Louis Gavelkal
About
CEO of Gavekal Research; investor and macro analyst specializing in China, geopolitics, and commodity markets
Cast within
No topic-region cast yet — this appears once Louis Gavelkal's compiled claims are aligned into a topic region's argument tree.
Claims by Louis Gavelkal (20 of 36)
Oil prices have been set in a 65-100 range by China's buffer stock strategy: China buys at 65 and below, stops buying at 100, creating a band that is great news for energy companies making healthy cash flows without needing capex; refiners at all-time highs with enormous free cash yield and no plans to build new capacity.
The problem in sovereign debt crises is never debt-to-GDP; it's always the percentage of debt owned by foreigners; when foreigners wake up and decide to exit (as happened in Thailand, Argentina, Greece), there's no buyer and the currency collapses; risk exists for France and Britain (both 30% foreign-held) but not the US because US assets will always have buyers at some price.
China was forced into open-sourcing AI because the US prohibited chip sales to China, preventing the closed-model strategy that requires ever-larger compute and capital spending. This forced workaround—having global software developers improve the product—contradicts China's historical preference for throwing money at problems.
US tech companies (Microsoft, Amazon, Google, Facebook) had zero cost of capital due to their great run and massive valuations, allowing them to throw capital at AI at scale; this was possible only because of 18 trillion in government debt yielding negative yields, which turned capitalism on its head.
China was happy to back away from seaborne oil market by tapping its 1.3 billion barrels (officially) or 1.8 billion (actually) of oil storage during the Russia-Ukraine war, and increased capacity to import from Russia on the sly; this was a unique capability allowing China to stabilize the oil market when it chose.
Central banks were huge net sellers of gold for 25 years, but after the Russia asset seizures, they became massive net buyers; Chinese individuals and foreign central banks (Indonesia, Thailand, Saudi Arabia) all started buying gold as a result of realizing treasuries and OECD bonds were not safe.
The US no longer controls the world's sea lanes because warfare has changed; you cannot protect billion-dollar ships with million-dollar missiles against $10,000 drones—the math no longer works—so the US couldn't control Red Sea or Persian Gulf, a massive shift from 80 years of assumed US naval dominance.
When Russia and Ukraine invaded each other and seize treasuries (including Russian oligarchs' assets like Chelsea Football Club without court proceedings), the West undermined the rule of law, its biggest comparative advantage; this signals to China and other nations that Western asset protection cannot be relied upon.
The lesson of the Iran/Strait of Hormuz crisis is that nations need to stockpile commodities (fertilizer, oil, rare earths, pharmaceuticals) rather than hold treasuries or gold, because you cannot buy what you need when sea lanes are disrupted; India learned this when unable to obtain fertilizer despite having 700 billion in treasuries.
Three factors are driving extreme crack spreads (refining margins) globally: Russia and Ukraine bombing refineries regularly; six major Gulf refineries taken out (including second-largest in Bahrain); China stopped exporting refined products domestically and prevented their export to India and Australia, as refining capacity globally is constrained.
We live in an inflationary boom where governments run 4-7% budget deficits of GDP (US at 7% with full employment), making investors search for scarcity; Q1 scarcity was energy (oil stocks up), Q2 was semiconductors (7 of 10 trillion global market cap increase), Q3 likely refining capacity.
China's strategy in AI is to deliver a Toyota (good-enough product at 80% cheaper) rather than a Ferrari; this makes them competitive globally especially with bootstrapping startups; most VC-backed startups use Chinese LLMs because they're essentially free and most startups can't afford expensive proprietary models.
We are in a structural bond bear market that has been ongoing for 5 years, kicked off by COVID policy failures (lockdowns and free money transfers); the easiest way for governments to pay for these costs is inflation; today all mentioned countries have higher-than-comfortable inflation rates, making them prone to inflationary bust if shocks occur.
My Notes
Loading notes...