Video summary

Banks, Gold, Biotech & AI: The Rotation Trade

Main summary

Key takeaways

Finance

Finance-focused summary (rotation trade: AI → banks/healthcare → gold; macro & positioning)

Macro / rates narrative & contrarian positioning

  • The hosts frame the current market rotation as driven by crowded positioning around the idea that rates will be raised 2–3 times (“rates up” macro narrative).
  • Jason’s contrarian view: when “everyone is on board,” the trade may be the most crowded—and potentially the most wrong.

Trade implication (macro pairs-style)

  • Consider being:
    • Long the short end of the bond market
    • And/or short the US dollar
  • Complement with:
    • Long gold
    • Long silver
    • Possibly long Bitcoin (“starting to work well”)

Additional positioning cues

  • Currencies people are massively short, including:
    • Swiss franc (CHF)
    • British pound (GBP)
    • They have “started to move the other way.”
  • “Short end of the curve” positioning is highlighted as crowded.

Market rotation / sector leadership (what’s outperforming)

  • Rotation is described as occurring in “full gear” around summer trading, after rebalancing into leading stocks.

AI leadership test (and wobble)

  • AI names are described as “crowded trades.”
  • They faced a “first true test” and saw multi-day drops, tied to Meta news (ticker not given).

  • As AI leadership wobbles, the rotation favorites discussed include:

    • Banking stocks / small banks / small caps (Jason’s point: they do better if rates don’t rise as expected)

    • “Industrial” Dow stocks (also framed as benefiting from a lower-rate environment)

Core logic: rates sensitivity shows up later

  • The hosts’ recurring idea: very high-growth AI-like companies may appear less sensitive to rates at first (“rates do not affect them” as much), but eventually they will.
  • They prefer to trade when market action confirms the contrarian thesis (rather than only relying on narrative).

Gold / miners divergence & timing notes

Divergence setup: GDXJ vs GLD

  • A gold setup is built around comparing:
    • GDXJ (junior gold miners ETF)
    • GLD (gold ETF price proxy)
  • Key observation:
    • Gold made lower lows in June (noted as June 24th),
    • but junior miners (GDXJ) did not make lower lows,
    • described as bullish divergence / relative strength.

Gold price level range mentioned

  • Gold’s move range cited: ~497 down to ~369 (high-to-low).

Timing remark

  • Gold “topped exactly on the day that Wars was appointed” (name appears as “Wars”; spelling unclear),
  • later moved as expectations shifted between rate cuts vs rate hikes.

Positioning tools & constraints (COT + flows)

  • COT (Commitment of Traders) oscillator interpretation:
    • Separates “commercials” vs “traders”
    • Green/red mapping referenced in the discussion (green = traders; red = commercials).
  • It’s described as very oversold, but not at “max oversold”, because readings improved over the last couple weeks.
  • Data timing note:
    • “Data comes out again today… but it was delayed on July 4th.”
  • ETF flow implication:
    • Gold ETFs had very strong outflow, framed as supportive of potential future strength (contrarian positioning logic).
Explicit constraint
  • Jason: his system cannot get long gold unless the COT changes this week.
  • If he can’t long gold, he may express the view via short dollar instead.

AI pullback risk management & “redeploy” framework around the 50-day

  • AI is treated as highly correlated:
    • idiosyncratic news (e.g., Meta compute remarks) can trigger broad selloffs.
  • Risk guideline:
    • During correlated selloffs, avoid being at “max exposure.”
  • Redeploy approach:
    • If adding/redeploying, consider the 50-day moving average area as a “normal” place to start in an uptrend.

Examples / tickers mentioned in the AI/semis discussion

  • NVIDIA (NVDA) (also referenced for market-cap scale)
  • Micron (MU) (“micron divergence” mentioned)
  • NBIS (stronger AI-related name hit hard; ticker spelling unclear)
  • AMD (described as very strong; “one closing session away from all-time closing highs”)
  • Dell (ticker not shown; described as holding in well)
  • TSM (“trillion dollar club” referenced later)
  • SMH (Semiconductor ETF; noted around/near the 50-day)
  • QQQ / “triple Q’s” (Nasdaq-100 ETF referenced)
    • correcting sideways for >1 month and “not broken down”

Timing/expectation

  • They suggest “coming days” are pivotal:
    • to determine whether there’s a back half of the summer rally
    • or if it’s “too fast” digestion.

Biotech & healthcare rotation (watch XBI / XLV; caution on crowded chasing)

  • Rotation includes:
    • XBI (biotech ETF): strong, but “tough” due to clinical trial dependence
    • XLV (healthcare ETF): also strong, but “not the same kind of strength”

Caution: don’t chase when extended

  • Don’t necessarily chase XBI if it’s extended:
    • risk described as “worst place at the worst time two times in a row.”
  • What to monitor:
    • how XBI behaves
    • how financials perform (rotation confirmation)
    • how semiconductors react off the 50-day (SMH)

Credit sentiment / “calm underneath”

  • JNK (junk bond ETF) is cited as a sign credit stress is not severe:
    • “still calm out there underneath the surface.”
  • This is framed as supportive of rotation, not a full breakdown.

Junk-bond / insto cashflow logic + 50-day explanation

  • Why the first pullback often resolves near the 50-day:
    • the 50-day is described as a proxy for the average price of the last quarter
    • buying interest can reappear when price revisits that area (investors who “missed it” reposition).

AI “end of trend” discussion (volatility & position sizing)

  • They argue AI’s rise is not over:
    • belief that there will never be enough compute (or enough fast enough),
    • implying structural tailwinds, but also “roadblocks.”
  • Macro/resource constraints (e.g., energy, construction speed) imply continued volatility.
  • Risk management principle:
    • Because some AI names rose dramatically (e.g., “700, 800, 900%”), volatility should rise, so position sizing should shrink.
  • Market mechanism disclaimer:
    • they reject “AI will solve trading” or making markets predictable
    • markets are framed as reflexive/self-adjusting.

Explicit “rotation likely continues” logic (mass-cap distribution math)

  • AI infrastructure leaders may not rotate out quickly due to market-cap scale:
    • NVIDIA ~ $5T explicitly stated
    • idea: moving large allocations (e.g., “world decides to sell 10% of Nvidia”) takes longer than a day
  • This supports:
    • AI-led uptrend may pause but not instantly reverse
    • other groups staying strong is interpreted as broader distribution occurring later.

Assets / tickers / instruments mentioned

ETFs / Funds

  • GDXJ (junior gold miners ETF)
  • GLD (gold ETF)
  • XBI (biotech ETF)
  • XLV (healthcare ETF)
  • SMH (semiconductor ETF)
  • JNK (junk bond ETF)
  • QQQ / “triple Q’s” (Nasdaq-100 ETF referenced)

Stocks

  • Meta (referenced; no ticker given)
  • NVIDIA (NVDA)
  • AMD
  • Dell (ticker not shown)
  • Micron (MU)
  • TSM
  • NBIS (mentioned; ticker spelling unclear)

Currencies

  • Swiss franc (CHF)
  • British pound (GBP)

Crypto

  • Bitcoin

Frameworks / step-by-step decision processes mentioned

Contrarian positioning confirmation (macro)

  1. Identify a crowded narrative (e.g., rates up / dollar up).
  2. Identify crowded positioning (e.g., short end of curve, CHF/GBP short).
  3. Check whether the market confirms the contrarian thesis during rotation.
  4. Express exposure via correlated trades:
    • If not long gold → consider short dollar (system/correlation logic).

Relative strength divergence method (gold)

  1. Compare miners ETF (GDXJ) lows vs gold ETF (GLD) lows.
  2. If miners don’t confirm gold’s lower lows → treat as divergence / early relative strength.

Redeploy timing in uptrends (50-day approach)

  1. In an uptrend, consider adding/redeploying after pullbacks into the 50-day moving average area.
  2. Prefer the strongest names within a correlated group when volatility hits.
  3. Avoid rebuilding exposure at “max drawdown” moments for already-correlated AI baskets.

COT/positioning gating (gold)

  1. Use COT oscillator interpretation (traders vs commercials).
  2. Wait for COT change for “system permission” to long gold.
  3. Cross-check with gold ETF outflows as supportive contrarian risk/positioning input.

Key numbers / levels / explicit metrics mentioned

  • Gold levels (range): ~497 down to ~369
  • Gold timing: “topped” on the day a person (“Wars”) was appointed (exact date not provided)
  • COT oversold logic: “least long in six months” and compared to February of 25
    • readings improved since then
    • data delayed on July 4th
  • AI performance scale: “700, 800, 900%”
  • Rebalancing / rotation timeline notes:
    • “heart of summer trading”
    • “tail end of last week”
    • debate about the “back half of the summer rally”
  • AI technical timing:
    • pullbacks bringing names toward the 50-day
    • SMH around/near the 50-day
    • AMD near all-time closing highs (“one closing session away”)
  • Market-cap fact: NVIDIA ~ $5T

Disclosures / disclaimers

  • No explicit “not financial advice” line appears in the provided subtitles.
  • Personal uncertainty emphasized (e.g., “I don’t know the future any better than anybody else,” plus “take with a grain of salt”).
  • System constraint disclosure:
    • gold long requires COT to change “this week” before Jason’s system allows it.

Presenters / sources (mentioned)

  • Matt Russo (co-host)
  • Jason Shapiro (co-host)
  • Donald Trump (referenced regarding appointments and rate-cut expectations)
  • Soros (mentioned via “reflexive” market concept; no further details provided)

Original video