Hyperscalers Are Collapsing. Is the AI "Bubble" Crashing or Consolidating?
Hyperscalers Are Collapsing. Is the AI "Bubble" Crashing or Consolidating?
YouTube56 min 14 sec
Watch on YouTube
Note: AI-generated summary based on third-party content. Not financial advice. Read more.
Quick Insights

Investors should rotate away from expensive Mag-7 hyperscalers like Microsoft (MSFT) and Alphabet (GOOGL) toward the "receivers" of AI capital, such as Micron (MU) and Vertiv (VRT). Eli Lilly (LLY) is a high-conviction long-term play, positioned to potentially become the world’s largest company as it leverages AI for drug discovery and dominates the GLP-1 market. While Micron (MU) remains a strategic asset in the AI memory shortage, focus on its transition to subscription-like revenue models rather than expecting rapid 10x gains. Avoid "catching the falling knife" in Bitcoin (BTC) until it clears its 200-day moving average, though you should prepare for a 2025 surge driven by AI agents using crypto for autonomous transactions. For broader exposure, look to Small Caps (IWM) and industrial names like Caterpillar (CAT), which are benefiting from a market rotation that favors infrastructure and domestic growth over high-multiple tech.

Detailed Analysis

Hyperscalers & Mega-Cap Tech (MAG-7)

The market is currently experiencing a mid-cycle slowdown, characterized by a rotation away from high-multiple "spenders" (Hyperscalers) toward "receivers" of AI capital. • Hyperscalers (Microsoft, Alphabet, Amazon, Meta, etc.) are down significantly month-to-date (approx. 14%). • The "MAG-7" is facing multiple compression: despite growing profit margins and earnings, their stocks are under pressure because of massive CapEx spending and uncertainty regarding when that spend will translate into revenue. • This is a rotation, not a fundamental collapse; the equal-weight S&P 500 and New York Stock Exchange breadth remain near all-time highs.

Takeaways

Short Hyperscalers relative to the thematic AI trade: The speaker prefers being on the short side of the big spenders while staying long the companies receiving that money. • Avoid High PEs: In this phase of the market, high Price-to-Earnings (PE) ratios are under pressure. Look for names where earnings growth justifies the price. • Expect Consolidation: A consolidation resets success through price, time, and volatility. The market is currently "shaking out weak hands" to allow moving averages to catch up to prices.


Micron Technology (MU)

Micron is cited as a primary example of the AI infrastructure trade and a "canary in the coal mine" for the mid-cycle slowdown. • Reported a "massive blowout" of earnings but the stock finished down for the week, signaling a slowdown in the rate of change of growth. • Supply Shortages: Management indicated that memory and storage shortages will take considerable time to improve, with no clear "line of sight" for supply catching up to demand before 2028. • Business Transformation: Micron is moving toward a subscription-like model with Strategic Customer Agreements (SCAs)—take-or-pay deals that now represent a significant portion of future revenue.

Takeaways

Valuation: At roughly 7.5x 2027 earnings, the stock is viewed as fairly priced, though the "10-bagger" (10x return) era for this specific name may be over. • Memory as a Strategic Asset: Memory is now architecturally essential for AI; demand is accelerating faster than hardware can be built. • Watch for "Violent" Downside Later: Once supply eventually catches up to demand (likely years away), the price correction will be sharp.


Eli Lilly (LLY) & Biotech

Biotech is entering a "renaissance" driven by AI-accelerated drug discovery. • Eli Lilly is predicted to become the largest company in the world within five years. • Growth is driven by GLP-1 (and upcoming GLP-3) medications, which address massive markets in obesity and addiction. • AI is shifting market sentiment toward healthcare as a "defensive growth" sector.

Takeaways

Application Layer Opportunity: Shift focus to the "application stage" of AI, where companies like Eli Lilly use AI to improve ROIC (Return on Invested Capital). • Biotech Consolidation: The sector has consolidated after a bull run and is now starting to move higher; look for "wicks" in charts signaling accumulation.


Bitcoin (BTC) & Crypto

The speaker maintains a bearish short-term view but a highly bullish long-term thesis based on "Agentic Commerce." • Current Status: Bitcoin is in a bear market, trading below its 200-day moving average and making new lows. • The "Debasement Trade" Unwind: Assets without traditional valuations (Bitcoin, Gold, Silver) are currently being sold off. • Future Catalyst: AI agents will eventually need to transact autonomously. Crypto is the "native rail" for these agents.

Takeaways

Don't be a hero: There is no sign of a bottom yet. Wait for positive momentum before building large positions. • Watch the Ecosystem: The speaker is tracking a 40-name "Tokenized Index" of AI-related crypto projects, which is showing relative strength compared to Bitcoin. • Timeline: Expect a potential bottom later this year, with a major focus on crypto for 2025.


AI Investment Themes & Sectors

The discussion highlights a shift from infrastructure to agentic workflows.

AI Agents & "Loops"

Agentic Loops: The next abstraction layer where AI agents repeatedly run tasks (code review, debugging, marketing) without human intervention. • Token Usage Explosion: Unlike human users, "loops" run 24/7. This will drive exponential demand for compute and memory (benefiting chip makers). • Job Market Impact: Profit margins will expand not by replacing all workers, but by "power users" employing loops to replace the workflows of the bottom 98% of staff.

Small Caps (IWM)

• Surprisingly, Small Caps (IWM) have outperformed recently despite the tech sell-off. • This suggests that the market is not fearing a recession, but rather rebalancing away from overextended tech giants.

Takeaways

Focus on the "Receivers": Invest in the companies providing the "food" for AI (Memory, Power, Industrials) rather than just the big cloud providers. • Industrial AI: Names like Caterpillar (CAT) and Vertiv (VRT) are part of the "Industrial Momentum" trade related to data center build-outs.

Ask about this postAnswers are grounded in this post's content.
Video Description
Visit : https://ai.22vresearch.com/ Contact : mwhaling@22vresearch.com for more information In this week's video, I walk through a brutal June for the hyperscalers and the Mag 7, and why I think investors have to adapt to a midcycle slowdown, a pause inside an expansion where growth stays positive but the rate of change comes down. The S&P was off 2% on the week, but this is not a fall. Healthcare ran +8%, real estate +4%, utilities higher, while the Mag 7 complex, consumer discretionary, and info tech sold off. The Q's dropped nearly 5%, yet IWM outperformed for a third straight week. This is rotation, rotation, rotation. Hyperscalers are down 14% month-to-date, and I still want to be short them relative to the thematic AI trade, not on fundamentals, but as continued benchmark arbitrage, a reweighting away from high-multiple spenders whose profit margins keep growing even as the stocks get pounded. The receivers of the money are driving everything. This remains an earnings-driven bull market, not a speculation-driven one: we're tracking the exact 15-year annualized pace, yet it's still called a bubble. The real story underneath is memory, Micron's blowout, the Anthropic deal, take-or-pay agreements, and no line of sight on supply until 2028 and beyond. With RSI now live and loops arriving, token demand only accelerates. Timestamps • (00:00–04:46) Maine intro and the midcycle slowdown framing. Week's scorecard -S&P -2%, healthcare +8%, real estate +4%, utilities up, Mag 7 / discretionary / info tech down, Q's -5%, IWM +1.5%. Inflation expectations falling across swap, 10-year, and 2-year rates; Kevin Warsh's hawkish moment amounted to nothing. • (04:46–07:32) Hyperscalers -14% MTD. Short the spenders relative to the thematic trade. This is benchmark arbitrage, not a fundamental break. The S&P sits at its 14.7% 15-year annualized pace, yet gets called a bubble, Nikkei +38%, Kospi +100%, TX +53% on heavier AI benchmark weighting. • (07:32–11:14) The receivers of the money are driving everything. What a consolidation actually is. The token index has likely peaked; 60% of companies are migrating to cheaper models, but Jevons paradox keeps compute climbing. • (11:14–15:12) Thematic-portfolio rate-of-change rolling over (30-day peaked in May, 50-day in early June), slowing growth, not a top. Micron's blowout with the stock down on the week is the symbol of that slowdown. Consolidations let moving averages catch up to price. • (15:12–18:29) Pick-your-poison retracement levels off the Opus 4.5 breakout. Long thematic vs. hyperscalers is still a bull market; equal-weight S&P is fine. Don't get mega-bearish forward EPS has gone parabolic and the S&P's 7.4% YTD is mostly the spenders. • (18:29–21:30) The earnings facts: S&P 600, global EPS, info-tech long-term growth at 38%, semis justified by earnings. Token demand growing exponentially. Japan's $2.3T plan. Multiple compression = uncertainty on revenue timing while margins still grow. • (21:30–26:04) Apple as the canary, no Mac minis or Mac studios to run GLM 5.2 privately; hardware can't keep pace with model capability. Micron–Anthropic deal, memory shortage with no line of sight to 2028. RSI is here; smartphones, autos, robotics, and humanoids all add memory demand. • (26:04–31:28) Take-or-pay SCAs, 16 strategic agreements, ~20% of volume, ultimately half of revenue. L2+ vehicle mix doubling; humanoids carry 10x a car's memory. Sovereign AI as necessity. The application stage is where better Sharpe ratios appear hunt PEG ratios below 1, like Eli Lilly. • (31:28–37:09) Loops as the next abstraction layer after agents (Boris from Claude, Steinberger's OpenClaw post, Karpathy). The workforce math: power users teach agents the workflow, then the other 98 go away. Token usage explodes; Claude tags join the team. • (37:09–43:44) The AI mindset paper build your own knowledge brain. The thematic-portfolio loop architecture. Healthcare and the biotech renaissance; why I think Eli Lilly becomes the largest company in the world inside five years on GLP-1s and GLP-3s. • (43:44–49:32) Aura Ring and HRV. Frontier-lab cash-flow timing as speculative — I'm not permabull on the spenders. AI driving net wealth (Johnson Redbook), housing vs. parabolic tech capex (Warren Pies). Stripe Sessions: the singularity moment, vibe coding to vibe deploying. • (49:32–55:54) The agentic sequence agents write code, operate infrastructure, deploy, then transact. Crypto is the financial guardrail in layer five of the cake. Bitcoin's in a bear market; don't be a hero. Tokenization releases the dormant two-thirds of $700T, with Bitcoin as collateral. The 40-name tokenized index is leading Bitcoin out. Crypto YouTube launching Sept/Oct with OpenBB. See you from Maine.
About Jordi Visser
Jordi Visser

Jordi Visser

By @jordivisserlabs

Empowering seasoned professionals to navigate the future of finance, technology, and AI. What We Offer: - Cutting-edge ...