BTC’s face-melting rally is framed as a debasement-trade squeeze driven by Fed Treasury buybacks, not a clean-cycle confirmation: the key test remains a break of prior highs and multiple weekly closes back above the 50-week moving average, with June lows called the cycle low but a retracement still expected. At the same time, the macro setup is cast as a dangerous Treasury-bond experiment around $40 trillion of debt and $1.37 trillion in annual interest, with Scott Bessent’s buybacks, rate-suppression and forward-guidance fight against the Fed potentially worsening the long end rather than fixing it. The AI complex got a short-term relief bid from NVDA’s blowout quarter, but the video ties that to circular financing, an OpenAI/Anthropic valuation bubble, and a possible liquidity pullout from crypto if the IPO window opens. The back half shifts to financial repression, hard assets and the Clarity Act vote, with gold, BTC and real assets favored while cash and long-dated bonds are the assets to avoid.
BTC odds that the lows are in improved sharply after the rally, but the bear-market nail in the coffin is taking out prior highs; historically, multiple weekly closes back above the 50-week moving average mean the bear market is over.
BTC could still be a 2018/2015-style rally that reaches near the prior high without fully breaking it, so the move is not yet a cycle-confirmation by itself.
BTC indicators including supply in profit/loss and weekly and monthly RSI fully reset, but realized price, balance price and the MBRVZ score did not look fully reset.
Confidence that the BTC low is in requires the next pullback, a higher low and acceptance above the prior high; until then, DCA through the second half of the midterm year is the default BTC play.
Before the rally, more than 50% of BTC supply was in loss, long-term holders were accumulating, and realized P&L sat at a 43-month low.
The BTC melt-up was driven by Fed Treasury buybacks, a debasement trade that also catalyzed gold; the move triggered a historic short squeeze and large short liquidations.
Last week saw the highest BTC ETF accumulation since October last year, with spot ETF buying still adding about $3 billion in August so far under the short squeeze.
The June lows were labeled the lows for this BTC cycle, but a retracement is still expected because profit-taking should follow a vertical move.
Trump said the US government could buy BTC, but that was treated as something to take with a grain of salt; a separate possible BTC headwind may be easing because the first quantum-resistant BTC transaction was apparently executed successfully.
Scott Bessent’s Treasury buyback move did not lower 30-year yields for long; yields round-tripped almost immediately, leaving the $40 trillion debt load and $1.37 trillion in yearly interest as the structural problem.
Bond vigilantes could punish any sign that Treasury is panicking by selling harder and forcing higher yields; bond fund managers and Stanley Druckenmiller’s camp called the intervention ill-advised because it interferes in the Treasury market and encourages more debt issuance.
The Treasury is issuing debt to buy back the long end, effectively using T-bills to fund the short-dated side in an Operation Twist-like move on the curve, while Bessent acts like a hedge fund manager using the whole Treasury general account.
The market is reacting to Bessent correctly, but the administration’s actions do not fix why the long end rose and could even be inflationary; the long-end move is framed as the market pricing in inflation that will be allowed to run hot.
The Fed side of the policy fight is being set up as a clash between Bessent and ‘Walsh’: Walsh wants no forward guidance and prefers leaning on the long end to tamp down inflation, while Bessent is speaking openly and trying to suppress the yield he thinks should prevail.
Financial repression favors real assets — commodities, energy, infrastructure and farmland — while cash and long-dated bonds are the main assets to avoid; shorter-duration bonds or TIPS fit better if fixed income must be held.
Energy has been one of the better investments this year, cash holdings are at record lows, and the default long-term approach remains buying index funds monthly, though the domestic/international split has shifted from roughly 2/3-1/3 to about 50/50 because of more uncertainty.
NVDA reported $96.2 billion in revenue, beat expectations and grew more than 100% YoY; next-quarter guidance of $108 billion was slightly light versus estimates and sold off first, then the CFO’s next-year revenue guide of up 70% vs 45% analyst expectations pushed the stock higher again.
AI stocks broadly rallied after NVDA’s report, but the complex still has circular-financing questions: a $500 billion securitized data-center package allegedly backstops debt so customers can buy NVDA chips, and NVDA may also be extending invoice terms to finance customers.
OpenAI is expected to stay private until 2027 with a targeted $1 trillion valuation, while Anthropic may IPO within the next month or two aiming for a $2 trillion listing and claiming a $30 trillion revenue TAM; that IPO could pull liquidity out of crypto and other markets back into AI.
OpenAI executives have been quitting, Claude has been criticized for underperforming, and Chinese models are increasingly seen as matching the task for a fraction of the price, setting up a potential AI unwind that could be a bloodbath when the music stops.
Mktar says the September 15 Clarity Act closure vote has a high probability of failure and is priced to fail, which could trigger a selloff; if it passes, the catalyst would matter more for altcoins than BTC.
Druckenmiller’s Wall Street Journal piece attacking Bessent’s bond strategy appeared AI-written, and he admitted, ‘I write everything using AI now’; AI ghost-writing is spreading among executives and public figures, but the name on the piece still makes it the author’s responsibility.
AI is becoming the new subeditor and a useful tool for internal memos, code and legal work, but using it for public opinion columns is a lower-status use of a privileged platform; if today’s WSJ increasingly runs on AI, AI will train on AI-written text in a doom loop.