S&P 500's Strongest Decade Since 1990s, AI Investing Concentration, and Emerging Geopolitical Tensions

Key Takeaways

1. S&P 500 annualized 15.6% since 2020, the best decade since the 1990s, but 19 double stocks this year are AI-related, signaling narrow speculation and threat to software industry. — via 1 2
2. 2020s is the worst decade for 10-year Treasury bonds, with annualized total return -0.6%, low yields causing low returns and high duration risk. — via 1
3. US government debt has increased by over $3 trillion in less than a year after a $5 trillion debt ceiling hike; government spending as share of economy at all-time high, potentially hitting debt ceiling again by 2027. — via 1 2
4. Next-generation AI frontier models may come from application companies with last-mile distribution, not from OpenAI or Anthropic. — via 1
5. The anti-AI data center organization "Humans First" is actually funded by Facebook co-founder Dustin Moskovitz, exposing a political "Trojan horse." — via 1
6. US stock ownership rate is ~55%, highest globally, rooted in a "ownership culture" where ordinary people share in economic growth. — via 1

1. Market and Economy

1. S&P 500 has annualized 15.6% since the start of 2020, marking the strongest decade-like period since the 1990s. However, this year 19 of the stocks that doubled are AI-related, indicating a narrow speculative frenzy where the market views AI as a threat to the software industry rather than a broad-based growth story. — via 1 2 3
2. The 2020s have been the worst decade for 10-year Treasury bonds, with an annualized total return of -0.6%. Low starting yields led to low returns and high duration risk. — via 1
3. Charlie Bilello questions the Fed's decision to expand its balance sheet (QE) again while the stock market is at all-time highs, credit spreads are tight, and inflation has been above target for 63 consecutive months. — via 1
4. Less than a year after a $5 trillion debt ceiling increase, US national debt has already grown by over $3 trillion. Government spending as a share of the economy is at an all-time high, and at the current pace, the debt ceiling will be hit again by 2027. — via 1 2
5. Capital flows show stark divergence: South Korea's stock market gained 126% in 2026 (likely a typo, but data as given), Taiwan +73%, while China fell 12%. — via 1
6. The Memory ETF ($DRAM) became the fastest ETF to reach $10B, $15B, and $20B in assets under management. — via 1
7. Meb Faber highlights that the US stock ownership rate is about 55%, the highest in the world, rooted in an "ownership culture" where ordinary people expect to share in economic growth. — via 1

2. AI and Technology

1. A key insight from @Jason: next-generation AI frontier models are likely to come from application companies that control the last-mile distribution, not from pure research labs like OpenAI or Anthropic. This suggests the value capture in AI shifts to companies with direct user access. — via 1
2. Marc Andreessen amplifies a report revealing that "Humans First," an organization protesting AI data centers and appearing conservative, is actually funded by Facebook co-founder Dustin Moskovitz and staffed by left-wing activists – a "Trojan horse" undermining AI infrastructure. — via 1
3. Andreessen also shares a "woke" perspective on AI value distribution: open-source models cause frontier labs to lose value, and eventually everyone will run their own startup operated by a "guardian angel" AI, democratizing entrepreneurship. — via 1
4. Elad Gil notes that the AI world is rapidly transforming code, product development, research, and mathematics, calling it an exciting era of constant change. — via 1

3. Geopolitics and Culture

1. Lyn Alden highlights a growing concern: the US is also moving toward a VPN ban, following trends in other countries. — via 1
2. Clifford Asness criticizes the New York Times for promoting wealth redistribution and demonizing wealth accumulation, calling out the myth that Elon Musk has vast unused resources. — via 1
3. Asness also points to a double standard in US foreign policy: attacking drug-exporting countries is acceptable, but opposing Israel's strikes on rocket-launching neighbors is not. — via 1
4. Marc Andreessen weighs in on the Ezra Klein controversy, arguing that the principle "attending an event doesn't mean endorsing participants" died in 2017, and sarcastically imagining scenarios. — via 1