2026-09-28 13:40
Post-Close Brief — 2026-08-17

type: earnings-brief session: PM date: 2026-08-17 daily_note: "[[Daily/2026-08-17]]" tags: [earnings, sellside]


EarningsBrief PM — 2026-08-17

← [[Daily/2026-08-17|Back to the daily note]]

Information cutoff: 21:12 ET / America/Toronto. Workflow status: BLOCKED — REPORT DELIVERY VERIFIED; SENTIMENT HISTORY NOT RENDERED.

PM executive decision sheet

[[FN]] — WAIT, Tier 1, PROVISIONAL — RELEASE ONLY. Fabrinet delivered $1.316B of revenue and $4.10 of non-GAAP EPS, about 0.3% and 4.3% above the dated ChartMill consensus, and guided fiscal Q1 revenue to $1.375B–$1.425B. The operating result is strong, but the stock entered at an exacting valuation and fell from a $598.58 regular-session close to $556.27 at 19:59 ET. The buried negative is cash conversion: FY2026 free cash flow fell to $4M as inventory and receivables absorbed cash and capex doubled. Wait for the full call and a better risk/reward; a price below $500 or evidence that working capital normalizes would improve the setup.

[[XP]] — HOLD, Tier 2, PROVISIONAL — RELEASE ONLY. Net inflow accelerated and EBT margin expanded, while retail take rate remained below last year and active clients were flat sequentially. At $15.53 after hours, the roughly 7.6x annualized adjusted-EPS multiple is undemanding, but a release contradiction over net inflow and the absence of a full transcript keep conviction capped. Add only after management reconciles the inflow figure and shows that flow quality, not wholesale cyclicality, can sustain 22%+ ROAE.

[[HTHT]] — HOLD, BMO catch-up remains PENDING_TRANSCRIPT. The morning release-only call is directionally confirmed by the $46.34 19:51 ET price versus $41.88 on August 14, but the full-session move is not evidence about management tone. No complete transcript or fully reviewed Q&A record was available; the AM judgment is therefore PARTIAL, not final.

Coverage Triage

Ticker Report verified TIF position / Ledger Market cap Reaction Transcript status Tier Reason
[[FN]] Yes; company IR release and SEC 8-K on Aug. 17 Ledger file located but unreadable because the TIF volume did not materialize; no verified open threshold $20.43B $598.58 close; $556.27 at 19:59 ET Official call held 17:00 ET; complete transcript not public at cutoff TIER 1 — FULL UNDERWRITE >$10B, >5% after-hours move, AI-optics and contract-manufacturing read-through
[[XP]] Yes; company IR schedule and SEC 6-Ks accepted Aug. 17 Ledger file located but unreadable; no verified open threshold $8.20B $15.70 close; $15.53 at 19:49 ET Call-highlights article exists, but no complete transcript/Q&A TIER 2 — DETAILED UPDATE Qualifying reporter with complete primary release but lower portfolio/materiality rank

Calendar exceptions

Ticker Resolution
GRFS Excluded. Grifols released H1 2026 results on July 28, not August 17.
CHRN Excluded. ChronoScale IR still shows Q1 2026 as its latest result, and SEC submissions show no filing after July 24. No Aug. 17 release was verified.

Tier 1 — Full underwrites

[[FN]] — Fabrinet

1. PM decision line

The pre-print stack was a $1.25B–$1.29B company revenue guide, $3.72–$3.87 non-GAAP EPS guide, and dated ChartMill consensus of $1.312B revenue and $3.93 EPS. Fabrinet delivered $1.316B and $4.10, then guided fiscal Q1 2027 revenue to $1.375B–$1.425B and non-GAAP EPS to $4.10–$4.25. The business delta is positive but the stock delta is less attractive: at $556.27 after hours, Fabrinet still discounts substantial AI-optics durability while FY2026 free cash flow was nearly zero. Action: WAIT; confidence medium-low until the complete call and working-capital explanation arrive.

Status: PROVISIONAL — RELEASE ONLY. Price: $556.27 at 19:59 ET, down 7.1% from the $598.58 regular close; this is an after-hours observation, not a settled next-day reaction.

2. Pre-print expectations and variance

Layer / metric Prior guide or expectation Actual / new guide Variance and classification
Q4 revenue Company $1.25B–$1.29B; ChartMill $1.312B, 12 analysts $1.3158B +$45.8M vs guide midpoint; +$3.8M / +0.3% vs consensus; STRUCTURAL POSITIVE, though only a small Street beat
Q4 non-GAAP EPS Company $3.72–$3.87; ChartMill $3.93 $4.10 +$0.305 / +8.0% vs guide midpoint; +$0.17 / +4.3% vs consensus; STRUCTURAL POSITIVE
Q1 FY27 revenue No company guide before print $1.375B–$1.425B Midpoint +6.4% q/q and +43.1% y/y; STRUCTURAL POSITIVE if supply and ramps convert
Q1 FY27 non-GAAP EPS No company guide before print $4.10–$4.25 Midpoint +1.8% q/q and +56% y/y; STRUCTURAL POSITIVE with tax-quality caveat
Gross margin Q3 GAAP gross margin about 11.9% Q4 12.0%; non-GAAP 12.2% Stable sequentially, -20 to -30 bp y/y; MIXED because scale is not yet expanding gross margin
Operating margin Q3 GAAP about 9.9% Q4 GAAP 10.2%; non-GAAP 10.9% Sequential operating leverage; STRUCTURAL POSITIVE
FY free cash flow FY2025 $207.3M FY2026 $4.2M -98%; STRUCTURAL NEGATIVE / CAPACITY RAMP, pending proof of reversal

The buy-side hurdle is not verifiable from a credible public source. The collector's yfinance revenue actual of $1.214B is stale Q3 data and is rejected in favor of the official release. The valuation-implied hurdle is high: using a $556 price and a rough fiscal-2027 EPS sensitivity of $17–$19, the stock trades near 29x–33x forward earnings. That multiple requires sustained high-20s earnings growth, stable low-double-digit operating margin, and eventual cash conversion—not merely another guide beat.

EPS quality gate. GAAP EPS was $3.83 and non-GAAP EPS $4.10, a $0.27 adjustment. Share compensation, legal, severance and restructuring added back roughly $0.27, while a $1.56 non-marketable-equity gain was offset by a $1.58 Pillar Two tax provision. The net adjustment is only about 16% of the $0.17 non-GAAP EPS beat versus consensus when measured on absolute reconciling items it is much larger and economically noisy. The reported non-GAAP beat is not classified as low quality solely under the >30% net-contribution test, but investors should not treat the near-canceling equity gain and tax provision as recurring operating power.

3. Release-only read and operating engine

The release confirms a powerful capacity-and-program ramp: quarterly revenue rose 44.6% y/y and 8.4% q/q, while operating income rose 50.8% y/y. Q1 guidance implies another sequential step-up. That is the good news the market expected from AI-driven optical systems and adjacent high-performance computing. The less comfortable truth is that Fabrinet is financing the ramp through its balance sheet. Receivables rose 34% y/y to $1.018B and inventory rose 76% to $1.021B, both faster than annual revenue growth of 36%. Accounts payable rose 58%, cushioning—but not eliminating—the cash drain.

Causal KPI Evidence and rate of change Financial transmission Compound assessment
Revenue / program ramp $1.316B, +44.6% y/y, +8.4% q/q; Q1 midpoint $1.400B More optical and HPC units absorb fixed manufacturing overhead and raise operating profit Positive, but customer/program concentration remains
Operating margin 10.2% GAAP and 10.9% non-GAAP; +40/+20 bp y/y Scale and SG&A discipline offset stable gross margin Positive operating leverage
Inventory $1.021B, +76% y/y and +$440M Component commitments and ramp inventory consume cash; excess/obsolete risk rises if demand changes Negative unless turns recover
Receivables $1.018B, +34% y/y Growth funds customer working capital and delays cash conversion Neutral-to-negative; consistent with growth but material
Capex $252.5M FY26 vs $121.1M FY25 Adds capacity for future programs but suppresses near-term FCF Positive only if utilization and returns arrive
Free cash flow $4.2M FY26 vs $207.3M FY25; Q4 -$36.9M Weak conversion limits buybacks and raises the cost of an execution miss Buried negative

The compound positive is causal: customer programs drive volume, volume absorbs fixed overhead, and higher operating profit can ultimately monetize the new capacity. The compound risk is equally causal: program ramps require inventory and equipment before cash receipts; if schedules slip, both utilization and working-capital turns deteriorate together. Buried signal: the income statement is accelerating while cash conversion is deteriorating sharply. A durable bull case needs both to converge, not a permanent excuse that growth consumes all cash.

4. Guidance and FY1/FY2 bridge

Fabrinet's Q1 midpoint implies approximately $5.6B annualized revenue. A reasonable fiscal-2027 sensitivity is $5.6B–$6.0B revenue, 10.5%–11.1% non-GAAP operating margin, and $17–$19 EPS. Fiscal 2028 can reach $6.2B–$6.8B and $19–$22 EPS if program ramps persist and gross margin remains near 12%; it falls below those ranges if supply, customer concentration, or utilization reverses.

Estimate Revised view Driver algebra
FY1 revenue $5.6B–$6.0B optical/HPC units × program share × price/mix; Q1 midpoint anchors the floor
FY1 operating margin 10.5%–11.1% non-GAAP gross margin near 12.2% less SG&A; upside from fixed-cost absorption
FY1 EPS $17–$19 revenue × operating margin + interest/other income − normalized tax, divided by ~36.3M shares
FY1 FCF $200M–$400M sensitivity operating profit + D&A − inventory/AR build − capex; the main uncertainty is working-capital reversal
FY2 revenue $6.2B–$6.8B FY1 run rate × 10%–15% growth, not the current 36% base effect
FY2 EPS $19–$22 revenue growth plus modest leverage, bounded by customer pricing and tax

Valuation sensitivity at the $556.27 after-hours price: 25x $18 FY1 EPS yields $450; 30x $20 yields $600; 22x $16 yields $352. Weighting bear/base/bull at 25%/50%/25% gives roughly $463. The market is still pricing a bull-leaning outcome. A $500 entry corresponds to about 27.8x $18 and offers a more defensible risk/reward; below $450 the base case becomes attractive if cash conversion is intact.

5. Call forensics, debate, thesis and action

Transcript provenance: the official 17:00 ET webcast was identified, but no complete speaker-attributed transcript with full Q&A was public by 21:10 ET. A webcast replay alone was not fully reviewed. Status is PENDING_TRANSCRIPT; tone, Q&A answer quality, prior-call language deltas, omissions, pressure delta, and management credibility are PENDING — CALL and unscored.

The next AM catch-up must answer:

  1. How much Q1 growth comes from telecom, datacom, DCI and high-performance computing, and which customer/program ramps are supply constrained?
  2. When do inventory turns and FY free cash flow normalize, and how much inventory is non-cancellable or customer protected?
  3. Does the $1.375B–$1.425B guide assume gross margin expansion, and how should Pillar Two tax affect FY2027 GAAP versus non-GAAP EPS?
Live claim entering print Evidence required Evidence received Verdict Next resolution
AI optics and adjacent compute ramps can sustain >25% growth Revenue and next-quarter guide above Street without margin erosion Q4 +45% y/y; Q1 midpoint +43% y/y; gross margin stable Strengthened Segment bridge and customer concentration on call
Scale will convert into higher earnings quality Operating leverage plus FCF conversion Operating margin improved; FY FCF collapsed Unresolved / weakened on cash Inventory turns and FY27 FCF framework
Premium valuation is supported by durable program wins Multi-quarter visibility and $20+ FY28 EPS Release gives only one-quarter guide Unresolved Backlog/program-duration evidence
Balance sheet de-risks expansion Net cash and capacity funded internally $875M cash/investments, but working capital and capex absorbed cash Partly strengthened Cash-conversion timing
Thesis pillar Delta Status
Demand / volume Q4 and Q1 guide show accelerating demand REINFORCED
Pricing / mix Gross margin did not expand with scale UNRESOLVED
Margin / cost architecture Operating margin improved IMPROVED
Competitive position Program ramps and high-complexity manufacturing remain strong REINFORCED
Balance sheet / capital allocation Net cash remains, but FCF and buybacks weakened WEAKENED
Management credibility Guide beaten; cash-conversion explanation absent from release UNRESOLVED
Catalyst timing Full transcript next AM; Q1 FY27 results likely November UNCHANGED

Old narrative: Fabrinet is a scarce, high-quality outsourced manufacturing beneficiary of AI optical infrastructure. After release: the demand and operating-leverage thesis strengthened, but the cash-conversion and valuation debate became more important. After call: PENDING — CALL. Settled reaction: unavailable until the next regular session. Business delta: positive. Estimate delta: positive for FY1/FY2 EPS, uncertain for FCF. Stock delta: negative risk/reward at the regular close and only balanced after the after-hours decline.

Decision card: WAIT; no sizing increase. Confirmation: Q1 revenue at least $1.40B, non-GAAP operating margin at least 10.8%, and FY2027 FCF framework above $250M. Falsification: Q1 revenue below $1.375B, gross margin below 11.8%, or inventory remains above 75% of quarterly revenue without customer protection. Catalyst: complete transcript by 08:00 ET August 18 and Q1 FY27 results. Entry trigger: below $500 with confirmed working-capital normalization; reduce/avoid trigger: above $600 without a $20+ FY2027 EPS path. 10-second PM line: exceptional growth and guide, but almost no FY cash conversion leaves too much of the bull case embedded above $550.

Tier 2 — Detailed updates

[[XP]] — XP Inc.

Decision, expectations and variance

XP's official release showed R$5.056B gross revenue, R$4.884B net revenue, R$1.565B EBT, R$1.384B adjusted net income and R$2.67 adjusted diluted EPS. The dated Benzinga public estimate was about $0.50 EPS and $987.9M revenue; at the Aug. 17 USD/BRL rate near 5.22, EPS equates to roughly $0.51 and net/gross revenue to about $936M/$968M. EPS was approximately in line to slightly above, while revenue was below that vendor's USD estimate. Currency and gross-versus-net definitions make the revenue comparison lower confidence.

KPI Actual YoY / QoQ Interpretation
Client assets R$1.535T +12% / flat Asset growth is positive, but the sequential base stalled
Net inflow Table: R$28B; narrative: R$20B Table +188% / +94%; narrative +28% / +7% Primary-release contradiction; direction positive, magnitude unresolved
Retail take rate 1.20% -5 bp / +2 bp Monetization stabilized sequentially but remains below last year
Retail revenue R$3.881B +8% / +3% Equities, funds and new verticals offset fixed-income mark-to-market pressure
Wholesale revenue R$1.175B +32% / +3% Strong, but cyclicality and segment reclassification matter
EBT margin 32.0% +172 bp / +209 bp Clear operating leverage
Adjusted ROAE 22.5% -189 bp / +76 bp Sequential improvement, still below last year's return profile
Capital ratio 20.3% +15 bp / -39 bp Ample capital above management's 16%–19% target range

The buy-side hurdle and TIF Ledger threshold are not verifiable. Market positioning is not reliably sourced. The valuation-implied bar is modest: annualizing R$2.67 quarterly EPS and translating at 5.22 yields about $2.05 per ADS, so the $15.53 after-hours price is approximately 7.6x earnings. That price does not require heroic growth, but it does require that Brazil-rate normalization, flow quality, and capital return offset take-rate pressure.

EPS quality gate: the release presents adjusted net income/EPS, but no evidence available at cutoff shows a one-time item contributing more than 30% of the variance versus the $0.50 estimate. The main quality issue is not below-the-line engineering; it is whether wholesale growth and fixed-income marks are repeatable. Share repurchases of roughly R$1B in H1 aided per-share progression, so EPS growth should be separated from underlying net-income growth.

Operating engine, model bridge and debate

XP is a spread-and-fee platform. Client assets and inflows expand the revenue base; take rate converts assets/activity into retail revenue; wholesale capital-markets activity adds cyclical upside; operating discipline converts gross profit into EBT; and excess capital funds buybacks. This quarter, flow and margin improved faster than the customer base. The buried signal is mixed: net inflow accelerated despite flat active clients, suggesting deeper wallet capture or institutional/large-account skew, while the release itself reports two different inflow figures.

The positives compound if inflow lifts assets, stable take rate monetizes those assets, and EBT margin remains above 31%. The negative compound loop is take-rate compression plus headcount growth: headcount rose 13% y/y while active clients rose only 1%; if asset growth is market-driven rather than net-new money, operating leverage can reverse.

Estimate Revised sensitivity Algebra
FY1 net revenue R$19.5B–R$20.2B client assets × retail take rate + wholesale + banking/cards/insurance
FY1 EBT margin 30.5%–32.0% gross profit − personnel/admin/technology costs
FY1 adjusted EPS R$10.4–R$11.2 EBT − tax, divided by post-buyback diluted shares
FY2 net revenue R$21.0B–R$23.0B 8%–14% growth from flows, activity and cross-sell
FY2 adjusted EPS R$11.5–R$13.0 revenue growth + margin stability + buybacks

At 5.22 BRL/USD, FY1 EPS sensitivity is about $1.99–$2.15. A 6x bear multiple on $1.80 gives $10.80; an 8.5x base multiple on $2.10 gives $17.85; a 10x bull multiple on $2.35 gives $23.50. A 25%/50%/25% weighting yields about $17.50, roughly 13% above the after-hours price.

Debate claim Evidence Verdict
Flows are reaccelerating Table and narrative both improve, but disagree on magnitude Strengthened directionally; magnitude unresolved
Take-rate pressure is bottoming +2 bp q/q, still -5 bp y/y Partially strengthened
Operating leverage can offset slower client growth EBT +15% and margin +172 bp y/y Strengthened
Capital return supports per-share value R$1B H1 buybacks; another R$1B authorization open Strengthened, subject to price discipline

Call, thesis and action

Transcript provenance: the official call was held at 17:00 ET. A call-highlights article is discovery evidence only and cannot establish complete Q&A. No full transcript with speaker anchors was available by cutoff. Status is PENDING_TRANSCRIPT; sentiment is unscored.

The call must answer: (1) Is Q2 net inflow R$28B or R$20B, and what channel/client cohort drove it? (2) How much wholesale growth is reclassification versus organic activity, and what is a normalized run rate? (3) Why did headcount rise 13% against 1% active-client growth, and when does that investment earn through?

Thesis pillar Status Evidence
Demand / volume IMPROVED inflow acceleration and assets +12% y/y
Pricing / mix UNCHANGED take rate +2 bp q/q but -5 bp y/y
Margin / cost IMPROVED EBT margin 32.0%
Competition UNRESOLVED flat active clients and no market-share proof
Capital allocation IMPROVED buybacks and capital above target range
Management credibility UNRESOLVED contradictory inflow disclosure requires reconciliation
Catalyst timing UNCHANGED full transcript next AM; Q3 results likely November

Old narrative: XP is a low-multiple Brazilian financial platform whose earnings depend on rate/activity normalization and better monetization. After release: flow and margin recovery are more credible, but flat client count and take-rate pressure prevent a clean growth reacceleration call. After call: PENDING — CALL. Business delta: modestly positive. Estimate delta: low- to mid-single-digit positive if EBT margin holds. Stock delta: valuation is supportive, but the incomplete disclosure and macro sensitivity justify patience.

Decision card: HOLD; consider adding below $15 after call verification. Confirmation: quarterly net inflow at least R$20B, retail take rate at least 1.20%, and EBT margin at least 31%. Falsification: net inflow below R$10B, take rate below 1.15%, or ROAE below 20%. Catalyst: complete transcript by 08:00 ET Aug. 18 and Q3 print. Valuation trigger: add below 7x a verified $2.10 FY1 EPS path; reduce above $20 without >R$12 FY2 EPS visibility. 10-second PM line: cheap and improving, but the release's inflow contradiction and absent transcript keep the call at HOLD.

BMO release-to-call and full-session reconciliation

[[HTHT]] — morning judgment PARTIAL

The AM report expected the call to quantify the raised HWC/M&F guide, separate HWI geopolitical pressure from new-hotel mix, and explain the $2.5B capital-return split. By 21:10 ET no complete public transcript or fully reviewed Q&A record had been obtained, so management's answers cannot be asserted. The $46.34 late price versus $41.88 on Aug. 14 is consistent with the market capitalizing the guide and shareholder-return plan, but causal attribution remains a TIF inference. The release-only HOLD is neither disproved nor fully confirmed: estimate and balance-sheet evidence strengthened, while same-hotel HWC RevPAR (-3.0%) and HWI RevPAR (-3.8%) remain the decisive unresolved operating questions.

Morning call grade: PARTIAL because the business/estimate judgment held and the stock reacted positively, but the call-evidence and sentiment gates remain open. Required next source: complete speaker-attributed Q2 transcript or fully reviewed official replay with Q&A by 08:00 ET Aug. 18.

Cross-company causal read-throughs

  1. Growth without cash or client breadth needs a second proof point. FN's income statement outran cash conversion; XP's flows and margin outran client growth. Both prints improve estimates, but each has a quality variable the headline does not resolve.
  2. AI infrastructure demand is still strong at the manufacturing layer. FN's Q1 guide implies >40% y/y revenue growth, supporting optical component and contract-manufacturing demand. The read-through is conditional on segment mix and inventory protection.
  3. Capital allocation separates the low- and high-multiple stories. XP can retire shares at a single-digit earnings multiple; FN must fund capacity while trading near 30x forward earnings. The same dollar of cash has very different per-share value.

Next-morning transcript queue

Ticker Deadline Required evidence
[[FN]] 2026-08-18 08:00 ET Complete current transcript/Q&A plus prior Q3 call; segment ramp, inventory protection, FCF and tax bridge
[[XP]] 2026-08-18 08:00 ET Complete transcript/Q&A plus prior Q1 call; inflow reconciliation, wholesale normalization, headcount productivity
[[HTHT]] 2026-08-18 08:00 ET Complete current transcript/Q&A plus Q1 comparison; same-hotel assumptions, HWI bridge, capital-return mechanics

Completion Audit

Ticker Tier Status Analytical words Causal KPIs Q&A Debate claims Prior-call deltas Omissions FY1/FY2 bridge Transcript provenance Sentiment Tone delta Answer quality Pressure delta Tracker read-back Failed / deferred gates
FN 1 PROVISIONAL — RELEASE ONLY >1,700 6 0 4 0 0 Complete via sensitivity Official call identified; complete transcript absent PENDING_TRANSCRIPT PENDING — CALL PENDING — CALL PENDING — CALL BLOCKED fn.call-forensics, sentiment-delivery
XP 2 PROVISIONAL — RELEASE ONLY >1,000 7 0 4 0 0 Complete via sensitivity Official call identified; highlights rejected as transcript PENDING_TRANSCRIPT PENDING — CALL PENDING — CALL PENDING — CALL BLOCKED xp.call-forensics, sentiment-delivery
HTHT BMO catch-up PENDING_TRANSCRIPT Morning full underwrite + PM reconciliation 7 in AM 0 4 in AM 0 0 Complete in AM sensitivity Official replay exists; complete Q&A not reviewed PENDING_TRANSCRIPT PENDING — CALL PENDING — CALL PENDING — CALL BLOCKED current transcript and tracker delivery

Sources