Hand-picked finance shows,
read for you.

Keller argues that AI adoption remains early on its S-curve, supported by double-digit revenue growth, open capital markets, and continued hyperscaler investment, even as infrastructure financing becomes more debt-dependent. Market performance is becoming more dispersed, with weaker reactions to AI announcements, greater scrutiny of data-center economics, and potential opportunities outside crowded technology strategies. He sees unresolved risks in corporate adoption, valuation, leverage, and construction constraints, while favoring cash, lower exposure, differentiated stock picking, and selective lower-multiple AI beneficiaries.

• Keller expects hyperscalers to keep spending because stopping could create greater strategic risks than preserving capital. He warns that a 4% or 5% return on invested capital would be weak for shareholders, though not necessarily capital-destructive.
• Keller views Meta's willingness to rent excess compute primarily as a message to investors rather than a major business change. Three-month contracts could generate a couple billion dollars in profit, but that would be modest relative to Meta's valuation.
• Keller argues that AI adoption is still climbing its S-curve rather than approaching a blowoff top. He cites double-digit growth at Apple Services and Google Search, and growth in the 20% range at Meta. Some AI labs could sustain double-digit growth for 10 to 15 years.
• Keller expects AI to expand the software market and increase demand for database software because every application requires an underlying database. Large language models may increasingly drive database selection instead of human developers, leaving the competitive mix uncertain.
• Keller believes traditional, differentiated stock picking may outperform as crowded technology strategies unwind. He points to 2022, when many technology investors rode favored names downward.

The video examines why estimates of oil moving through the Strait of Hormuz differ, with government data showing a roughly 8-million-barrel-per-day decline from prewar levels and later IEA data favoring the government's estimates. It explains how AIS, optical imagery, SAR, hyperspectral imaging, and military surveillance differ in their ability to detect dark ships and reconstruct hidden transfers. Shin also assesses the maritime-intelligence industry and argues that higher Hormuz flows may affect prices, while declining strategic reserves, Iran's expected resistance before the midterm elections, and potential attacks on ships keep disruption risks unresolved.

• Shin describes three tiers of maritime intelligence: free consumer tools with major gaps, commercial providers that charge thousands of dollars, and military systems available only to the U.S. government. Military systems are superior because of unhindered satellite access and the ability to decrypt ship-to-ship communications.
• Vortexa and Windward are smaller than Kpler, with estimated revenue of roughly one-fifth to one-fourth of Kpler's. They specialize in areas including risk, compliance, deceptive-practice profiling, and AI-driven analysis of waterborne energy flows.
• Shin identifies Kpler as a leading dark-fleet and shadow-inventory tracker because it fuses SAR, AIS, and customs data. Its customers include major oil companies, trading houses, investment banks, and hedge funds, although its commercial coverage remains below military capabilities.
• Commercial shipping-intelligence providers sell timely data to investment firms, hedge funds, and major oil companies for investing, planning, and strategic decisions. Shin describes the major firms as substantial businesses with roughly $8 billion to $12 billion or more in combined revenue.
• The reported 65 billion barrels of proven Venezuelan oil would not provide immediate relief. Developing the infrastructure needed to extract and ship it would take at least a couple of years.

Gareth Soloway examines technical setups in oil, gold, silver, and natural gas, concluding that oil remains trapped in a wedge while gold and silver are testing important support levels. He identifies natural gas as his favorite commodity because of a potential breakout, seasonal demand, European supply deficits, and data-center energy needs, while arguing that refiners such as Valero and Phillips 66 appear overbought and vulnerable to declines.