Chapters
00:00 Intro
05:31 FOMC Day: 637, the VIX & a Demand Zone
07:27 $NOW's Strength & $IBM's Cup and Handle
12:17 QQQs: Target Hit, Max Pain & 700
19:12 $SMH Measured Move & Fibs
20:59 A 2021 Analog? Breadth, Rate Hikes & Gold
27:04 Gold, $COST & Copper
29:33 $SNDK & SK Hynix
33:48 $LRCX at the 200-Day
36:18 What Would Break the Chip Thesis?
41:19 GP's Chip Reset: 539, 513 or 420
49:19 SpaceX, the Lockup & the QQQs
57:23 Gap Fills: $NFLX, $GOOGL & $IBM
58:34 The Yahoo Story: Taking Profits Without Giving Up
Market Hours #196 went out live on Wednesday, July 29 — FOMC day, with Chair Warsh speaking after 2 p.m. ET — with Ben, Nick and GP. The market had been sold hard into the Fed, Micron was swinging wildly in overnight trading, and the QQQs had hit Nick's measured-move target that morning. The crew covered whether this was a washout or the start of a V, gold as the leading indicator, where the chip names reset, software and AI moving together, SpaceX's lockup, and a story about taking profits without giving up on a stock.
Nick's setups, levels and trades
Nick said he was more bullish than the day before. On the QQQs, the conservative target he'd posted on StockTwits was hit almost exactly that morning around 660–665. He'd expected a lift into the Fed because he didn't think people would want to carry shorts into it. They had built a descending wedge, and he was leaning toward a V out of it. Part of him would rather see the QQQs dice the 200-day for a complete washout, but if he were coming from cash, he said, this is where you'd start taking a stab. If they wanted max pain, he thought the magnet was the high and low of the high-volume November 20, 2025 candle, which would undercut the 630 level that many on FinTwit were waiting to buy. The big steps higher are reclaiming 685 and then 700, which would also break the wedge. "We're on week 13," he said; he hopes it resolves up but is keeping an open mind.
On the SMH, the measured move to the downside technically completes at 490, and the 61.8% retracement drawn from the April low sits at 480. He likes to draw Fibonacci levels both ways and look for confluence, and when a measured move hits he expects a countertrend move as traders take gains. He also pointed to the trick Cookie shared: when price takes out the prior month's high, it usually goes after the 10-period moving average, which on the SMH was around 450–460. From Ben's Fibonacci levels, he had the 61.8% at 479, and a trip to the 78.6% would likely fill the gap. If chips churn for months, he thinks the IGV could bounce, retesting at least 100.
On gold, he likes the big descending wedge and the undercut of the 200-day, then using the 200-day as the neckline. He used Costco as an example of the pattern, a setup an old iBankCoin trader, Mike Lee, loved: price gapping over a major moving average after a big decline. On GLD, above 385 is where you start paying attention. Monthly, gold was sitting on its 12- and 18-month moving averages after a massive expansion, pulling in on low volume. He also likes copper through FCX and SCCO. SanDisk had been the momentum leader and hadn't tested its 200-day once in the run; he'd been eyeing about 750, with the SanDisk and Micron reports over the next two weeks as big tells for growth and momentum.
On LRCX, it was in line with the SMH and pulling into the 200-day, and 236–237 is the high-volume candle on the daily if they really want to beat it. Weinstein-style trend traders, he said, would be "juiced up" over names pulling into the 30- and 40-week moving averages. On SpaceX, the measured move from about 215 to 150, then the bounce to 175, completed around 105–110. He'd look for a short-term trade back to 130–135 or the declining 20-day, maybe 140–150. The prettiest scenario for bulls is a bounce, then the market absorbs the lockup with a minimal pull-in — the first ABCD leg up. He thinks space is the new frontier long term, but "timing is important more than time in." On earnings reactions, he wants beaten-down names bought up quickly: Netflix had filled its gap, Google was close and IBM was getting there.
His closing advice: stay small and nimble, and when you exit a name in a big base, set alerts at the pivot high and at a stink bid near the 200-day, then ignore it until one triggers.
Ben's setups, levels and trades
Ben said he'd put money back into ServiceNow on its earnings day, something he never does, and he was glad he did. He's been a semis guy as long as the crew has known him. His two buckets are AMD plus Nvidia, and Broadcom, which builds custom chips for the hyperscalers — if one side gets hit, he figures the other benefits. For him to call the SMH run over, technology would have to cannibalize itself or chips would have to stop mattering, and the main risk he sees is geopolitical — China taking Taiwan and TSM — since there aren't enough foundries to absorb that. He noted Micron is sold out through 2027. "You don't have to be the first one to exit. You just don't want to be the last one."
He prefers SK Hynix over SanDisk or Micron and pulled up the Korean listing (000660) instead of the Nasdaq ADR to see the full history. It had filled a gap through a thin volume-profile pocket, and a big buy candle suggested support, but he was waiting until after FOMC to buy. He also asked about Lam Research, with his area of interest at about 233–234, where the 200-day and 50-week line up.
On SpaceX, six to eight weeks after its IPO it was lower while Microsoft was higher — his "I drink your milkshake" argument about IPOs pulling capital from the rest of the market. He pointed out SpaceX was already about 1% of the QQQs, roughly the 24th-largest holding, so locked-up insiders have exit liquidity and will be selling from far lower prices. Index buying may absorb it, but the capital has to come from somewhere.
He closed with a story from the "Einstein of Wall Street," a longtime NYSE floor trader. During the dot-com bubble, everyone rode Yahoo from $3 to $600 and back to $3, while one calm trader bought and sold it over and over on the way up and made $20 million. The lesson for Ben, who tends to hold and accumulate, was that you don't have to give up your thesis to take profits. He gave his own example: after selling Marvell from a $17 basis, he had plenty of chances to buy back before it topped in the 300s.
GP's setups, levels and trades
GP said he got into TQQQ August 70 calls. As he saw it, one of two things happens: either the market collapses and tests the 637 zone, or it had already de-risked into FOMC through the "murder algo" selling all week. The VIX finally popped over 20 and above its 200-day, which he called a demand zone like March. If nothing bad materializes, the asymmetric upside could be huge because people are out of position — so he took the shot with time, not same-day calls. He said ServiceNow's strength was unprecedented and assumed its buyback, roughly 10% of its market cap, had kicked in. IBM looked like an amazing cup and handle right at the gap-down level and had just tagged its 2013 highs after 13 years of going nowhere. He called the first candle "the oh-I-was-wrong candle" before the "oh-my-God-I-was-right candle." He was more bullish because of the destruction into FOMC, though he said nobody knows yet how Warsh behaves after only one meeting.
He pushed back on a 2021 top analog: back then breadth was collapsing and only a handful of mega caps held the market up, while now the advance-decline lines were near 52-week highs — "weakness begets weakness," not strength. The 2022 bottom came when the Fed finally got aggressive with a three-quarter-point hike in June, and rates now are mostly stable. His leading indicator has been gold reclaiming its March lows near 4,100: if the NQ breaks out while gold breaks down, the NQ breakout fails, and vice versa. On the day, gold bottomed and started leading well before the NQ did.
On the SMH, he said the 50% retrace of the breakout and heavy dark-pool activity sat around 539–541, the levels to reclaim — which he thought could happen by the end of the week. A full retrace back to 420 was possible without ending the move. His view was that Q3 is a chop-fest for the chip names that ran hardest — AMD, Intel, SanDisk and Micron — likely for another month to month and a half, until after August OPEX. The first investor base is taking profits, and the next leg is a slower grind with new buyers; for some, demand shows up at 513, for others at 420. He argued software and AI should rise together because most people use AI through tools like Copilot and ServiceNow. His plan: let the AI names reaccumulate this quarter while software shows relative strength, and then both rise into Q4. Datadog stood out on the day.
On SpaceX he'd hold only a small position, as a sort of hedge to Tesla, and treat it as a trade, not something to marry. He doesn't think the lockup will be as bad as people fear, because employees have had liquidity through private sales, as at Databricks. He nearly bought ASTS on the day but held off because of FOMC, and he likes that it's expected to have satellite mobile service in 2027. He also pointed out that Netflix bottomed a month before the market in 2022 and may be doing the same now.
Subscribe for the weekly live show at YouTube.com/@MarketHoursPod.
This is not financial advice. Market Hours is for education and entertainment only. Always do your own research.
Follow Market Hours: YouTube @MarketHoursPod · X @markethourspod · TikTok @markethourspod · StockTwits @markethourspod