🔥 How AGI Kills CAPE — And Why Warren Buffett's Favorite Metric No Longer Matters

📊 The CAPE Ratio and the AGI Era

Warren Buffett's preferred market timing metric — the Cyclically Adjusted Price-to-Earnings (CAPE) ratio — is flashing red. But in the age of artificial general intelligence, this 120-year-old indicator may have lost its relevance entirely.

The Shiller CAPE ratio, which measures the S&P 500 price against 10 years of inflation-adjusted earnings, currently sits at 41 — a level that historically signals overvaluation and weak forward returns. For context:

• Long-run average (since the late 1800s): ~16
• 1929 peak: 32
• Dot-com peak: 44
• Current level (September 2024): 41

Traditional wisdom suggests this signals an impending correction. But this framework was built for an industrial economy — not an era of exponential intelligence, infinite ROI potential, and rapidly compounding AI-driven businesses.

"CAPE does not work in the age of AGI. We are going into exponentiality... This isn't a railroad or the invention of the internet. This is way bigger."

🧠 Why Old Metrics Fail in the New Paradigm

The coming transformation isn't comparable to past technological shifts. AGI represents:

• Infinite intelligence at near-zero marginal cost
• Software development by AI — accelerating productivity beyond human limits
• Robots driven by AI — generating cash flows with minimal human input
• Abundance economics — where traditional scarcity-based models break down

Consider this: a Tesla Cybercab might cost $15,000 to produce but generate $50,000 per year in autonomous revenue. That's an infinite ROI scenario — the kind of cash-generating engine that renders traditional PE ratios obsolete.

📉 The Buffett Warning Sign — And Why It's Misleading

Warren Buffett himself has moved to over 40% cash in his Berkshire Hathaway portfolio, citing high valuations. But Buffett's approach has underperformed the S&P 500 for the last 18 years — a telling sign that his framework no longer captures the reality of modern markets.

Berkshire's portfolio is heavy on old economy plays: insurance companies, furniture retailers, candy manufacturers, and legacy banks like Bank of America. Buffett famously avoided technology during the internet boom — and continues to shy away from exponential growth sectors today.

"The legend ran out of juice quarter of a century ago... because he was afraid of technology. He was afraid of exponentiality."

Over the past 20 years, the QQQ (Nasdaq-100 ETF) has delivered a 31x return — a trajectory that will only steepen as fewer companies capture more profit in the age of AI.

⚡ The Concentration of Wealth is Accelerating

The AI revolution is driving a winner-takes-most dynamic. More profit is flowing to fewer companies, and those with the most compute, the best infrastructure, and the strongest network effects will dominate. This makes traditional diversification strategies less effective and conviction-based portfolio construction more critical.

The firms building the picks-and-shovels infrastructure of AI — semiconductor companies, data center operators, cloud hyperscalers, and compute providers — are positioned to capture this exponential value creation.

🔐 The Risk of Holding Too Much Cash

One question raised the issue of sitting on 30% cash while valuations appear stretched. While cash currently offers decent returns — 6% risk-free in traditional markets and up to 12% in crypto-linked stablecoin yields — the opportunity cost of staying sidelined is significant.

"Scared money don't make money... 30% cash is way too high."

The recommendation: deploy capital strategically, especially during pullbacks, rather than waiting for a massive correction that may never come. The exponential compounding potential of the current cycle far outweighs the risk of short-term volatility.

🚀 Tesla and SpaceX: What Happens in a Merger?

Speculation around a potential Tesla-SpaceX merger has intensified. Prediction markets on Kalshi currently show:

• 37% odds of a merger before May 1, 2027
• 72% odds by the end of 2028

When prediction markets hit the 70-80% range, it's typically a strong signal that the event is highly probable.

What This Means for LEAP Holders

If the merger happens, here's what to expect:

• Stock deals: LEAPs (long-term equity options) get adjusted based on the exchange ratio. Intrinsic value shifts to reflect the new combined entity.
• All-cash deals: The time premium on LEAPs disappears. Holders receive only the intrinsic value, which can be a significant loss if the options were purchased out-of-the-money.
• Out-of-the-money LEAPs: These can go to zero even with years left on the clock if the deal doesn't favor your strike price.

"Never buy out-of-the-money LEAPs unless it's an extreme bargain... Always have intrinsic value."

For context, early Tesla LEAPs purchased when the stock traded near $106 with a $140 strike expiring in December 2026 are now delivering a 14x return as the stock approaches $450. That's the power of buying LEAPs during periods of extreme fear and dislocation.

📉 When to Buy LEAPs — and When to Avoid Them

LEAPs are not tools for speculation at market tops. They are instruments for capitalizing on oversold, high-conviction opportunities.

Key Rules for LEAP Trading:

• Buy when assets are beaten down — ideally during capitulation or broad market fear
• Always include intrinsic value — never go all-in on time premium
• Target a 50/50 split between intrinsic and extrinsic value
• Use indicators like IDSS to identify buy signals and avoid buying at resistance levels

Recent examples of successful LEAP trades include Marvel, which delivered an 8.35x return on LEAPs and a 33.4x return on synthetic longs. These trades were entered during clear buy signals and held through the recovery.

"You do not buy a LEAP at a top. You buy them when they're super beaten down cheap."

🤖 The China Factor: Humanoid Robotics and AI Competition

The global AI race is increasingly a two-horse race: USA vs. China. Both nations are aggressively building infrastructure, but China has two critical advantages:

• Faster production scaling — rivaled only by companies like Tesla
• Abundant electricity supply — essential for running massive compute clusters

Chinese Humanoid Robot Plays

Several Chinese companies are positioning themselves in the humanoid robotics supply chain:

• Sanwa: Actuators
• Invance: Servos, drivers, and body components
• HSAI: Vision sensors and LiDAR
• Xpeng (XPEV): EV company with a robotics division
• Ubtech (UB): Hong Kong-based with strong revenue growth
• Horizon Robotics: Chips for robots

However, Chinese stocks carry significant risks:

• Lower valuations but less transparency
• Delisting risks
• Governance concerns
• Difficulty trading options (e.g., covered calls)

"If you want to play humanoids, buy Tesla. It's that simple. Don't overthink it."

🔒 Open Source vs. Closed Models: The Real Battle

The debate between open-source and closed-source AI models is heating up. While closed labs like OpenAI and Anthropic currently capture more revenue, the trend is shifting:

• ~79% of AI tokens are now generated by open-source models
• ~21% come from closed-source providers

Big enterprises are increasingly building on open-source models to avoid vendor lock-in and protect their intellectual property. This shift moves margin from the model layer to infrastructure and applications — exactly where the IA13 thesis focuses.

"Your margin is my opportunity." — Jeff Bezos

Regulatory capture could temporarily favor closed labs, but open-source AI is unstoppable. Code spreads like wildfire, and attempts to ban or restrict it are futile — much like trying to ban drugs in a prison.

🧮 Shorting Anthropic: A Contrarian Bet

There's growing skepticism around the valuations of OpenAI and Anthropic. Both are projected to lose hundreds of billions of dollars over the next few years, despite high revenue growth.

Key concerns:

• No durable moat: Model weights fit on a thumb drive
• High costs: Closed models are expensive to run and expensive for customers
• Open-source competition: Margins will compress as alternatives proliferate

If these companies cannot sustain their current growth trajectories, their valuations could collapse — creating a potential short opportunity upon IPO.

"Closed labs have no future... They charge way more than the others, and their margin can't exist in the free world."

⚙️ SpaceX: The Compute King

One underappreciated advantage of SpaceX is its dominance in compute infrastructure. The company is projected to generate $100 billion in annual recurring revenue (ARR) from hosting compute for AI workloads.

Why SpaceX wins:

• No one else has enough compute — not Microsoft, Google, OpenAI, or Anthropic
• Fastest data center spin-up in the industry
• Owns its own energy infrastructure via Megapacks
• Controls 30-40% of all chip supply from key vendors

"He or she who has the most compute wins the AI game — not the weights of a model that fit on a thumb drive."

🎯 Final Takeaways

• CAPE and other traditional metrics are obsolete in the age of AGI and exponential business models
• Holding excessive cash (e.g., 30%) is a opportunity cost in a compounding environment
• LEAPs are powerful tools — but only when bought during fear and with intrinsic value
• Open-source AI is winning the long game, shifting value to infrastructure providers
• SpaceX is the compute king — and that's the real moat in AI
• China is a serious competitor — but Tesla remains the simplest way to play humanoid robotics

The age of exponentiality is here. The winners will be those who understand the new rules — and ignore the outdated playbooks of the past.