Moat: Aramco sits on the largest proven oil reserves on Earth. There is no substitute, no competitor, and no startup disrupting this. As long as the world runs on hydrocarbons (and despite what Twitter will tell you, it still very much does), Aramco prints money.

Big pivot: Aramco isn't just an oil company anymore. They're spending $50–55 billion in capex this year to build out a massive natural gas and petrochemicals business. Three major projects (Shaheen in South Korea, Amiral in Saudi Arabia, Hapco in China) are targeting a combined 4 million barrels/day liquid-to-chemicals capacity by 2030. So, they're turning oil into plastics, fertilizers, and industrial materials at a scale nobody else can touch.

Stock buyback: Aramco just launched a $2–3 billion share buyback in March. For a stock that already yields north of 5%, that's a signal that management thinks their own stock is cheap.

Biggest risk: It's a Saudi state-controlled company. Geopolitics is the risk. The current oil environment (Brent above $100) is great for Aramco, but the stock trades at the mercy of OPEC+ decisions, not earnings calls.

AI printing press: TSMC's AI chip revenue is projected to grow at a 60% compound annual growth rate through 2029. Every time a hyperscaler announces another $50 billion in AI capex, a meaningful chunk of that spend flows directly to TSMC's order book.

The 2nm edge: TSMC's next-generation N2 process entered volume production in Q4 2025. This matters because every new chip generation means higher revenue per wafer. Their customers literally cannot go anywhere else for this technology.

The numbers: Q4 profit hit NT$1 trillion (a record), January 2026 revenue surged 36.8% YoY, and the company guided for ~30% full-year revenue growth. At a forward P/E of ~23x, it's not cheap, but it's not absurd for a monopoly growing 30% a year.

Biggest risk: Taiwan. That's it. That's the risk. The geopolitical tension around Taiwan is the single biggest overhang on what is otherwise the most important company in the semiconductor supply chain. TSMC is building fabs in Arizona and Japan to diversify, but the majority of cutting-edge production still sits on an island 100 miles from mainland China.

AI revenue is exploding: Broadcom reported $8.4 billion in AI revenue last quarter. That's 106% year-over-year growth and 43% of total company revenue. Q2 guidance? $10.7 billion in chip revenue, which would represent 140% YoY growth. The acceleration is getting faster, not slower.

A $100 billion target: CEO Hock Tan dropped a bomb on the last earnings call: Broadcom has "line of sight to achieve AI revenue from chips, just chips, in excess of $100 billion in 2027." That figure is backed by a $73 billion backlog with secured TSMC production capacity at 3nm and 2nm through the end of the decade.

Custom silicon moat: Here's what makes Broadcom different from every other chip company: their biggest customers (Google, Meta, ByteDance) don't want to buy Nvidia's one-size-fits-all GPUs for every workload. They want chips custom-designed for their specific AI models. Broadcom is the company that builds those chips. And once you've spent two years co-designing a custom accelerator with Broadcom, you don't switch vendors. That's a moat measured in years, not quarters.

Biggest risk: Valuation. At a trailing P/E of ~60x, a lot of growth is already priced in. If AI spending decelerates, or if even one major hyperscale customer pulls back, the multiple compresses fast. Broadcom is a high-conviction bet that AI infrastructure spending doesn't just continue, but accelerates.

Saudi Aramco is the energy floor. It’s the company that keeps the world running on the fuel it actually uses today, while quietly building a petrochemicals empire for tomorrow.

TSMC is the manufacturing monopoly, the single point of failure (and single point of profit) for every advanced chip on Earth.

Broadcom is the AI infrastructure play hiding behind a boring name. It has the custom silicon, networking, and software that every hyperscaler needs but Nvidia doesn't provide.