Momentum Is Crashing, Bitcoin Is Bottoming, AI Agents Are Rising
Momentum Is Crashing, Bitcoin Is Bottoming, AI Agents Are Rising
YouTube50 min 40 sec
Watch on YouTube
Note: AI-generated summary based on third-party content. Not financial advice. Read more.
Quick Insights

Investors should use current pullbacks to the 50-day moving average as entry points for high-quality AI infrastructure stocks like NVIDIA (NVDA) and Marvell (MRVL). The "easy money" phase has passed for memory leaders like Micron (MU), so focus should shift toward the AI application layer, specifically Healthcare and Insurance companies like Eli Lilly (LLY) and Chubb (CB). Technical indicators suggest a potential bottom for Bitcoin (BTC), making it a high-conviction rotation play alongside Gold and Silver to hedge against U.S. debt debasement. Monitor regional banks via the KRE ticker, as AI-driven automation is expected to trigger a wave of acquisitions and margin improvements within the sector. Despite hawkish market fears, the deflationary impact of AI productivity suggests the Federal Reserve may be less aggressive with rate hikes than currently anticipated.

Detailed Analysis

AI Infrastructure & Semiconductors

The AI infrastructure trade is currently undergoing a "volatility shakeout." While the long-term narrative remains intact, the market is flushing out retail momentum traders and hedge funds due to high Value at Risk (VAR) and crowded positioning.

  • Consolidation Playbook: The sector is experiencing a mid-cycle slowdown. A retracement to the 50-day moving average is viewed as a healthy reset rather than a bubble bursting.
  • Compute Scarcity: Despite concerns about capital expenditure (CapEx) cuts, industry leaders like Greg Brockman (OpenAI) suggest there will "never be enough compute" to satisfy global demand.
  • Memory Sector: While SK Hynix and Micron (MU) have seen massive gains, the analyst suggests the "easy money" (7-10 baggers) has been made. Future gains may be more modest (doubles or 200% over three years).
  • Key Tickers Mentioned: NVIDIA (NVDA), SK Hynix, Micron (MU), Marvell (MRVL), Entigris (ENTG), Fluence (FLNC).

Takeaways

  • Buy the Pullbacks: Use consolidations as opportunities to enter high-quality infrastructure names.
  • Manage Expectations: Shift focus from speculative "moonshots" to companies with proven earnings growth, as the trade is now earnings-driven rather than purely hype-driven.
  • Monitor Bottlenecks: Watch for data center power and cooling constraints, as these physical limits may temporarily cap the upside for hyperscalers.

The AI Application Layer (Software & Agents)

The focus is shifting from building AI (infrastructure) to using AI (applications). We are moving from "AI tools" to "AI workers."

  • AI Agents: The next phase involves "mission agents" capable of hours or days of uninterrupted work, moving beyond simple chatbots.
  • Enterprise Adoption: We are at an inflection point where companies are seeking ROI. This will lead to labor substitution and massive productivity leverage.
  • Open Source vs. Frontier Models: Chinese open-source models like GLM 5.2 are challenging U.S. frontier models (Anthropic, OpenAI). While frontier models will still capture 90% of economic value for complex tasks, open source will dominate general workflows.
  • Sectors Benefiting: Healthcare and Insurance are identified as the primary beneficiaries of the application layer.

Takeaways

  • Look for Productivity Gains: Invest in companies that can successfully integrate AI agents to cut costs and improve margins.
  • Healthcare & Insurance Focus: Watch for breakouts in these sectors as they automate paperwork and call centers. Key names mentioned: Eli Lilly (LLY), McKesson (MCK), Travelers (TRV), Allstate (ALL), and Chubb (CB).
  • Diversify AI Exposure: Don't just own the chipmakers; start looking at the software companies (SaaS) that are successfully deploying agentic workflows.

Bitcoin (BTC) & Precious Metals

Bitcoin is showing signs of a potential bottom despite being below its 200-day moving average. The "debasement trade" remains a core theme due to worsening U.S. federal debt paths.

  • Volatility Bottom: Bitcoin's 60-day volatility is at historical lows, which often precedes a significant price move.
  • Macro Correlation: Bitcoin has recently followed interest rate hike expectations. If the Fed remains status quo or shifts dovish, Bitcoin, Gold, and Silver are expected to rally.
  • Political Factors: Mentions of the "Clarity Act" and political shifts (Trump/Warsh) are creating a backdrop for potential regulatory clarity.

Takeaways

  • Bullish Divergence: Technical indicators are starting to show lower lows in price but higher highs in momentum, a classic sign of a trend reversal.
  • Rotation Strategy: The analyst has rotated a majority of "AI infrastructure money" (specifically from semiconductors) into Bitcoin, Silver, and Gold names.
  • Hedge Against Debt: Maintain exposure to these assets as a hedge against the long-term "math" of U.S. mandatory spending and interest costs.

Macroeconomic Outlook & The Federal Reserve

The discussion centers on Kevin Warsh (potential Fed Chair candidate) and his view on AI-driven productivity.

  • AI & Deflation: Warsh views AI as a "Hyper-Moore’s Law" event that brings deflationary pressure and job disruption. This suggests the Fed may not need to be as hawkish as the market fears.
  • Labor Market: The jobs market is weaker than headline numbers suggest. If you strip out Healthcare, there is almost no job growth. The "Quits Rate" is down, indicating workers are less confident.
  • Inflation: Headline CPI is expected to trend lower due to falling energy prices (gasoline futures).

Takeaways

  • Ignore the "Rate Hike" Noise: The analyst disagrees with banks predicting three more rate hikes, citing restrictive current policy and AI-driven productivity.
  • Financials (KRE/KIX): Watch for consolidation in regional banks (KRE). Larger, AI-ready banks are likely to acquire smaller ones to automate their books and reduce headcount.
  • Monitor Revisions: The bull market remains healthy as long as earnings estimate revisions stay positive. Currently, U.S. and European revisions remain strong.
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 break down why the midcycle slowdown within AI continues to dominate the narrative. Equal weight S&P is breaking out, IWM made new all-time highs, NYSE cumulative breadth hit an all-time high, and the S&P is consolidating in a triangle after spending June flat. Meanwhile, momentum volatility, now running hotter than the dot-com era, is flushing out hedge funds with VAR limits and retail traders chasing breakouts. Headwinds don't stop things. Planes don't hit a headwind and go straight down. You get turbulence, and that's what we're going through. The deeper story is threefold. First, the podcasts: Greg Brockman says there will never be enough compute in the world to satisfy demand, Scott Wu describes agents moving from tasks to missions, and Gavin Baker says GLM 5.2 forced him to rethink his framework, the future is composable models, with frontier tokens capturing 90% of economic value while open source handles the workflow layer. Second, Kevin Warsh: he called AI improvement "hyper Moore's law stuff" so why would a pro-AI Fed chair hike into deflationary pressures and job disruption? The market pricing hikes makes no sense against negative CPI prints, a 3.5% median wage, and a quits rate signaling no labor market strength outside healthcare. Third, Bitcoin: sitting at its lowest 60-day vol, with the first divergences since July of last year, lower lows, higher highs, on bad news. Timestamps • (00:00–01:42) Intro from Maine: midcycle slowdown continues, the volatility shakeout, and why the application layer and Warsh matter for Bitcoin. • (01:42–04:39) Markets: Equal weight S&P breaking out, IWM at all-time highs, S&P in a consolidation triangle, NYSE breadth at all-time highs. Hyperscalers (Meta, Amazon, Microsoft, Google) stuck in multiple compression despite growing earnings; financials setting up to break higher. • (05:00–08:25) The megaphones: Semi/software rotation ranges, Morgan Stanley momentum index trading outside 5% bands, TMT volatility above dot-com levels. This is how crowded trades get uncrowded. Plus, why the margin debt "bubble" chart is lame analysis relative to US market cap, we're below the post-GFC midpoint. • (08:25–09:53) Consolidation types and the headwinds: data centers, power and cooling, memory bottlenecks, regulation. Turbulence, not stoppage the headwinds matter more now because the earnings surprise is gone and the trade is crowded. • (09:53–14:34) Podcasts and open source: All-In with Gavin Baker on GLM 5.2 and composable models; frontier tokens capture 90% of economic value while open source processes 80%+ of tokens. Dean Ball's paper on the regulatory limbo risk. • (14:34–16:26) The memory architecture rumor: a claimed breakthrough in memory efficiency went viral, but with no details on timing or applicability, don't sell your memory stocks and don't expect another 7-to-10 bagger after you've had one. • (16:26–23:35) Scott Wu and Greg Brockman: agents moving from seconds to hours of work, less like a tool, more like a junior coworker. Companies aren't adopting AI tools, they're adopting AI workers. Enterprise adoption at the inflection point, "show me the ROI" phase. There will never be enough compute and the final seven minutes on AI in healthcare are must-listen. • (23:35–26:52) The playbook and the data: AI consolidation playbook re-posted, estimate revisions still strong in the US with Europe posting its biggest upturn in two years. Aggregate weekly payrolls sitting at boring post-GFC levels. • (26:52–38:26) Kevin Warsh deep dive: rejected forward guidance, "hyper Moore's law," and why a Fed chair who believes AI brings deflation and job disruption wouldn't hike. Real two-year rates driving the debasement trade selloff; crowded long-dollar positioning; negative headline CPI prints coming; Cleveland median PCE and Atlanta median wage both back at 2010–2020 levels; the CBO debt path everyone forgot $7 trillion in net interest by 2036. • (39:01–41:56) Application layer: insurance and healthcare benefiting most as infrastructure consolidates. KBW insurance breaking out (Travelers, Allstate on the 52-week high list), regional bank M&A coming as AI-capable banks buy up smaller ones. 8% of the S&P at 52-week highs, 13 of them financials. • (41:56–47:19) Meta reality check: the compute-resale story is a valuation play, not excess capacity, they said themselves they're still compute constrained with demand outrunning expectations. Physical capacity, not demand, is what limits hyperscaler capex surprises. Plus GLM 5.2 as a realistic risk to the Anthropic ARR multiple narrative. • (47:49–50:40) Bitcoin: lowest 60-day vol in the cycle, rate-cut expectations chart driving price, rate of change getting less bad, and the first divergences since July of last year arriving on bad news. If the Fed-hike positioning unwinds, it's good for gold, silver, and Bitcoin.
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 ...