U.S. Pre-Market Brief | 2026-08-07 (ET) Friday

Coverage window: 2026-08-06 (Thursday) 16:00 ET regular-session close → 2026-08-07 08:34 ET (20:34 Beijing time on August 7). The July nonfarm payrolls report released at 08:30 ET is already included — see Section 9.5. Scope: Thursday after-hours earnings (16:00–20:00 ET), the overnight Asian and European sessions, and this morning's pre-market. All times are marked in Eastern Time (ET). Single-stock pre-market quote snapshots were taken at 08:00–08:10 ET, cross-checked against two sources — the CNBC quote API and stockanalysis.com; index and single-stock closing prices come from yfinance daily bars; verification depth grading for earnings figures: TEAM / NET / TTD / ABNB have been checked line by line against the original 8-K / 10-Q filings on SEC EDGAR (every discrepancy versus media paraphrase is flagged below); DOCS / PUBM / RMD / SPCX come from company IR press releases, earnings calls and media paraphrase, with no line-by-line comparison against the original SEC filings, and are therefore lower confidence, as flagged throughout. Pre-market quotes move continuously; actual opening trades govern.

---

0. Today in One Line

1. Today's protagonist is not any single stock — it is the July nonfarm payrolls print at 08:30 ET, and its reaction function is inverted. The Fed held rates at 3.50%–3.75% on 7/29 by 9–3, and the dissents from three regional Fed presidents (Hammack, Kashkari, Logan) were all calls for a 25bp hike (primary source: Fed statement of 7/29¹). The statement explicitly says inflation "remains above the 2% objective, partly reflecting supply shocks including energy." The current Chair is Kevin Warsh (sworn in 2026-05-22 as the 17th Chair), and the market is pricing one to two hikes this year, not cuts. Therefore today "strong jobs = room to hike opens up = bearish for stocks and bonds" — BofA said bluntly ahead of the data that an in-line payrolls print would mean downside risk to employment has faded, enough to give the Fed room for three hikes over the remainder of this year. This is the exact opposite of the 2024 "bad news is good news" regime, and it is the easiest place to stand on the wrong side today.

   ✅ The data was released at 08:30, and the outcome was the opposite extreme: nonfarm payrolls −23,000 (negative growth), below every mainstream forecast; May–June revised down by a combined 103,000; the unemployment rate fell to 4.1% but only because the labor force participation rate declined. The market read it as "the hike threat has been removed" — 2-year Treasury −5.0bp, dollar −0.34%, equity index futures modestly higher. Full data, market reaction and this brief's honest reconciliation are in Section 9.5.

2. ⚠️ The most important finding in this brief: the names the media lumped together this morning as "the guidance-raise group" differ enormously in the quality of those raises — the only one that genuinely qualifies as a substantive full-year raise is ABNB. After checking the original SEC filings one by one:
   • TEAM strictly speaking did not "raise" anything — FY27 is the first annual guide, so there is nothing to raise or cut; the correct description is an initial guide below consensus: the revenue guide is only 1–2% above consensus, while the non-GAAP EPS guide is about 10% below consensus (implied $5.44 vs consensus $6.01–6.10, −7% year over year).
   • NET did raise, but the magnitude was inflated roughly 6x — the truth is a raise from the company's own May guide of $1.19–1.20 to $1.25–1.26, +5%; the widely quoted "vs consensus $0.96, a 30%+ beat" is an incorrect comparison (see Section 0.5, item 4).
   • DOCS raised its revenue guide by only $6 million (about +0.9%); ⚠️ and there are reports that the midpoint of its full-year adjusted EBITDA guide was cut and that non-GAAP gross margin falls from about 91% to 88% on AI compute costs — if true, that is "revenue up, profit down"; this brief could not obtain the prior value for that item and lists it as pending verification (see Section 0.5, item 20).
   • PUBM's Q3 guide is also 7.3% above consensus (listed alongside to avoid an arbitrary sample).
   • Only ABNB is a substantive dual raise on a full-year basis: the revenue growth bracket was raised from "low-to-mid double digits" to "at least mid teens," the full-year adjusted EBITDA margin from 35% to "at least 35.5%", and "we are raising our full-year outlook" is verbatim from the company's shareholder letter. In other words, this morning's +16% to +90% moves are mostly not buying "earnings upgrades" — they are buying a valuation re-rating after a narrative was falsified. The persistence of those two things is completely different.
3. The rule for earnings-driven trading: the market buys acceleration and sells deceleration — and it sells without mercy. Yesterday DDOG beat on both lines and even raised full-year guidance, and still fell 19.03% in a single day, purely because the Q3 guide implied growth was no longer accelerating; APP −19.66%. That rule continued this morning: TTD −28.52% pre-market, while TEAM/NET/DOCS/PUBM are +31%/+16%/+85%/+31% pre-market. Seeing +85% and −28.5% on the same day inside the same sector is itself the single most important conclusion today.
4. ⚠️ TTD's real detonation point is not revenue, it is margin. Put on a consistent year-over-year basis using the Q3 guide: the revenue guide of $650M implies −12.0% year over year, while the adjusted EBITDA guide of $160M implies −49.5% (versus $317M in the year-ago quarter), with margin collapsing from 42.9% to 24.6% — the profit collapse is roughly 4x the revenue decline. (The other frequently quoted "−3%" is the Q2 actual revenue miss versus consensus, which is not the same basis as the year-over-year figures above and cannot be compared side by side.) Revenue deceleration can be blamed on the macro; margin collapse cannot. Over the same period gross margin fell from 81.1% to 74.2% in two years, and platform operations costs rose +22.1% while revenue rose only +3%.
5. ⚠️ Ad tech is not collapsing as a whole — the demand side is collapsing while the supply side rallies. TTD (demand-side DSP) is −28.52% pre-market, while on the same night PubMatic (supply-side SSP) is +31.31% pre-market, and PUBM's Q3 guide is 7.3% above consensus. Do not extrapolate TTD's crash into "the whole ad market has caved" and short the entire chain — this looks much more like TTD's own share and execution problem.
6. DOCS +84.95% pre-market is not because results were good — it is because it had been priced for death. Revenue grew only 7.3%, adjusted EPS of $0.29 missed by a penny, and the full-year guide was raised by only $6 million. The real reason is the starting point: DOCS fell from $75.12 on 2025-09-29 to yesterday's close of $20.66 (−72.5% from the high), and the market had already fully priced in "killed by medical AI search like OpenEvidence" — and this report proved it is not dying. This is valuation repair plus short covering, not a growth stock re-accelerating.
7. Geopolitics: the Hormuz "peace trade" is still advancing, but nothing has landed. Iran and Oman have reached agreement on transit routes (inbound through the Iranian-controlled channel, outbound through the Omani-controlled channel), but Iran is simultaneously demanding a ban on U.S. and Israeli vessels and compensation from hostile states — this is the core unresolved dispute and cannot be treated as a positive already delivered. Oil rose more than 3% yesterday (energy was yesterday's only bright spot: OXY +4.14%, XOM +2.12%, CVX +1.51%), and gave a little back this morning pre-market (all three about −0.7%), with Brent $81.85 (−0.78%) and WTI $76.66 (−0.82%).
8. Pre-market state (08:10 ET): S&P futures 7,749.00 (+0.18%), Nasdaq 100 futures 29,624.00 (+0.46%), Dow futures 54,046.00 (+0.06%), Russell 2000 futures 3,020.0 (+0.52%) — risk appetite is mildly positive but the magnitude is tiny, which says money is waiting for 08:30. The 10-year Treasury is at 4.662%, the 2-year at 4.224% (2s10s +43.8bp), VIX 15.22, dollar index 99.85, gold $4,380.90 (+1.89%). ⚠️ Gold up 1.89% while oil falls and the dollar is flat — that is classic "inflation hedge / policy-risk hedge" buying, and it diverges from the mild optimism in equity futures.

⏱ Snapshot-basis note (applies throughout): every single-stock pre-market quote in this brief was taken at 08:00–08:10 ET, all of it before the July nonfarm print at 08:30 ET. That data resets all pre-market pricing — no gap% in this brief can be used directly for order entry, and everything must be re-checked against live quotes before the open.

---

0.5 This Brief's Verification Results and Open Items (the cost of honesty)

Four items verified that overturn the media's version (the most valuable part of this brief)

Materially omitted information verified and added (essentially unreported by the media)

Still unresolved / conflicting

---

1. News Overview

---

2. Strongest Themes, Descending

---

3. Single-Stock Strength Ranking (bullish direction, descending by total score)

Pre-market quote snapshot 08:00–08:10 ET, cross-checked against two sources. All of it precedes the 08:30 ET payrolls print.

---

4. Single-Stock Scoring Model (out of 100)

Notes on adjustments versus the first draft (cut after verification):

• TEAM cut from 75 to 65: expectation gap 8→6 (guidance was not raised and the EPS guide is about 10% below consensus), earnings elasticity 13→9 (about 45% of the revenue beat and the above-100% profit flow-through came from pulled-forward term license recognition), persistence 8→6 (reported growth in the second half of FY27 will fall to single digits), risk deduction −12→−14 (earnings downgrades and multiple expansion are happening at the same time).
• NET cut from 74 to 66: expectation gap 7→5 (the "30% beat" is a data error, the real raise is only +5%), moat 13→12 (gross margin −560bp in two years), risk deduction −11→−14 (P/S 46.8x is the 100th percentile of its own history + the 8/15 convertible maturity turns net cash negative).
• ABNB trimmed from 78 to 77 but still ranks first: verification against the original SEC filing confirms it is the only name whose guidance was genuinely raised today (in the company's own words), it has the smallest gap, and its TTM FCF margin of 37% is the strongest on this list; earnings elasticity went 11→10 because the primary documents show net income includes a $77 million one-off tax benefit, revenue growth includes roughly 3–4pp of FX tailwind, and Q3 margin is guided slightly down year over year.

---

5. Detailed Analysis of Top Names

1) ABNB — Airbnb (total 78, priority deep-dive)

• Related news: 08/06 16:05 ET, Q2 2026 earnings. Every figure in this section has been checked line by line against the original SEC 8-K Ex-99.1 shareholder letter²¹. Revenue $3,608M (+17% YoY); net income $816M (versus $642M a year ago); GAAP diluted EPS $1.37; adjusted EBITDA $1,261M (+21%), margin 35% (34% a year ago); free cash flow $1,253M (35% margin), TTM FCF $4.86 billion (37% margin); GBV $27.2 billion (+16%); Nights and Experiences Booked 148.3M (+10%, accelerating from Q1 2026); implied take rate 13.2%, flat versus a year ago. Q2 buybacks $1.1 billion, with $3.4 billion of authorization remaining.
• Guidance (primary source, the basis for this brief's core claim):
  • Q3 2026: revenue $4.69–4.77B, +15%~17% YoY.
  • Full-year 2026: "we now expect year-over-year revenue growth to improve to at least mid teens"; "we now expect to deliver a full-year Adjusted EBITDA Margin of at least 35.5%".
  • The company writes explicitly at the top of its Outlook section: "we are raising our full-year outlook for both revenue growth and Adjusted EBITDA Margin" — "raising full-year guidance" is the company's own wording, not a media interpretation.
• Catalyst logic: after checking the original SEC filings one by one, ABNB is the only name this morning whose full-year guidance was genuinely raised — TEAM's was not raised (its EPS guide is in fact about 10% below consensus), NET's was raised only 5%, and DOCS added only $6 million. In today's "buy acceleration only" market, a real full-year raise is the hardest currency there is. It also ends three consecutive quarters of misses.
• ⚠️ But the primary documents contain three offsetting items the media broadly missed, and they must be read alongside:
  1. The $816M of net income includes a $77 million one-off tax benefit (verbatim from the shareholder letter: "a $77 million benefit recorded in the current period related to recently published guidance impacting prior year taxes"). Ex that item it is about $739M, still up year over year, but the "net income +27%" figure is propped up by a one-off.
  2. FX contributed a meaningful share of growth: Q2 revenue was +17% but only +13% ex-FX; the +15%~17% Q3 guide includes about 3 percentage points of FX tailwind (the company's own words), i.e. organic growth of roughly 12%~14%.
  3. Q3 adjusted EBITDA margin is guided "down slightly" year over year ("Adjusted EBITDA Margin to be down slightly compared to Q3 2025, due to timing of investments"). A full-year margin raise coexists with a next-quarter margin decline — precisely the "implied next-quarter deceleration" structure that got DDOG cut 19% yesterday, only far milder in magnitude.
• Theme and phase: consumer/travel, launch phase.
• Fundamental verification: closed yesterday at $151.64, only −1.0% from the 12-month high of $153.11 (2026-07-28), with a low of $111.54. Net margin 23% (21% a year ago), TTM FCF margin 37% — the strongest cash generation on this list (for contrast: NET's EV/FCF is about 370x).
• Moat: two-sided network effects plus a supply barrier in listings; the combination of moat and liquidity is among the best on this list.
• Pre-market and technicals: +6.97% to $162.21, on track for a 52-week high. The moderate gap is its single biggest advantage — it has not front-loaded the good news in one shot the way TEAM/DOCS have.
• Final judgment: among the names this brief covers, it has the most complete evidence chain and the smallest gap, and it is the only one that withstands the question "was guidance actually raised?" But it is not flawless: the one-off tax benefit, roughly 4pp of FX tailwind, and a Q3 margin guide slightly down year over year mean its "real quality" is below the headline +17% growth. It is also a consumer name, so if 08:30 shows employment weakening, the consumer narrative takes a direct hit — that is its largest risk exposure.

2) NET — Cloudflare (total 66, watch closely)

• Related news: 08/06 16:05 ET, Q2 2026 earnings (SEC 8-K Ex-99.1⁸). Revenue $696.061M (+35.87%), 4.76% above the company's own $664–665M guide (the largest beat in recent years); non-GAAP diluted EPS $0.29 ($0.21 a year ago). Q3 revenue guide $736.0–737.0M; FY26 revenue guide $2,864–2,870M, non-GAAP EPS $1.25–1.26.
• Catalyst logic (the real part): year-over-year revenue growth has trended up over the past six quarters: 26.5% → 27.8% → 30.7% → 33.6% → 33.5% → 35.9%, a cumulative acceleration of about 9.4pp (⚠️ note the fifth quarter went 33.6%→33.5%, marginally flat/lower, so strictly speaking this does not constitute "consecutive acceleration"; this brief's first-draft phrasing of "four consecutive quarters of acceleration" has been corrected). This is still the clearest re-acceleration since 2022, and it is corroborated by three independent metrics, not single-quarter noise: net retention DBNRR jumped from 114% to 120% (the most underrated number in this print), 4,698 customers with annualized revenue >$100K (+26.6%), and U.S. revenue +40.7%. cRPO +35% is in line with revenue growth, with no divergence of "revenue running ahead of bookings." Three straight quarters of beating its own guide by 3–5% mean the assumption that "guidance is a floor" holds up.
• ⚠️ The part that must be discounted (the market narrative this brief's verification overturns):
  1. "FY26 EPS guide $1.25–1.26 vs consensus $0.96, a 30%+ beat" is a false comparison. The company's own guidance history is February $1.11–1.12 → May $1.19–1.20 → August $1.25–1.26, so this raise is only +$0.06 (+5%). If part of the +16% pre-market move was ignited by that wrong number, that part has no fundamental support.
  2. The bill for the second-half margin jump was paid by employees, not by AI revenue: the company announced a roughly 20% workforce reduction on 5/7 and took a $150.7M restructuring charge in Q2. The Q3/Q4 guides imply non-GAAP operating margins of 17.6%/18.3%, versus an actual 13.8% in Q2.
  3. Gross margin has fallen about 560bp in two years (non-GAAP 73.1% vs 78.7% in FY24), driven by third-party technology service fees +$30.1M and other AI-related costs. Management commits only to "stabilizing" it.
  4. SBC is 20.2% of revenue, 1.30x non-GAAP net income.
  5. pay-per-crawl currently generates essentially zero revenue, and the CEO explicitly said the approach to malicious AI crawlers is "block them and charge nothing extra."
• Theme and phase: software/AI infrastructure, crowded phase.
• Valuation (after verification): at the pre-market $330.50, market cap is $117.6B, with P/S (TTM) 46.8x, EV/S (FY26E) 40.7x, non-GAAP P/E (FY26E) 263x, and EV/FCF about 370x. Percentile within its own history: about the 100th (median of 23.4x over a three-plus-year sample, all-time high 47.2x). Peer comparison: DDOG P/S 20.6x, SNOW 21.9x, and both have TTM free cash flow roughly 3.4–3.5x NET's.
• ⚠️ A certain event 8 days out: the 2026 convertible notes mature 8/15, with principal of $1,293.8M settled in cash plus the excess delivered in stock (about 2.8–2.9 million shares, ≈0.8% dilution). Net cash goes from +$869 million to roughly −$400 million (estimated).
• Pre-market and technicals: +16.20% to $330.50, on track for an all-time high (prior high $301.33, 2026-08-04).
• Final judgment: the acceleration is real and the quality metrics are real, but "a big guidance raise" is fake, and the valuation position is the highest on this list. "New high in price" and "new high in valuation" are happening at the same time here. What needs attention is the asymmetry of that valuation: the current P/S of 46.8x is closing in on its own historical extreme of 47.2x, while the median over three-plus years is 23.4x — this brief makes no forecast about future price direction, but readers should recognize that this position leaves extremely little room for error if growth rolls over, and is the most sensitive to rising rates.

3) TEAM — Atlassian (total 65, watch closely)

• Related news: 08/06 16:05 ET, FY2026 Q4 earnings (SEC 8-K 0001650372-26-000031). Revenue $1,766.5M (+28%); cloud revenue $1,213.5M (+30.8%); GAAP diluted EPS $0.55 (−$0.09 a year ago, positive for the first time); non-GAAP EPS $1.87 ($0.98 a year ago). Subscription ARR $6,606M (+23%), RPO $4,817M (+44%).
• ⚠️ The most important unpacking in this brief: both "+28% high growth" and "only 13% in FY27" are illusions; underlying growth is stable at 19–20%.
  • In September 2025 the company announced Data Center end-of-life (EOL) in March 2029. Under ASC 606 the license portion of a term license is recognized upfront at contract inception, and the EOL announcement pushed customers to lock in multi-year contracts early, slamming future revenue into FY26 all at once.
  • FY26 segment base: Cloud $4,410.6M, Data Center $1,830.9M, Marketplace $330.7M, totaling $6,572.3M (DC is 27.9% of the mix). FY27 segment guidance: Cloud +25.5%, Data Center −17.0%, Marketplace +12.0%, which back-solves to $5,535M + $1,520M + $370M = $7,425M, i.e. +12.98%, consistent with the official 13.0%.
  • Data Center's direct drag on total FY27 growth is about −4.7pp (1,830.9 × −17% ÷ 6,572.3); if DC is excluded entirely, Cloud+Marketplace grow +24.6% combined. ⚠️ Note that the 11.6pp gap between 24.6% and 13.0% is caused by two different denominators (the ex-DC base is $4,741M) and must not be read as "DC drags growth down by 11.6 percentage points" — this brief's first draft said exactly that, and it has been corrected.
  • Management's own words: they expect revenue growth to re-accelerate in FY28, and proactively suggested "using Subscription ARR as a cleaner measure."
  • FY25→FY27E two-year CAGR = 19.3%, almost identical to FY25's 19.7%. The FY27 Q1 guide implies +19.0%~+19.7%, which also falls in that range.
• ⚠️ About half the Q4 beat is accounting: measured against the company's own 4/30 guide, revenue came in $108.3M above, of which Cloud contributed 45% and Data Center 45%; management conceded the DC portion was "recognized greater-than-expected term license revenue." Incremental profit / incremental revenue = a 121% flow-through rate — anything above 100% can only be zero-marginal-cost term license. So a substantial share of the big $1.87 EPS beat is likewise an accounting contribution.
• ⚠️ Guidance was not raised at all: the FY27 revenue guide of $7,427M is only 1–2% above consensus; the non-GAAP EPS guide implies about $5.44, roughly 10% below consensus of $6.01–6.10, −7% year over year. The FY27 non-GAAP operating margin guide of 25.0% is below FY26's 30.4% (of which about 4pp is the fading DC accounting windfall and about 3pp is the company deliberately shifting compensation from equity to cash — a move that is directionally shareholder-friendly).
• The genuinely high-quality parts: Cloud guided +25.5% → actual +30.8%, a third straight quarter of acceleration (26%→26%→29%→31%); GAAP positive for the first time; FY26 buybacks of $1.80 billion cut net share count −4%; CEO Cannon-Brookes has committed to start a 10b5-1 plan buying up to $250 million as early as after today's close; short interest is 7.0% of float, requiring 5.2 days to cover.
• ⚠️ Cash flow is the most overlooked blemish in this print: revenue +26% and non-GAAP EPS +59%, but FY26 free cash flow fell −6.8% year over year ($1.319 billion), with FCF margin down 7pp to 20.1%. Accounts receivable rose +63.2% to $1,269.9M and DSO deteriorated from 54 days to 71 days — the inevitable consequence of upfront recognition: revenue is recognized first, cash collected later. SBC of $1.607 billion > FCF of $1.319 billion (122%). The company gave no FY27 FCF target this time.
• Balance sheet: net cash plunged from $1.950 billion to $250 million (DX acquisition −$1.23 billion, buybacks −$1.80 billion), and marketable securities have gone to zero. Solvency risk is not high, but the financial cushion is already thin.
• Moat (the hardest on this list): covers more than 85% of the Fortune 500 and 350,000+ customers; 57,334 customers with Cloud ARR >$10,000, accounting for more than 85% of Cloud ARR. The most damaging evidence against the short thesis: MCP monthly actives passed 1 million with call volume +400% QoQ, while 98% of MCP users were still active in the Jira UI that month, and "MCP adopters expand seats faster and grow ARR 2x faster" — pointing to agents currently being an amplifier of seats rather than a replacement for them. ⚠️ But these are all company-selected, unaudited metrics, Rovo has no separately disclosed revenue, and the FY27 gross margin guide of 86.5% is below FY26's 87.9%, which the company explicitly attributes to "the negative impact of Rovo usage and hosting costs" — AI is currently a net drain on gross margin.
• Valuation: at the pre-market $144.20, market cap is $36.59B, with P/S 5.57x, P/ARR 5.54x, and a forward non-GAAP P/E of ≈26.5x. Historical percentiles diverge sharply: the 13th percentile over the past 4 years, but the 51st percentile over the past 12 months — P/S has already been crushed from ~9.5x to 2.6x over the past year, and the old double-digit P/S belongs to 2021 bubble pricing and cannot serve as an anchor for mean reversion.
• Pre-market and technicals: +30.99% to $144.31; still −68.5% below the 5-year high of $458.13, but +152% off the 4/10 low of $57.15, and already up 28.9% in the month before earnings.
• Final judgment: the fundamentals really are improving (Cloud accelerating, GAAP turning positive, share count shrinking, founder buying), but the price is changing much faster than the fundamentals are. Almost the entire move is multiple expansion: P/S +31% while FY27 revenue estimates go up only 1–2% and EPS estimates need to come down about 10% — earnings downgrades and multiple expansion happening together is the most concentrated combination of valuation risk there is. And FY27 is "the year the numbers look worst," with reported H2 growth falling to single digits.

4) DOCS — Doximity (total 62, watch only)

• Related news: 08/06 16:05 ET, FY2027 Q1 earnings (BusinessWire⁹). Revenue $156.6M (estimate $151.7M, +7.3%); adjusted EPS $0.29 (estimate $0.30, a small miss). The full-year FY27 revenue guide was raised from $664–676M to $671–681M (only +$6 million); the adjusted EBITDA guide is $309–329M (midpoint margin about 47%). Operations: 165 health systems signed as AI customers, AI Search queries +25% QoQ, and workflow-active prescribing physicians +30% year over year.
• Catalyst logic (the one thing this brief most needs to make clear): there is an order-of-magnitude mismatch between +85% and "revenue +7.3%, EPS a penny short, guide raised only $6 million." The explanation is not in this quarter's numbers but in the starting point: DOCS fell from $75.12 on 2025-09-29 to yesterday's close of $20.66 (−72.5%), with a May low of $18.01. The market's prior pricing assumption was "medical AI search will destroy Doximity's ad monetization" — and this report falsified that assumption, so the entire valuation framework was overturned. This is valuation repair from "the death thesis was falsified" plus short covering, not growth acceleration.
• Theme and phase: healthcare AI, oversold-repair launch.
• Fundamentals: an EBITDA margin of about 47% shows the business model itself is still extremely profitable — which is exactly the core evidence that it is "not dying." But +7.3% is objectively slow.
• Moat: the verified-identity U.S. physician network has real network effects, but generative AI is eroding it from the entry point — this quarter only proves the erosion is slower than imagined, which is not the same as the threat being removed.
• Pre-market and technicals: +84.95% ($38.21) at 08:00–08:10 ET, and by 08:37 ET it had reached +88.72% ($38.99). ⚠️ There is also an after-hours quote of +81.46% ($37.49). Note that after-hours (16:00–20:00) and pre-market (04:00–09:30) are two different trading sessions, and comparing prices across sessions cannot be called "pre-market drift" — on the pre-market series actually obtained, the shape is a one-way climb rather than back-and-forth chop. This brief's first draft inferred "the price is extremely unstable" on that basis; the evidence was insufficient and it has been withdrawn. Arguing that pre-market volume is thin would require cumulative pre-market volume and bid-ask spreads, and this brief did not obtain either.
• Final judgment: the logic holds, but the price no longer allows any margin for error. An +85% gap front-loads several quarters of repair in a single move. A textbook case of "even if you understand it, today is not the day to buy."

5) PUBM — PubMatic (total 56, watch only)

• Related news: 08/06 after hours, Q2 2026 earnings. Revenue $78.59M (+10.5%) beat; non-GAAP EPS $0.12 a substantial beat; adjusted EBITDA $19.6M, FCF $13.7M. The Q3 revenue guide midpoint of $76M is 7.3% above consensus. CFO Steve Pantelick announced his retirement, serving through Q1 2027.
• Catalyst logic: its significance is larger than its market cap. On the same night TTD collapsed, a supply-side platform delivered guidance above consensus, directly falsifying the reading that "the ad market has caved" and pinning TTD's problem to company-level share and unit economics.
• Theme and phase: ad tech (supply side), fermentation phase.
• Fundamentals: closed yesterday at $13.48. Revenue is only about $78.59M per quarter, the smallest on this list. The SSP space is highly commoditized, giving it the weakest moat (scored 7/15).
• Pre-market and technicals: +31.31% to $17.70. Small cap + a 31% gap = high liquidity-trap risk, with wide bid-ask spreads.
• Final judgment: its main value is at the information layer — it lets you read the nature of the TTD event — supply-side guidance beat consensus, falsifying the reading that "ad demand has caved wholesale." Its information content as an observation indicator exceeds its tradability as a position (small cap, +36.8% pre-market gap, wide bid-ask spreads).

---

6. Bearish / Avoid List

---

7. Intra-Theme Rankings

Theme 1: SaaS "narrative falsified" re-rating

Theme 2: Ad tech (the split between demand side and supply side)

Theme 3: The geopolitical "peace trade" (energy)

---

8. Opening Verification Signals

Pre-market (before 09:30 ET)

• The primary signal is the 08:30 ET payrolls print itself, not any single stock. Note that the reaction function is inverted: payrolls well above expectations (>150k) with firm hourly earnings → hike pricing heats up → 10Y up + valuation compression in growth, and at that point NET's pre-market gap (P/S at the 100th percentile) is the most fragile. Payrolls well below expectations (<50k) → hike pressure eases but growth worries rise → falling rates favor duration assets, but consumer names (ABNB) come under pressure.
• Watch whether TEAM's and NET's gaps hold after 08:30. These two are the theme leaders, and if they cannot hold, the whole theme fails for the day.
• DOCS moved from +84.95% (08:05) to +88.72% (08:37) pre-market, a one-way climb. Actual volume and bid-ask spreads in the final 15 minutes before the open (not obtained for this brief) are the key evidence for whether it can hold this price.

Intraday (09:30–10:00 ET)

• The dividing line between gap-and-go and gap-fill: for the three extreme gap names TEAM (+31%), DOCS (+85%) and PUBM (+31%), breaking below the pre-market low in the first 30 minutes essentially confirms a gap-fill structure.
• Sector confirmation: if TEAM/NET rally but IGV (the software ETF) does not follow, that says this is a single-stock event rather than a theme, and persistence must be discounted.
• TTD short verification: watch whether a snapback comes on volume. 17.61% short interest means any marginal good news gets amplified; for a stock that has just broken its 52-week low, day one is often not the best entry for a short.

Options sentiment

• Today is a concentrated day for post-earnings IV crush. TEAM, NET, DOCS and TTD are all the day after earnings, so implied volatility will fall sharply — getting the direction right can still lose money on long options.
• VIX is now 15.22, at a low level, meaning the market is pricing payrolls optimistically, and if the data surprises, the room for volatility expansion is asymmetric.

Risk signals

• ⚠️ The divergence between gold +1.89% and equity futures +0.18%: gold is strengthening on an inflation/policy-risk narrative while equities are mildly optimistic. They cannot both be right.
• ⚠️ Breadth was already deteriorating yesterday: on 8/6, 59% of NYSE and 57% of Nasdaq stocks declined while the S&P fell only 0.18% — the index is being screened by a few heavyweights while the internals are already weakening.
• ⚠️ A lone runner: if NET/TEAM rally but semis and mega-cap tech do not follow, that is a textbook "earnings lone runner."
• ⚠️ The Dow ended a five-day winning streak yesterday and pulled back from a record high, on top of the 10Y at a high 4.66% — the divergence between rates and equities remains unresolved.
• ⚠️ NET's 8/15 convertible maturity (8 days out) is a known, certain cash flow and dilution event, and should not be obscured by this morning's optimism.

---

9. Final Conclusions

① The 5 stocks most worth watching today

② The 3 strongest themes today

③ What to avoid today

1. Establishing any directional position before 08:30 — the data was released at 08:30 (payrolls −23,000, a large miss and negative), so this item is void. With hindsight the judgment was correct: the actual print was below even the most pessimistic institutional forecast (Vanguard +18,000), so betting ahead of the data really was a coin flip. New risk points are in Section 9.5.
2. Chasing the extreme gap names (DOCS +85%, TEAM +31%, PUBM +31%) — quotes on thin pre-market volume do not represent absorption capacity after the open.
3. Trading this morning's rally as an "earnings upgrade" for persistence — verified against the original SEC filings, only ABNB of the four had guidance genuinely raised; the rest is valuation re-rating, which is one-off by nature.
4. The three energy names (XOM/CVX/OXY) — direction is entirely determined by the Hormuz negotiations, and one round of risk premium was already realized yesterday.
5. The technical bounce in storage/AI hardware (WDC/SNDK) — the theme is in its ebb phase and the bounce has no fundamental support.
6. Extrapolating TTD's collapse into shorting the whole ad sector — PUBM's 7.3% guidance beat has already falsified that extrapolation.

④ The final one-line judgment

Today is a day when the macro overwhelms everything, and the 08:30 answer landed on the side the market did not expect: July nonfarm payrolls −23,000 (negative growth), below the most pessimistic institutional forecast, with May–June revised down by a combined 103,000; the unemployment rate did fall to 4.1%, but it was bought with a declining labor force participation rate. The market read it as "the hike threat has been removed": 2-year Treasury −5.0bp, dollar −0.34%, equity index futures modestly higher. Note that this brief's framework ("what the market trades is the rate path") holds, but the side I warned about was "strong jobs," and what arrived was the opposite extreme — readers who cut long-duration growth exposure in the direction this brief warned about had the direction wrong (see the reconciliation in Section 9.5). At the single-stock level, this brief's most valuable finding is that the "guidance raise group" the media lumped together shrinks to ABNB alone once checked against the original SEC filings — about half of TEAM's beat came from pulled-forward Data Center accounting recognition and its EPS guide is 10% below consensus, NET's raise was inflated 6x, and DOCS added only $6 million for the full year. This morning's +16% to +90% moves are buying a one-off valuation re-rating from "the narrative was falsified," not a sustainable earnings upgrade — falling rates gave them a tailwind, but the problem of a stretched valuation has not changed one bit; and the clearest bearish story (TTD) is not suitable to chase short today either, given 17.6% short interest and a broken, oversold chart. Conclusion: the rate tailwind makes today's extreme gaps easier to hold, but it does not change the judgment that "the one that has run the most (DOCS, already +89.69%) carries the most risk"; the best risk-reward is still the most restrained mover, ABNB — except that one of its bull points, "resilient consumer demand," was just weakened by this employment report.

---

9.5 ⏰ Post-finalization addendum: the 08:30 ET payrolls print is out — a large miss, and negative

The data in this section is taken directly from the original BLS news release USDL-26-1291⁷ (THE EMPLOYMENT SITUATION — JULY 2026), not from media paraphrase. Written at 08:34 ET.

Actual vs expected

⚠️ Three decisive details that matter more than the headline number:

1. The prior two months were revised down by a combined −103,000: May was revised from +129,000 to +63,000 (−66k) and June from +57,000 to +20,000 (−37k). Combining the current month's −23k with the revisions, the shortfall versus the market's prior understanding is roughly −126k. The average monthly gain over the past 12 months has fallen to +34,000.
2. The drop in the unemployment rate to 4.1% "fell the wrong way": BLS explicitly records a labor force participation rate of 61.4%, down 0.7 percentage points since January, and an employment-population ratio of 58.9%, down 0.5 percentage points over the same period. The decline in the unemployment rate comes mainly from people leaving the labor force rather than from improved employment — that is not a healthy decline. At the same time temporary layoffs rose by 153,000 to 921,000.
3. Structurally it is bleeding in three places: government education, retail, and finance: local government education −50,000, retail −19,000, financial activities −14,000 (a cumulative decline of 121,000 from the May 2025 peak); healthcare +22,000 is still the only pillar, but it is already slower than the +36,000 average pace of the past 12 months.

Market reaction (measured 08:33–08:34 ET)

How to read it: this is a textbook repricing of "a hawkish risk removed" — the 2-year fell more than the 10-year (the front end is most sensitive to the policy path), the dollar weakened, gold held, and equity futures moved modestly higher. The market's first instinct was to read it as the bullish "hike threat removed," not as a bearish "recession alarm."

What this means for this brief's conclusions (an honest reconciliation)

• ✅ The framework that the reaction function "is inverted" was validated, but the direction is the opposite of the side I emphasized. What Section 0 repeatedly warned about was the risk on the "strong jobs → hike pricing → valuation compression" side. What actually arrived was the other extreme: extremely weak jobs → hike pricing unwinds → rates fall. The framework (the market trades the rate path) holds, but the side I weighted did not happen — readers who cut long-duration growth exposure in the direction this brief warned about had the direction wrong. That has to be stated clearly.
• ✅ The biggest risk exposure for NET / TEAM / DOCS (rising rates) is removed for now. This brief's core risk for NET was "P/S at the 100th percentile of its own history, most sensitive to rising rates"; falling rates are in fact a tailwind for all three. As of 08:33, DOCS is at +89.69%, PUBM at +36.80%, and TEAM at +31.61% — the extreme gaps not only failed to fill, they widened. But this does not change the judgment that "the valuation is stretched and chasing is extremely risky" — that is a valuation problem, not a rates problem.
• ⚠️ ABNB's risk exposure went from "potential" to "actual." This brief wrote explicitly that it "is a consumer name, so if 08:30 shows employment weakening, the consumer narrative takes a direct hit" — and that is exactly what happened. Although it actually ticked slightly higher pre-market to +7.17%, negative job growth + falling participation + rising temporary layoffs act directly on travel consumption demand. It remains this brief's top-scoring name, but the "resilient consumer demand" leg of the case has weakened, and today the observation focus should shift from "delivering on earnings" to "the durability of consumer demand."
• ⚠️ The inflation problem was not solved by this data — it got thornier. The Fed's 7/29 statement attributed elevated inflation to supply shocks such as energy, not to wage pull. A weak employment report does nothing to ease energy-driven inflation — it merely removes "the economy is overheating" as a reason to hike. Weak growth + supply-driven inflation = a stagflationary combination, and the Fed is squeezed from both ends. The case made by the three regional Fed presidents who wanted a hike is weakened, but inflation itself has not improved.
• The next key dates: the preliminary benchmark revision for 2026 is released on August 28, a systematic re-estimation of the full year's employment data, and the risk is asymmetric against a backdrop of successive downward revisions; the August employment report comes on September 4; and the FOMC meets September 15–16.

Single-stock pre-market quote update (08:37 ET, versus the 08:00–08:10 snapshot)

How to read it: after the data, the extreme gap names did not fill — they broadly moved higher (PUBM alone added 5.5pp). That is consistent with "falling rates → long-duration assets benefit." The only exception is NET's modest pullback — it is the name at the highest valuation percentile on this list. RMD and TTD kept sinking, which says a rate tailwind cannot rescue names whose guidance has weakened — and that in turn confirms this brief's core judgment that "the market sells deceleration."

⚠️ All pre-market quotes in Sections 3, 5 and 6 were taken at 08:00–08:10 ET, before this data was released; the table above is the 08:37 ET update. Pre-market quotes continue to drift, and every gap% judgment should be re-verified against live quotes before the open.

---

Operations Log (not sent to clients)

Data collection status this run: the main pipeline was fine, but yfinance behaved inconsistently on the sub-agent side.

• yfinance worked in the main process, but batch download() across multiple tickers produced intermittent YFRateLimitError: in one batch pull of 21 tickers, MSFT / ^DJI / ^IXIC failed, and all succeeded after switching to one-by-one Ticker().history(). Conclusion: throttling triggers on request frequency, and splitting into serial single-ticker calls gets around it.
• ⚠️ But two sub-agents reported yfinance being fully blocked on their side (the TEAM agent reported 429s on both query1/query2, purely IP level; the NET agent reported both yfinance and the CNBC quote API returning 429/Access Denied). Three agents pulling data concurrently push this machine's IP over the throttling threshold — next time they should be started staggered, or given an explicit instruction to prefer SEC EDGAR. After the two agents switched on their own to SEC EDGAR + the Nasdaq API + stockanalysis, data quality actually improved, and this backup path should be locked in as the first choice.
• The CNBC quote API was the workhorse for this brief's pre-market data, and it was stable on the main-process side. ⚠️ The | separator must be URL-encoded as %7C, or it returns an empty response and JSON parsing fails (that is exactly how the first attempt failed this run).
• Futures symbols: @ES.1 / @NQ.1 / @YM.1 return None on the CNBC API; the correct symbols are @SP.1 / @ND.1 / @DJ.1, cross-verified as consistent with yfinance's ES=F/NQ=F/YM=F.
• WebFetch failure list: Benzinga (403), indexbox (403), CNBC article pages (403), stocktitan (DNS timeout), businesswire (60s timeout), investors.doximity.com/news-releases (404). The Benzinga and CNBC 403s reproduce reliably, so future runs should prefer search snippets or SEC EDGAR.
• SEC EDGAR's data.sec.gov/submissions/CIK*.json works well, and was used this run to establish that EXPE's 8-K filing date was 8/5 (not 8/6 as the media implied), avoiding misreporting the prior trading day's earnings reaction as a fresh catalyst this morning — this should be locked in as the standard method for verifying earnings dates. The same approach with Archives/edgar/data/<CIK>/<acc>/ directory listings retrieved ABNB's Ex-99.1 shareholder letter, from which the verbatim "raising our full-year outlook" and the three offsetting items the media missed were read directly. WebFetch on SEC pages returns 403; you must use curl with User-Agent: <name> <email>.
• BLS payrolls: bls.gov/news.release/empsit.nr0.htm returns 403 on WebFetch but works with curl plus a UA; after the 08:30 release the URL updated to M07 immediately (previously M06), but search engine indexing clearly lagged (at 08:33 searches still returned only pre-release preview articles). Conclusion: time-critical macro data must be pulled directly from the primary source and cannot wait for search. That is how the full −23,000, the −103,000 revision and the participation-rate detail were obtained within 4 minutes this run.
• Waiting strategy: using a sleep 1800 background task to block data collection until after 08:30, while running sub-agents and drafting in parallel, wasted no waiting time, and this pattern is reusable.
• Assessment of sub-agent output (all three agents hit substantive errors this run, extremely high value):
  • TEAM agent: saw through "FY27 growth of 13%" as an ASC 606 accounting reverse payment from the Data Center EOL, and quantified that 45% of the beat came from pulled-forward term license recognition with a 121% profit flow-through rate; also dug out the CEO's $250 million buying plan in 8-K Item 8.01 (missed by all media).
  • NET agent: overturned the "EPS guide beat by 30%" figure that nearly all media were quoting, proving the real raise was only +5% and that $0.96 was most likely a data vendor mis-tagging the FY25 actual; also dug out the 5/7 20% layoff and the 8/15 convertible maturity.
  • TTD agent: corrected the false premise in my prompt about a "10:1 stock split in 2025" (it was 2021), and pointed out that the real detonation point was the −49.5% EBITDA guide rather than the revenue miss.
  • Lesson: "background numbers" given to agents in the prompt must be marked as unverified, or they will contaminate the agent's judgment — the TTD agent corrected me on its own this run, but that cannot be counted on every time.
• Names with no agent assigned: DOCS / PUBM / RMD were not verified against the original SEC filings, and the confidence difference is disclosed to clients in Section 0.5, item 18. ABNB was additionally verified against SEC Ex-99.1 by the main process before finalization (because it was pushed to the number one slot and could not be staked on unverified data).
• ⚠️ The risk-auditor caught an error type this round that must be remembered: reading only the press release and not the earnings call. I asserted from ResMed's press release that "the company gave no FY27 guidance," and from that derived the causal chain "refusing to guide is a negative signal" — in fact the call gave FY27 underlying revenue +5%~7% and earnings +12%~14%, plus a $75 million revenue / $0.15 EPS drag from the Astral recall. A press release without guidance ≠ the company gave no guidance; on earnings day, guidance often appears only on the call. From now on, any time I write "company X gave no guidance," I must first confirm I have checked the call transcript.
• Other substantive issues the risk-auditor caught: NET's "four consecutive quarters of acceleration" being falsified by the sequence itself (one quarter went 33.6→33.5); TTD's "the EBITDA collapse is 4x revenue" mixing two different denominators; TEAM's "DC drags growth by 11.6pp" treating the difference between two growth rates on different bases as a drag; TEAM's CEO buying being elevated from "intends to establish a plan" to "committed to buy" (and the 10b5-1 cooling-off period meaning no near-term buying is possible); the three SPCX lockup numbers being mutually inconsistent; and DOCS mixing after-hours and pre-market sessions in a price comparison. All corrected and flagged in the body as "the first draft said this, since corrected."
• The whole-brief pivot that "only ABNB raised guidance" was flagged as too strong (NET did raise +5%, DOCS did raise revenue, and TEAM's first annual guide does not fit a "raise" framework), and was changed to "the quality of the raises differs enormously, and only ABNB is a substantive dual raise on a full-year basis," with PUBM added to avoid an arbitrary sample.
• Compliance: the audit judged that the label system (priority deep-dive / watch closely / watch only / avoid / short watch) constitutes ratings. That label system is an output structure mandated by SKILL.md and is not altered unilaterally by this brief; but the phrasings that most resembled trading advice ("the cleanest long," "downside far exceeds upside," "not a good entry point / wait for the bounce to fade," "don't use it to place a bet") have been rewritten as falsifiable observations and verification conditions.

---

⚠️ Risk disclosure: this list is a pre-market information review and observation only, and does not constitute investment advice. U.S. equities carry high volatility and pre-market gap risk, and post-earnings IV crush and guidance reversals occur; automatically generated content may contain stale information or factual errors, so company disclosures and SEC filings govern, and this may not be used directly as a basis for trading.