What We’re Reading (Week Ending 29 March 2026) - 29 Mar 2026
Reading helps us learn about the world and it is a really important aspect of investing. The late Charlie Munger even went so far as to say that “I don’t think you can get to be a really good investor over a broad range without doing a massive amount of reading.” We (the co-founders of Compounder Fund) read widely across a range of topics, including investing, business, technology, and the world in general. We want to regularly share the best articles we’ve come across recently. Here they are (for the week ending 29 March 2026):
1. What the closure of the Strait of Hormuz means for the global economy – Lutz Kilian, Michael Plante and Alexander W. Richter
From the point of view of the rest of the world, a disruption of oil exports from the Persian Gulf is equivalent to a disruption of oil production in the Gulf. From the point of view of oil producers in the Gulf, the difference is largely academic because as soon as local oil storage fills up, oil producers have no choice but to shut in their oil wells if the oil cannot be stored or exported. This is why many oil producers, starting with Iraq and Kuwait, started curtailing their production in early March 2026.
A complete cessation of oil exports from the Gulf region amounts to removing close to 20 percent of global oil supplies from the market, about 80 percent of which is shipped to Asia. Oil importers unable to access oil from the Persian Gulf have to turn to other oil suppliers, putting upward pressure on oil prices worldwide…
…Major oil supply shortfalls driven by geopolitical events such as wars or revolutions previously occurred following the Yom Kippur War in 1973, the Iranian Revolution in 1979, the outbreak of the Iraq–Iran War in 1980 and the Persian Gulf War in 1990. What makes the closure of the Strait of Hormuz different from these earlier oil supply shortfalls is first and foremost its magnitude. For example, in 1973 and 1990 only a little more than 6 percent of global oil supplies was removed from the market and in 1979 and 1980 only about 4 percent. Today, we are concerned with a shortfall close to 20 percent, making this geopolitical event three to five times larger.
This is the first time the Strait has been closed. While some observers in 1990 grew concerned that Iraq would take over Saudi Arabia and control of the Persian Gulf, these concerns never materialized…
…Regardless of the likelihood of the Strait reopening in the future, the model implies that a closure of the Strait of Hormuz that removes close to 20 percent of global oil supplies from the market during second quarter 2026 is expected to raise the average West Texas Intermediate (WTI) price of oil to $98 per barrel and lower global real GDP growth by an annualized 2.9 percentage points in second quarter 2026 (Chart 2).
The subsequent effects depend on when oil shipments resume. For example, if the Strait reopens after one quarter, the oil price drops to $68 per barrel and growth increases 2.2 percentage points in third quarter 2026. While the oil price drop causes growth to recover, the level of real GDP remains 0.2 percent below its pre-closure level even by year-end 2026 and 0.1 percent below its initial level by year-end 2027. The positive growth response in third quarter 2026 reflects the increased availability of oil and the resulting decline in the price of oil.
When the oil supply shortfall lasts longer than one quarter, richer dynamics arise. Extending the closure to two quarters causes the oil price to rise further to $115 per barrel in third quarter 2026 before falling to $76 per barrel in fourth quarter 2026 (Tables 1, 2). The impact on real GDP growth only turns positive in fourth quarter 2026. If shipping resumes after three quarters, the oil price will rise even further before declining, reaching as high as $132 per barrel by year-end. The impact on growth will remain negative through year-end 2026…
…While the model underlying these scenarios is global, the case can be made that the effects of higher oil prices on U.S. GDP growth will be of similar magnitude to the global effects. Although the U.S. economy for many decades was heavily dependent on imported petroleum, since the shale oil boom the U.S. petroleum trade balance has been close to balanced. This makes the U.S. economy not so different from a global economy model in which there is no trade in oil by construction…
…One way the oil supply shortfall could potentially be reduced is by Saudi Arabia increasing the flow of oil on the East-West pipeline from the Persian Gulf to the Red Sea. The capacity of the Yanbu port would allow Saudi Arabia to redirect about 4 million barrels of oil per day from the Persian Gulf for transport by oil tankers from the Red Sea, corresponding to about one-fifth of the global supply shortfall.
One obvious concern with this approach is the port in question is within range of both Iranian and Houthi missiles from Yemen, as are the waterways in the Red Sea. The other concern is that shipping this oil south past the Bab el-Mandeb Strait to Asia exposes oil tankers to attacks by the Houthis, while shipping it north through the Suez Canal limits the tanker size and requires redirecting the oil toward Europe rather than Asia where it is most needed.
There is also a short pipeline in the United Arab Emirates bypassing the Strait of Hormuz to the port of Fujairah on the Gulf of Oman. That pipeline as well as the port, however, have already come under Iranian attack, making it difficult for the existing flows to be maintained, never mind increased.
2. Warren Buffett Case Study – Dirtcheapstocks
J. Paul Getty was the richest man in America in 1957.
Five years later, you could buy a piece of his oil company for 63 cents on the dollar.
Warren Buffett took notice…
…Buffett’s Getty shares were marked at $18 at year end…
…On the surface Getty didn’t look especially cheap. Sure, it traded at a large discount to book, but the ROE was low and the stock was selling for 17x earnings.
But there was a larger issue at play.
Getty owned large stakes in three publicly traded, related party companies: Mission Corporation, Mission Development Corp., and Tidewater.
Without going into too much detail, the nature of these businesses was to produce, refine and market oil and manage other assets of the “Getty Empire”…
…Getty’s market cap was only $287mm.
Getty’s share of these three assets alone was $259mm.
Buyers at $18 were paying almost nothing for Getty Oil.
And it’s not like Getty had a bad business. It was earning $14mm of net income and had a pristine balance sheet. This excludes its share of income from Mission, Mission Development and Tidewater.
Getty had $398mm of total assets and only $50mm of total liabilities…
…Buffett was actually paying 1.7x earnings and 30% of book for Getty.
On a look through earnings basis (Getty earnings plus share of minority-owned earnings), Buffett paid ~7.5x for his Getty investment. It was cheap any way you slice it.
Getty had a steady history of growth…
…That’s an 11% CAGR over 11 years in BVPS.
Companies compounding book value at double digits should not trade at a discount to book…
…In 1949, J. Paul Getty ignored his advisors and bought a barren strip of desert in Saudi Arabia.
That piece of land produced 15,000 barrels per day in 1956.
By 1962 it was up to 100,000 barrels per day.
Getty’s agreement with Saudi Arabia called for a fixed royalty structure, allowing Getty to capture the vast majority of the field’s value.
In 1962, Getty produced 10x the oil volume of a decade prior…
…I don’t know how long Buffett held his shares, but he probably made money.
Shares traded up to $27.50 in 1963 and $32 in 1964.
Getty Oil was bought by Texaco in 1984 for $10.1 billion.
Adjusting for splits, the shares Buffett owned at $18 in 1962, would have grown to $625 in 1984.
Excluding dividends, the stock compounded at 17.5% for 22 years.
3. Meta’s Agentic AI Ambitions – Abdullah Al-Rezwan
One of the interesting bits from the blog post is that Meta mentioned for long-horizon workflow autonomy, Meta built REA on an internal AI agent framework called “Confucius” which they elaborated further on this paper back in February 2026. Often, when tech companies try to improve AI coders, they focus on making the underlying AI models (like GPT or Claude) smarter. However, the paper argued that the “scaffolding” i.e. the software environment, memory systems, and tools built around the AI is just as important. When working on big codebases, AI agents frequently get overwhelmed by reading too much code, forget their original plan during long tasks, or repeat the same mistakes.
The most interesting takeaway from the paper is that a great setup can compensate for a less powerful AI. The researchers proved that a weaker model (Claude 4.5 Sonnet) using the Confucius scaffolding successfully fixed more bugs (52.7%) than a stronger, more expensive model (Claude 4.5 Opus) using Anthropic’s standard setup (52.0%). When powered by the GPT-5.2 model, Confucius Code Agent successfully resolved 59% of the real-world bugs on the SWE-Bench-Pro test, beating both prior academic research and the official corporate systems built by OpenAI and Anthropic under identical conditions. If such scaffolding itself can consistently beat the more expensive SOTA models, it can provide a ceiling on SOTA model developers’ ability to exercise pricing power. It remains to be seen whether such scaffolding can outperform more expensive SOTA models in a wide range of scenarios. Nonetheless, the key takeaway is quite encouraging for all the tech companies that will not have a SOTA model and those tech companies may still be able to capture value from better scaffolding.
4. The AI Supply Chain Runs Through a War Zone. Nobody in Silicon Valley Is Paying Attention – Veron Wickramasinghe
The physical supply chain that powers every artificial intelligence system on earth passes through a single chokepoint that has been effectively closed since early March. Not a data bottleneck. Not a software constraint. A 21-mile strait between Iran and Oman through which 30 percent of the world’s LNG and 20 percent of its oil once flowed…
…Helium is the second most abundant element in the universe and one of the rarest on Earth’s surface. It is produced by the radioactive decay of uranium and thorium deep in the planet’s crust. It migrates upward through rock over billions of years and accumulates in the same geological traps that hold natural gas. You do not manufacture helium. You extract it as a byproduct of natural gas processing, or you do not have it.
Qatar’s three helium plants at Ras Laffan produce approximately 2.3 billion standard cubic feet per year: Helium 1 (660 million scf, online 2005), Helium 2 (1.3 billion scf, the world’s largest, online 2013), and Helium 3 (400 million scf, online approximately 2021). That is roughly one-third of total global helium supply, according to the US Geological Survey’s 2026 Mineral Commodity Summaries, which puts Qatar at 33.2 percent of world production.
All three plants have been offline since March 2, when Qatar halted LNG production following the outbreak of hostilities. The helium plants cannot operate independently of the LNG facility because helium is extracted from the natural gas stream during cryogenic liquefaction. When the gas stops flowing, the helium stops flowing.
QatarEnergy CEO Saad al-Kaabi confirmed on March 24 that the missile strikes reduced helium output capacity by 14 percent, with repairs expected to take three to five years. The planned Helium 4 plant, targeting 1.5 billion standard cubic feet per year and over 50 percent engineered before the crisis, has no confirmed restart timeline…
…The bottom line: helium is genuinely critical for specific, high-value fabrication steps, particularly plasma etching, where no substitute exists. It is not equally irreplaceable across all semiconductor applications. But the applications where it is irreplaceable happen to be the ones that define whether a chip gets made or does not…
…South Korea imports 64.7 percent of its helium from Qatar, according to Korea International Trade Association data for 2025.
South Korea is home to Samsung Electronics and SK Hynix, which together dominate global memory production. SK Hynix commands 62 percent of the High Bandwidth Memory market by shipment volume as of Q2 2025, per Counterpoint Research. Samsung holds 33 percent of global DRAM market share. Combined, these two companies produce the majority of the memory chips that go into every AI training system, every data centre GPU, and every high-performance computing cluster on earth.
HBM is the single most critical constraint in the AI hardware supply chain…
…South Korea imports approximately 70 percent of its crude oil from the Middle East. The Strait of Hormuz has been effectively closed to commercial shipping since early March, when war risk insurance premiums made transit economically unviable. Seoul implemented mandatory fuel rationing on March 25: a one-day-per-week vehicle ban for 1.5 million government vehicles, enforced by licence plate number.
QatarEnergy declared force majeure on long-term LNG contracts with South Korea on March 24. Gas generates approximately 26 percent of South Korea’s electricity. Those contracted molecules, which were supposed to flow reliably for decades, now carry a force majeure notice that could last five years.
South Korea is losing three supply lines simultaneously. Oil. Gas. Helium. All from the same chokepoint…
…SK Hynix has publicly stated it has diversified supplies and secured sufficient inventory. Samsung has not issued a public reassurance but is understood to hold approximately six months of stockpile and has deployed its Helium Reuse System, which reduces consumption by approximately 18 percent. TSMC says it does not currently anticipate notable impact and maintains helium from multiple suppliers with over two months of stock on hand. The Korea Semiconductor Industry Association says short-term supplies are sufficient.
There are reasons to take these reassurances seriously. Major fabs are not naive about supply chain risk. Over 70 percent of fabs in Taiwan and Japan already operate helium recycling systems. Six months of Korean stockpile buys time…
…The United States produces 42 percent of global helium but cannot rapidly scale. The former Federal Helium Reserve in Amarillo was privatised in June 2024 and can no longer serve as a government strategic buffer. Russia’s Amur Gas Processing Plant has design capacity roughly equal to Qatar’s entire output but faces Western sanctions. Algeria produces only 5 to 10 percent of global supply. Tanzania’s emerging helium projects are years from commercial production.
Phil Kornbluth estimates a minimum three-month disruption to helium supply chains, plus two months for logistics normalisation. If the conflict extends beyond six months, the structural deficit has no easy solution…
…South Korea does not just make chips. It builds the ships that carry the gas that the rest of the world needs to replace Qatar’s output.
South Korean shipyards, HD Hyundai Heavy Industries, Samsung Heavy Industries, and Hanwha Ocean, delivered 248 LNG carriers between 2021 and 2025, versus 48 from China. That is an 83.8 percent share of LNG carrier deliveries over the past five years, per BusinessKorea. Korean yards currently hold approximately two-thirds of the global LNG carrier orderbook by value, with LNG vessels accounting for 52 percent of their total backlog at $71.3 billion, per VesselsValue.
A single 174,000-cubic-metre LNG carrier costs $220 to 260 million at current pricing. Construction takes 30 to 36 months from steel cutting to delivery. Korean yards have orderbooks extending through 2028. New orders placed today face delivery in late 2028 or 2029.
Korean vessel exports hit $31.8 billion in 2025. Gas carriers make up over 60 percent of order composition…
…South Korea’s energy crisis, caused by the Hormuz closure and Qatar’s force majeure, puts pressure on the industrial base that builds the LNG carriers the world needs to transport replacement gas. If Korean industry faces sustained energy disruption, supply chain delays, or cost inflation, carrier construction timelines could slip. If carrier construction slips, the global LNG fleet grows more slowly at precisely the moment the world needs more ships. If there are not enough ships, the global gas shortage deepens. If the gas shortage deepens, energy prices rise further. If energy prices rise further, Korean industry takes a harder hit.
I want to be precise about the limitations of this argument. There are circuit breakers. South Korea is restarting five nuclear reactors and easing coal restrictions. Shipbuilding is moderately energy-intensive, far less than steelmaking or semiconductor fabrication. There is currently an oversupply of LNG carriers, with approximately 60 idle ships providing buffer. Any disruption to shipyard output today would only affect deliveries in 2028 to 2029, given build timelines.
5. Notes from the SaaS Funeral – Reid Hoffman
Just two weeks ago, a single tweet about Claude Code was enough to wipe five percent off SaaS stocks. I understand the instinct. But I think the inference most people are drawing is wrong, and it’s worth being precise about exactly where the logic breaks down…
…Most of the arguments here fundamentally misunderstand software businesses as just lines of code you generate once. They are living systems that require maintenance, verification, security, compliance, and ongoing refinement…
… A CRM company that ships a deeply intelligent set of agents that iteratively refine your sales workflow, that understands your pipeline more comprehensively than any human analyst, that comes with powerful backend libraries purpose-built for that domain has an extremely well-crafted moat…
…The business model will shift, too. We may see more models where customers prepay token budgets much like a utility. For example, a CRM company that reimagines its economic model around compute consumption and scale. We’ve experienced business model transitions like this before. We went from on-premises software to cloud SaaS and the world didn’t end; it expanded. We’re making a similar transition now, from cloud to AI-native…
…And Jevons’ Paradox will do what it always does… as the cost of building software drops dramatically, the demand for software will expand dramatically.
Disclaimer: None of the information or analysis presented is intended to form the basis for any offer or recommendation. We currently have a vested interest in Meta Platforms and TSMC. Holdings are subject to change at any time.
