Tech Sector Rally Hits Record, AI Cost Shift Looms, IPO Underperformance Confirmed

Key Takeaways

1. The Tech Sector ETF surged 47% in 9 weeks, the largest such advance ever, surpassing the late-1999 move. S&P 500 earnings are expected to grow 25% this year, and Apple trades at over 10x sales, a record high valuation. The S&P 500 dividend yield dropped below 1%, near the 2000 low. — via 1 2 3 4
2. Marc Andreessen predicts 80% of AI workloads will run on cheap models within 12-18 months, with energy and compute as constraints, not better models. — via 1
3. Historical data shows the average US IPO returned only 6% annually in its first 3 years, roughly half the market return. — via 1 2
4. The equity value held by those under 40 tripled in the 2020s, and the proportion of IRA holders under 30 doubled, a positive sign for young investors. — via 1
5. Meb Faber and Clifford Asness argue that bonds in 60/40 portfolios are more dilutive than diversifying, advocating for trend following and alternatives. — via 1 2

1. Market Extremes and Valuation Signals

1. Charlie Bilello reports multiple extremes: The Tech Sector ETF (XLK) surged 47% in 9 weeks, the biggest 9-week gain ever, surpassing the late-1999 parabolic move. S&P 500 earnings are expected to grow 25% this year, an unprecedented boom outside of post-recession rebounds, driven by big tech. Apple trades at over 10x sales, the highest valuation in its history. The S&P 500 dividend yield dropped below 1%, near the all-time low of 0.94% set in 2000. — via 1 2 3 4
2. Liz Ann Sonders notes that the NASDAQ 100 fell about 5% on Friday, the largest single-day drop since April 2025. The New York Fed’s May 1-year inflation expectation fell to 3.46% from 3.64%, and 3-year to 3.13% from 3.15%. — via 1 2
3. Ben Carlson highlights a positive generational shift: the equity value held by those under 40 tripled in the 2020s, and the proportion of IRA account holders under 30 doubled in the past decade, attributing this to technology and lower fees. — via 1

2. AI Cost Trends and Economic Transformation

1. Marc Andreessen outlines a cost-driven bifurcation in AI: 80% of workloads will run on cheaper models within 12-18 months, while 20% will require the latest models. The limiting factor will be energy and compute, not better models. He cites Coinbase routing prompts to cheaper models as an example. — via 1
2. Raoul Pal argues that GDP measurement is outdated and the new economy includes exponential growth components, with agents self-reproducing following Reed’s Law, leading to a Cambrian explosion of economic participants. — via 1 2

3. Historical Investment Patterns and Portfolio Construction

1. Charlie Bilello and Meb Faber both highlight poor IPO performance: The average US IPO returned only 6.0% annually in its first 3 years (half the market). Meb Faber echoes Buffett’s view that IPOs are a negotiation favoring sellers. — via 1 2
2. Meb Faber and Clifford Asness critique the traditional 60/40 portfolio, arguing that bonds are more dilutive than diversifying. Faber suggests true diversification includes trend following and market neutral strategies; Asness similarly states bonds dilute equity returns. — via 1 2