Daily Brief Archive: July 24, 2026 — U.S. & Hong Kong Stock Analysis

Technicals — The -14.5% close at $319.69 is the lowest in 11 months, just 6.8% above the 52-week low of $297.82; RSI(14) at 29.1 and RSI(7) at 17.6 are deeply oversold, but a breakdown candle on 3.6x volume is not a bottoming pattern. The 20-day at $391.8 is now distant resistance; $297.8-300 is the last technical line, below which the 2025 gap zone opens

Fundamentals — Q2 revenue of $28.24B (+26%) beat, but adjusted EPS of $0.33 missed consensus by 39%; auto gross margin ex-credits fell from 19.2% to 16.3%, and regulatory credits collapsed 67% YoY to $146M. Capex hit $5.79B (+142% YoY) with FCF at -$1.09B, guided to stay negative for several quarters — the growth is real, and so is the profit being devoured by AI and robotics spending

News — The 7/22 after-hours report triggered Thursday's crash, erasing over $140B in market cap in a day. All the highlights are long-dated: 1.48M FSD subscriptions (+56%), 55% of North American deliveries activating FSD, robotaxi in 7 metros with 2.5M paid miles, Cybercab in production at Giga Texas, Optimus first builds starting Q3. Musk's line — customers are 'buying FSD with a car thrown in' — but this quarter, the free car isn't making money

Technicals — The $100.23 close (-2.3%) is a pre-earnings coordinate — after hours it traded near $110. July had seen a 37% drawdown from the $142.35 peak to $89.65, with front-running money in for +8% the two days before the print. Post-gap, resistance sits at $115-121 (prior congestion); a low-volume retest of $103-106 (half-gap) would repair the structure

Fundamentals — Q2 revenue $16.1B (+25%) with non-GAAP EPS of $0.42 doubling estimates; DCAI +59%, Foundry +31%; the $11B GAAP loss is one-off items against $2.2B non-GAAP net income. Gross margin 41.8% (+12.1pp), 2026 capex raised to $20B — with 18A yields at 85% and Apple/Microsoft confirmed as design partners, the foundry story finally has hard numbers

News — The 7/23 after-hours print blew past estimates, after a dense run-up of catalysts: the Xeon 6700P memory-speed update, deeper Google Cloud ties, rumors of the external 18A CSP order; 14A work with Tesla's Austin 'Terafab' slated to supply Tesla/SpaceX/xAI. Restructuring continues (headcount down from 125K to 82.3K). With the US government holding 10% and the stock +170% YTD, options priced ±12% — this time the move resolved up

Technicals — The -22% close at $5.76 on 4.2x volume marks generational lows; -62.9% on the month, -80.5% off the high, RSI at 20.3 extreme oversold. Technicals degrade in low-price territory: moves are event- and liquidity-driven now, and the marginal odds of staying short have deteriorated

Fundamentals — The board voted on 7/20 to suspend the quarterly dividend (announced after hours 7/22) to prioritize deleveraging — the yield-support thesis is gone; the chief accounting officer departs 8/28; and after the $3.7B Getty merger formally died on 7/7, it's back to standalone life. The long thesis — AI-generated content devouring stock-media demand — is intact, but the valuation has moved from recession pricing to survival pricing

News — The trigger was the 7/22 after-hours double announcement — dividend suspension plus the CAO's resignation (which the company says involved no disagreement). Needham kept its Buy but cut $30 to $25; the $28.85 'consensus' target rests on a single analyst and is effectively meaningless

Technicals — The +0.9% close at $56.00 caps a +32.9% month driven entirely by the bid; RSI 75.1 is overheated, with the stock trading at a ~7.4% skepticism discount to the $60.50 offer. Event-driven names have no real technical levels: the range is $60.50-70 (deal-price bargaining zone) above, $47 (the bear script) below

Fundamentals — Deal dynamics dwarf the fundamentals: the consortium bid $60.50 on 7/15 (28% premium), the board rejected it on 7/20 as undervaluing the turnaround — hiring Goldman and Evercore and seeking near $70 — but notably did not reject a sale itself. Michael Burry publicly called the bid 'simply too low'; Polymarket at one point priced an 82% deal probability. The 7/28 pre-market Q2 print is the next variable

News — The consortium wants a deal by end-July — that's next week. Three scripts: a bumped bid closing at $65-70, a close near $60.50 (+8%), or talks collapsing (back to $47, -16%). The 7/28 earnings are negotiating leverage: a strong print arms the board's ask, a weak one undercuts it — earnings and the deal are levers on each other

Technicals — The -4.6% close at $120.04 predated the news — the DoD contract hit after hours, and the stock bounced about +3% on it post-close. Down -29.5% on the month and -65.3% off the high with RSI at 29.6; the 7/14 low of $113 is the double-bottom watch level, $135-140 the overhead gap. Volume at 0.8x says selling is exhausting — but buyers haven't shown up either

Fundamentals — The bull and bear ledgers are irreconcilable. Bears: FY26 FCF of -$23.7B, net debt of $97.6B (+20%), FY27 capex guided up to $95B, the vice chairman selling in June near the highs — and a backlog concentrated in a few customers led by OpenAI, whose IPO is reportedly delayed. Bulls: RPO of $638B (+363%) exceeding Microsoft's, FY2027 revenue guided to $90B, and a forward P/E around 16x

News — The 10-year, up-to-$7B DoD enterprise software agreement (a license consolidation that saves money, not new cloud revenue) was announced after hours on 7/23, lifting the stock about +3% post-close; the -4.6% session predated it. CLSA initiated at just a Hold, $145: 'the AI story isn't broken, but the build-out costs outweigh the balance sheet'; Piper Sandler is holding Overweight at $225 against the tape

Technicals — The -10.7% break at $170.42 on 2.5x volume sliced through a months-long shelf, now -34.8% off the 52-week high; RSI 36 isn't even oversold yet. Support sits at $160-165 (the 2024 base); any bounce first tests $181 (20-day, lower gap edge). Structural repair takes weeks

Fundamentals — Q2 EPS of $2.99 (a $0.40 beat), service revenue +9%, and a raised FY free-cash-flow guide — the profit engine is fine. What the market sold is the growth curve: Q2 postpaid net account adds of 277K fell 13% YoY, Q3 is guided to just ~250K versus 304K expected, and management concedes the premium-plan migration will suppress adds and raise churn risk near term

News — New CEO Srini Gopalan's second full quarter delivered the 'profit beat, growth miss' combo. The Street split: Wells Fargo's mid-July initiation at Equal Weight with a $170 target now sits right at spot, Seeking Alpha bulls called the reaction overdone, and GuruFocus models $226 of intrinsic value. Management waved off Starlink's satellite broadband as a near-term threat — but the premium-multiple rerating has already begun

Technicals — The +9.9% surge to HK$40.5 on 2.3x volume is one big candle against a -29.6% month, still -27% below the 50-day (HK$55.8); RSI 35.9 has only just left oversold. The July low at HK$35 is key support; HK$44-46 is dense overhead supply — a reversal needs both a held retest and a second volume push

Fundamentals — The H1 alert guides net profit to RMB 3.65-4.6B, swinging from a 530M loss, on higher lithium prices, own-resource ramp, storage-battery volume and PLS share disposal gains; solid-state work advances on two tracks (500Wh/kg in pilot production). Tianqi pre-announced a 33-49x jump the same day — the sector's interims collectively prove the cycle trough has passed; the argument is only about the slope of the recovery

News — On 7/23 A-share lithium names went limit-up in batches (Shengxin, Chuanneng, Guocheng), with H-share torque amplifying Ganfeng to +9.9%; the catalysts were lithium carbonate futures +5% above RMB 147K/t, a Goldman note (2020-21 cycle analogy, demand recovery in 2026 with Q4 tightening) and Citi's 250K Q3 call. Supply: weekly output -3.3% MoM in July. Demand: H1 storage-cell shipments up 80%+ YoY

Technicals — The +1.2% close at HK$199.3 sits below the mid-July low of HK$209.2 — lower lows are direct evidence the trend hasn't ended; RSI 31.7 is deeply oversold and 0.3x volume says the selling is fading. HK$209-230 is fresh overhead supply; below, every round number is psychology, not structure — the stock has no trading history down here

Fundamentals — January's listing was the peak: HK$165 issue price, +109% day one, HK$410B at the top. Then three blows: June's M3 launch with a shift to token-based billing sparked developer-churn fears (-15% that day), the 7/9 unlock freed 48.9% of shares, and the sector's cash-burn question looms (domestic foundation models losing ~RMB 18B a year collectively). In the 'top five' landscape it sells value-for-money multimodality against DeepSeek, Kimi and Doubao — all backed by deep-pocketed patrons, while MiniMax must raise from the market. That is its soft underbelly

News — Founder Yan Junjie announced a zero salary to shore up confidence; after the unlock flush, the banks lined up bullish in rare unison — Goldman HK$860, Citi 533, BofA 500, UBS and CICC at Buy, JPMorgan alone Neutral (1,100 cut to 400, then to 300 on 7/23). The chasm between those targets and a HK$199 print is the stock in one image: a war between narrative and float. It bounced 15% in a day on 7/15 over a next-gen video model tease — catalyst sensitivity is extreme