase margins as enterprise usage grows. Despite spending more than $115.9 billion on real estate, equipment, and data center expansion to expand capacity, Microsoft generated $67 billion in excess ith a solid balance sheet containing $20.9 billion in cash, an impeccable AAA credit rating, and a stable dividend payout ratio of 52%.
The company provides private investors with unparalleled protection against market volatility. Trading at a forward P/E multiple of 25, Microsoft is turning its massive AI investment into sustainable, high-margin growth without risk to shareholders' balance sheets . Moving from digital ecosystems to the physical grid, our second action is a clean energy monopoly that provides critical electricity for the AI revolution.
Constellation Energy Corporation, ticker CEG. Howard Marks has repeatedly emphasized that real-world implementation of AI requires physical reality, network stability, and reliable infrastructure. High-performance data centers operate nonstop , requiring 24/7 carbon-free energy that renewable sources alone cannot provide .
Constellation has unrivaled competitive protection as the largest nuclear power plant operator in the U.S., managing approximately 55 gigawatts of capacity following its strategic acquisition of Calpine. Recreating a nuclear fleet of this scale today would require three times more capital than Constellation's current value, creating an almost insurmountable barrier for competitors.
Constellation's business model provides predictability of cash flows, which perfectly fits Marx's concept of "safety margin." About 30% of nuclear power generation is secured by long-term power purchase agreements, the most significant of which is a 20- year contract with Microsoft to restart the Crane Clean Energy center.
In addition to private corporate agreements, public policy provides a reliable minimum level of revenue through tax incentives for nuclear power production. These inflation- linked tax credits provide up to $15 per megawatt- hour until at least 2032, protecting revenues from falling commodity prices and adding direct revenue if inflation remains high.
Financially, Constellation is demonstrating extraordinary dynamics : quarterly revenue amounted to $7.5 billion, and net profit doubled to $2.1 billion in the first half of the year. Operating cash flow for the same six-month period exceeded $1.55 billion , allowing us to confidently execute on capital allocation plans, including a new $5 billion share repurchase program.
Management is targeting over 20% CAGR in core earnings per share through 2029, supported by long-term contract demand and political guarantees. Trading at a forward price-to- earnings multiple of around 23.5, Constellation offers investors an irreplaceable monopoly in clean energy, directly tied to the expansion of hyperscale AI.
Moving from power generation to the hardware architectures that handle these massive workloads, our third stock is specialty semiconductor pioneer Broadcom Inc., ticker AVGO. While consumer-focused AI startups compete for attention, Broadcom supplies the critical hardware architecture that underpins the entire ecosystem.
The company has entered into development partnerships with six leading model developers, including long-term agreements to manufacture custom tensor processors for Google, as well as multi-gigawatt computing power for Meta, Anthropic, and OpenAI. In addition to dedicated accelerators, Broadcom has a dominant market share in high-speed networking chips, providing the necessary Tomahawk 5 and Jericho 3X switching platforms to prevent bottlenecks in huge data centers.
From a financial perspective, Broadcom demonstrates exactly the kind of safety margin and operating discipline that rational investors demand. Annual revenue exceeded $100 billion thanks to an impressive 86% year-over-year growth. While the increase in memory in specialty chips has led to a slight erosion of gross margins, Broadcom's huge operating leverage has increased net profitability to 44%, with management targeting 50% in the next 2 years.
Operating cash flow remains extremely strong : free cash flow margin exceeds 45%, and quarterly free cash flow reaches almost $14 billion. This huge cash generation easily funds aggressive capital spending, such as expanding packaging capacity at the Singapore plant and scaling up laser production, leaving room for shareholder returns.
Purchasing non- AI enterprise software, including VMware, provides a stable, highly profitable revenue base that insulates the business from hardware cyclicality. Broadcom trades at approximately a forward multiple of 20.7, creating an attractive discount to the semiconductor industry median of 22.1.
Focusing on core specialty chips and network backbones, backed by a long history of disciplined capital allocation. Broadcom offers investors direct access to the multi-trillion dollar AI market with an unmatched margin of safety. Whether you choose to invest in software, energy, or specialized hardware, these three pillars prove that you don’t have to sacrifice financial security to profit from the AI revolution.
Be sure to like, share your thoughts in the comments, and subscribe for in-depth analysis of the institutional market .
Comments 0