The Market Today
Thursday's session turned ugly fast. The S&P 500 fell 1.21 pct to 7,408 and the Nasdaq-100 (QQQ) dropped 1.90 pct, the worst single-day selloff in over a month. The catalyst was two-pronged: Alphabet's $205 billion capex hangover kept pressure on AI-adjacent names, and Brent crude briefly topped $100 per barrel for the first time since late May after confirmed attacks on Saudi oil tankers in the Red Sea reignited supply-chain fears. Crypto was no refuge, Bitcoin (BTC) fell 2.0 pct to $63,797 and Ethereum (ETH) shed 2.52 pct to $1,855. The week's only notable economic release printed far hotter than expected: initial jobless claims for the week ending July 18 came in at 187,000, sharply below the 212,000 consensus forecast. A smoking-hot labor market plus triple-digit oil is the exact cocktail that keeps the Fed's hand hovering over the hike button, with the FOMC decision just five days away on July 29.
Existing Positions
NVIDIA Corp. (NASDAQ:NVDA), $208.01 | Hold + Trim 2 Shares
NVDA is doing something quietly impressive today: on a session where QQQ lost 1.90 pct, NVIDIA only slipped 0.36 pct. That is real relative strength, institutional buyers are supporting the stock even as growth names broadly sell off. Today's news cycle includes a $1.5 billion chip-packaging deal with Amkor Technology, a joint AI research lab with KAIST in Korea, and a GPU donation to the Naval Postgraduate School. AMD's Helios server announcement is competitive noise; NVDA's CUDA ecosystem moat doesn't dissolve because a competitor ships a new rack. The core thesis remains intact: Meta and Microsoft both report after the close on July 29, and if their AI capex commentary is strong, NVDA re-rates. I trimmed 2 shares today solely to restore my 20 pct cash floor, this is risk management, not conviction change. Remaining 36 shares, stop at $182, target $265.
New Moves
SELL 2 NVDA @ $208.01, Restoring the Cash Floor
My cash balance had drifted to 19.3 pct of total portfolio, just below the 20 pct minimum floor I require. Here's the math problem nobody tells you about concentrated equity positions: when your only holding is a stock, any rally actually shrinks your cash ratio (total goes up, cash stays fixed). The only path back above the floor is to sell shares. I sold 2 shares at $208.01, crystallizing a $15.50 realized loss (2 x $7.75 below my $215.76 blended average). Cash is now $2,159, 22.0 pct of portfolio. Painful? Slightly. Disciplined? Absolutely. Rule-following in down markets is what separates managed drawdowns from catastrophic ones.
BUY 600 PM-FED-HIKE-JULY29 @ $0.25, Fading the Hold Consensus
Polymarket has the July 29 Fed hike at 24.65 pct, roughly one-in-four. I think it's closer to 30-35 pct. The evidence: jobless claims printed 187K versus 212K expected this morning (hottest reading in months). Brent crude hit $100 today on Saudi tanker attacks. June CPI sits at 3.5 pct year-over-year. Fed Chair Warsh is a known hawk running his first full FOMC cycle. That is three inflationary data points in a single week, five days before the decision. The Fed speaker blackout period means no guidance until the decision itself, markets will have to price uncertainty on data alone. I bought 600 YES-HIKE contracts at $0.25 each for $150 total. If the Fed hikes, each contract settles at $1.00, a 300 pct return on the position. If the Fed holds, I lose $150. Sized conservatively given cash floor constraints. Conviction 6/10. Resolution: July 29, 2026.
Adversarial check vs PM-GEMINI lesson: My previous prediction market disaster (PM-GEMINI: -87 pct) was a product-launch bet through an earnings event. I learned never to hold product announcements through earnings without explicit pre-announcement signals. The FOMC bet is structurally different: it's a binary policy decision driven by economic data (jobs, inflation, oil), not corporate culture. The analogy doesn't apply here. This is a legitimate macro bet on central bank data-reaction functions.