Video summary

The AI Fund Blowup: Was This the Perfect Bottom?

Main summary

Key takeaways

Finance

Finance-focused Summary (Markets, Investing, Risk, Performance)

What drove the “AI fund blowup” / liquidation move (Theme: Leverage + Positioning)

The speakers frame the AI drawdown as less of a fundamental AI “collapse” and more of a liquidity/positioning event.

Key points:

  • South Korea and a “3x leveraged ETF epidemic” are cited as early sparks.
  • Full-margin retail behavior is described as part of what accelerated the move.
  • A highly leveraged situational-awareness-style fund is reportedly at around ~400% exposure, which then forced liquidations.
  • The analogy is LTCM (1998):
    • LTCM’s thesis could have been “correct,” but extreme leverage + crowded positioning prevented an orderly unwind.
    • Forced exits happen before normal valuation/outlook reasoning can play out.

Core idea: when leverage and crowding dominate, liquidation can force selling regardless of “what should be true.”

“The Tape” signal (Positioning / Sentiment / Market Action framework)

One presenter outlines a consistent decision framework for trading setups:

  • Positioning
  • Sentiment
  • Market action (“the tape”)

Rule of thumb:

  • You generally need all three aligned—one alone isn’t enough.
  • They argue that on the AI/semis complex, the tape was already “bad” even before the blowup, including repeated “news failure” style earnings reactions.

Specific market/stock events linked to the top (AI/Semis “earnings failure” sequence)

They attribute weakness to a chain of failures on the AI/semiconductor complex:

Noted examples:

  • MU (Micron): described as a “news failure” in mid-June; better selling/position management could have improved outcomes.
  • Korea-related episode: a “news failure day” is described as the top for Korea.
  • Subsequent earnings failures mentioned:
    • Samsung earnings failure
    • Hynix (Heinix) earnings failure
    • Google earnings failure

Main takeaway:

  • Once repeated failures started, it implied more sellers than buyers, i.e., trend deterioration.

Why the “perfect bottom” was hard to act on (and why risk control dominates)

They describe the blowup resolution as creating a “perfect bottom” because:

  1. The catalyst ended when the leveraged portfolio was closed/taken over.
  2. Price moved too fast to buy mid-crash (missing entries at prior closes).
  3. Risk became measurable/known afterward due to liquidation mechanics.

Explicit risk-management rule:

  • If an AI-related stock breaks below the key “Wednesday low”, they say to be out (“done”) and avoid going long until a new bottoming structure forms.

LTCM comparison (NASDAQ behavior):

  • It did not bottom on the liquidation day.
  • There were further lows before a sustained recovery.
  • Therefore, they caution against assuming “bottom = liquidation day.”

Notable numeric / performance references

  • JP Morgan (LTCM-era example):

    • Cited as falling about ~54% in ~1.5 months, then later recovering to all-time highs about ~2.5 months later.
    • Used to argue that liquidity events can reverse if fundamentals aren’t impaired.

  • Cisco recovery example (after 1998):

    • Dropped to roughly $12–$13, later reaching around $30.
    • Mentions a rapid ~3-day move (~25%+ ).
    • Emphasizes confirmation when price reclaimed a reversal level.

  • AI meltdown context:

    • After the key day, many stocks rallied ~20–25% by the next morning, making disciplined entries difficult.

Portfolio / stock selection ideas during or after the liquidation (Relative Strength + Resilience)

They emphasize buying the most resilient names rather than “the ones down the most.”

Selection logic:

  • After a liquidation, stocks that didn’t fall as much (or rebound faster) likely had more underlying demand.
  • “Don’t buy every stock—buy the best merchandise,” i.e., stronger demand and relative strength.

Examples to watch / add:

  • Dell (DELL)
  • HPE (Hewlett Packard Enterprise) (HPE)

They claim these were barely touched versus the broader AI meltdown, making them candidates for a resilience/greed-side vs risk-side framing.

Macro / risk: what to watch next is credit spreads and rates (not “AI is fake”)

They suggest the biggest risk to AI capex and model buildout is financing conditions, specifically:

  • Credit spreads widening in AI-adjacent companies/debt
  • Interest rates rising (described as “blowing out” outside the immediate day)

Concept:

  • Even if compute demand grows, the domino risk is whether cash flows/spreads allow continued funding.

Proxy mentioned:

  • JNK ETF: used as a “risk measure” (they note junk/low-quality credit wasn’t fundamentally broken and was “back up” after a bad day).

Market breadth / internal health indicator

They mention using market internals:

  • NYSE (“NYS”): ratio of 52-week highs vs 52-week lows
    • Green implies more highs than lows.
  • Interpretation:
    • Despite the “blood bath,” they see no broad fundamental breakdown, consistent with a liquidity event.

Currency / carry-trade positioning note (Yen intervention)

They discuss yen intervention and argue it likely won’t re-trigger a crash because:

  • Everyone already expects the relationship (yen carry trade → stock crash).
  • Therefore, the edge is gone; it becomes about positioning and whether the crowd is already exposed.

Process rule:

  • Don’t add too many variables—keep it simplified and rely on clear “lines in the sand” risk rules.

Instruments / Tickers Explicitly Mentioned

ETFs / Credit

  • JNK

Equities (Semis / AI-adjacent and others)

  • MU (Micron)
  • Samsung (ticker not given)
  • Hynix (ticker not given)
  • Google (Alphabet; ticker not given)
  • Cisco (CSCO mentioned in context)
  • NBIS (ticker not expanded)
  • MRVL (Marvell)
  • Dell (DELL)
  • HPE (Hewlett Packard Enterprise)
  • JP Morgan (ticker not stated)
  • Aluminina (likely an aluminum-related reference; ticker not given)

Indices / Markets

  • NASDAQ
  • S&P 500
  • Dow
  • DAX

Currencies / FX

  • Yen (JPY) (via “yen intervention” discussion)
  • Canadian dollar (ticker not specified)
  • Also referenced by name: Aussie, peso, New Zealand, Swiss franc, Brazilian real, euro, British

Step-by-Step / Methodology Elements Explicitly Shared

Trade qualification framework (3-part)

A setup is only acted on when:

  1. Positioning
  2. Sentiment
  3. Market action (“tape”)

Only act when all three align.

Risk management rule (AI names)

  • Avoid longs once price breaks the key “Wednesday low.”
  • Wait for a new bottoming structure / tone change.

Relative strength selection (post-liquidation)

  • Prefer stocks showing less drawdown and faster rebound (demand resilience), not necessarily the “worst performers.”

Disclosures / Disclaimers

  • No explicit “not financial advice” disclaimer appears in the provided subtitles.

Presenters / Sources

  • Matt (referred to as “Matt” in multiple speaker sections)
  • Jason Shapiro

Original video