WKAP Radar Feed - 2026 - 10 - 02

WKAP Radar Feed

*October 2, 2026*

*China EVs: Close the Delivery Test, Keep the Earnings Gate Open*

*3 Thesis Objects: NIO, XPEV, LI*

Preheader:

Make your AI track X alpha.

WKAP Feed compresses earnings inflections, KOL flow, supply-chain
bottlenecks, and short-duration attention into daily AI-ready context.

Set up your AI to read this Feed.

Ask:

“What did I miss?”

## HUMAN USER — START HERE

Li Auto's September deliveries fell 15.6% sequentially, yet Q3
finished just 36 vehicles below its original guidance ceiling. All
three companies met their delivery ranges; none supplied Q3 earnings
with those counts. Retain NIO > XPEV > LI for economic diligence, not
expected returns. The hidden distinction is period and denominator:
NIO's September increment came mostly from FIREFLY, XPeng's latest
services category generated nearly half its gross profit, and Li's
positive operating cash did not cover investment. Close September's
quantity tests and reopen the model at mix, vehicle contribution and
cash conversion. The weakest shared assumption is that fulfilled
volume supports unchanged unit economics. Next Q3 financial
disclosures can change the ranking; October brand counts offer an
earlier, incomplete mix observation.

Suggested Prompt:
“Reconcile the original-guidance closure ledger, then populate the
three earnings worksheets. Separate facts, expectations,
interpretation and unknowns. Challenge NIO > XPEV > LI using the exact
promotion and invalidation rules; distinguish automotive from
consolidated conclusions. Return source-linked missing inputs and
executable 7-day/30-day workflows, without buy/sell advice.”

Or simply ask: “What did I miss?”

The goal is to turn today's information into trackable thesis objects
and explicit status-changing tests.

## FOR YOUR AI AGENT

This is a structured market-cognition input file. Follow delivered
units → vehicle revenue and mix → vehicle gross profit → services
contribution → operating expenses → operating cash and investment.
Delivery counts resolve only the first link. These are competing
research objects, not a demonstrated supplier/customer chain.

Process all three as trackable_wow updates to September 1 objects.
Keep actuals separate from original management guidance and analytical
thresholds. Preserve period, unit, definition, source and calculation
inputs. Close completed quantity requests; do not keep asking whether
an already reported quarter can meet its range. Rank diligence
separately from valuation, liquidity safety and suitability. Attention
may improve the question but cannot populate an undisclosed financial
cell.

Do not treat this Feed as investment advice.

## AGENT_OUTPUT_SPEC

1. Today in One Sentence: which quantity uncertainty closed and which
earnings uncertainty remains?
2. Thesis Object Summary: previous → current state and the changed
research assignment.
3. Fundamental Evidence vs Attention Flow: issuer counts/tables versus
independent interpretations.
4. Evidence-State Audit: REPORTED_FACT / MANAGEMENT_EXPECTATION /
INTERPRETATION / UNRESOLVED_QUESTION.
5. Period-and-Denominator Comparison: monthly movement, quarterly
guidance and financial conversion.
6. Cross-Object Ranking: test NIO > XPEV > LI and identify
adjacent-pair reversal conditions.
7. Business-Proof Comparison: vehicle margin, services attribution,
operating leverage and funded investment.
8. Attention Stage / Crowding Risk: observable discussion, without
invented positioning.
9. Weakest Assumption by Object: the one missing link capable of
defeating the thesis.
10. Most Important Data Point: exact metric, period, source and model cell.
11. Attention-to-Rerating Conditions: business proof followed by a
separate valuation assessment.
12. Attention Fade Conditions: narrative exhaustion versus actual
business deterioration.
13. 7-Day Research Workflow: reconcile inputs, retrieve updates,
preserve NO_NEW_DATA.
14. 30-Day Research Workflow and WKAP Daily Top 3 Source Follow-Up:
Validate/Upgrade/Downgrade/Invalidate.
15. Portfolio Fit: only with supplied holdings and constraints;
otherwise withhold sizing and suitability.

## TODAY_SUMMARY

### Part 1 — Main Market Thesis

NEW FACT — REPORTED_FACT: October 1 delivery releases complete the
quarter. UNCHANGED FACT: June-quarter financials remain the latest
actual margin/cash baselines. MANAGEMENT_EXPECTATION: original Q3
guidance supplies the historical comparison, not today's profit
result. INTERPRETATION: move from estimating volume attainment to
testing earnings conversion. UNRESOLVED_QUESTION: did price
competition, model mix and spending offset the delivered units?
Consumer financing conditions and pricing matter through realized
revenue and collections.

### Part 2 — Today’s Thesis Objects

NIO — Fundamental: Q3 109,178 closes the 108,000–111,000 range test;
September brand changes complicate a premium-mix inference. Attention:
@NioToday asks whether ONVO launch interest persists. Primary:
https://ir.nio.com/news-releases/news-release-details/nio-inc-provides-september-and-third-quarter-2026-delivery

XPEV — Fundamental: Q3 118,390 closes the 115,000–121,000 range test;
Q2 services economics cannot be assigned to car deliveries. Attention:
@Matteo_Inv requests separately disclosed non-car economics. Primary:
https://ir.xiaopeng.com/news-releases/news-release-details/xpeng-announces-vehicle-delivery-results-september-and-third-5

LI — Fundamental: monthly reports sum to Q3 99,964, inside
95,000–100,000 despite September's decline. Attention: @Jas0nYu uses
the original range rather than only the latest month's direction.
Primary: https://ir.lixiang.com/news-releases/news-release-details/li-auto-inc-september-2026-delivery-update

### Part 3 — Attention Flow Today

The selected discussion moves from delivery headlines toward brand
persistence, attribution and guidance-relative execution. Jason's
comparison is the clearest period-correction signal:
https://x.com/Jas0nYu/status/2105532543240221122 . Karl explicitly
discloses being long NIO since 2020. No holding disclosure was visible
for Matteo or Jason; that does not establish no holding. These three
posts are curated evidence of discussion, not measured market-wide
fund flows or consensus.

### Part 4 — The Better Question

Surface question: which company delivered the strongest month? Better
question: which completed quantity test permits which model update,
and what still prevents an earnings promotion? Replace estimated Q3
units with actuals now. Do not simultaneously promote price/mix,
vehicle margin, technology value and cash assumptions without their
own disclosures.

## MARKET_REGIME

RISK_TONE — INTERPRETATION: Selective post-delivery diligence, not an
index-level risk call.
MAIN_DRIVER — NEW REPORTED_FACT: Three management volume ranges closed
successfully.
MARKET_CONTEXT — UNCHANGED FACT: Q2 economics have not become Q3
economics merely because the quarter ended.
ATTENTION_ENVIRONMENT — INTERPRETATION: Monthly declines, quarterly
growth and AI language invite different shortcuts; positioning is
unmeasured.
WKAP_VIEW — UNRESOLVED_QUESTION: Which updated earnings case survives
mix, expense, working-capital and valuation scrutiny?

## STATE_CHANGE_LEDGER

NIO — Previous: September needed 36,230–39,230 deliveries to meet
original Q3 guidance. New fact: 37,408. Current: quantity gate closed;
mix/operating-leverage gate open, rank 1. Decision impact: retire the
volume forecast; investigate the composition of added units. Next
changer: Q3 vehicle margin with GAAP operating loss, not another
total-delivery headline.

XPEV — Previous: September needed 37,866–43,866. New fact: 41,256.
Current: quantity gate closed; two-engine earnings attribution open,
rank 2. Decision impact: build separate vehicle and services
gross-profit bridges. Next changer: vehicle margin/contribution plus
evidence about service milestones; group margin alone cannot settle
both.

LI — Previous: September needed 26,853–31,853. New fact: 31,817.
Current: quantity gate closed near the upper end; funded-repair gate
open, rank 3. Decision impact: reject a quarterly-miss inference from
September's sequential decline. Next changer: vehicle-margin
improvement and operating cash covering actual investment, with timing
effects explained.

The original economic-diligence order is unchanged; the questions
advance. These are research-state changes, not new purchases or
automatic public recommendations.

## CROSS_OBJECT_EVIDENCE_MATRIX

Audit axis | NIO | XPEV | LI
REPORTED_FACT: July / August / September units | 35,934 / 35,836 /
37,408 | 38,027 / 39,107 / 41,256 | 30,468 / 37,679 / 31,817
INTERPRETATION: Q3 sum | 109,178 | 118,390 | 99,964
MANAGEMENT_EXPECTATION: original Q3 range | 108,000–111,000 |
115,000–121,000 | 95,000–100,000
INTERPRETATION: versus midpoint | −322 | +390 | +2,464
REPORTED_FACT: Q2 vehicle margin | 18.5% | 12.1% | 9.4%
REPORTED_FACT: Q2 group margin | 18.4% | 20.7% | 11.0%
UNRESOLVED_QUESTION | Mix and operating leverage | Vehicle/services
profit attribution | Cash after actual investment
Separate attention field | Holder's brand-persistence critique |
Separate economics requested | Quarter-versus-month comparison

Guidance worksheet: September requirement = original Q3 range minus
July and August. Compare actual with both endpoints; never replace the
original range with a hindsight target. LI's 30,468 + 37,679 + 31,817
= 99,964; upper-end shortfall 100,000 − 99,964 = 36, or 0.036%.
September change = 31,817 / 37,679 − 1 = −15.5577%. Both are true. No
interim revision was found in the reviewed issuer release sequences;
this is an original-guidance audit, not an analyst-consensus
comparison.

NIO mix worksheet — REPORTED_FACT inputs: August NIO/ONVO/FIREFLY
21,174/8,810/5,852; September 21,318/8,763/7,327. INTERPRETATION:
changes +144/−47/+1,475 sum to +1,572; FIREFLY accounts for 93.8% of
the net increment, not sales. Full Q3 brand totals are
62,500/27,728/18,950. NIO-brand share rose to 57.25% from Q2's 56.61%.
Therefore the monthly observation neither proves premium uplift nor a
full-quarter premium decline. Brand margins and intra-brand price mix
remain undisclosed.

XPEV attribution worksheet — Q2 reported RMB millions: vehicle revenue
17,046.476 minus cost 14,987.590 = gross profit 2,058.886;
services/other revenue 2,697.117 minus cost 672.521 = 2,024.596. The
latter is 49.5801% of total gross profit 4,083.482 and 13.6607% of
revenue. This category includes parts/accessories and technical R&D
services, not simply robotics. Reconcile the same lines in Q3 before
inferring repeatability.

LI funding worksheet — Q2 reported RMB millions: CFO 15.025 minus
capex 1,315.790 = FCF −1,300.765. At that fixed historical capex, CFO
would have needed another 1,300.765 to fund investment; that is
arithmetic, not a cash forecast. For Q3 use actual capex, not this
frozen denominator, and reconcile collections/payables before calling
the coverage durable.

## RANKED_UNANSWERED_QUESTIONS

1. Does NIO retain at least 18.5% vehicle margin while narrowing GAAP
operating loss below Q2's RMB347.2 million? Retrieve the next
financial release together, not separate favorable lines. This can
remove the leader's promotion.
2. Does XPEV raise vehicle margin above 12.1% and vehicle gross profit
while services remain economically credible? Retrieve revenue/cost
lines and milestone descriptions. Return automotive and consolidated
verdicts separately.
3. Does LI exceed 9.4% vehicle margin and fund actual capex from CFO
without unexplained working-capital assistance? Retrieve the cash-flow
statement and collection/inventory-payment explanation. This can
promote repair ahead of unresolved attribution.
4. What liabilities, dilution and normalized earnings support
valuation at the observed prices? Only after business-proof updates,
compare valuation cases. A lower share price or larger broad cash
balance alone cannot answer portfolio fit.

## SCENARIO_BRANCHES

Confirm: NIO meets both its margin and operating-loss rules; XPEV and
LI remain economically unresolved. Preserve NIO > XPEV > LI, with no
allocation conclusion. Successful delivery attainment is already in
the baseline and cannot be counted twice.

Mixed: NIO margin holds but loss widens, or vice versa: keep under
review. XPEV group profitability improves through services while
vehicle margin fails its test: consolidated progress may be real, but
do not promote the automotive case. LI margin improves while CFO still
fails to cover actual capex: repair advances without becoming funded
repair.

Break/reorder: NIO's margin declines and operating loss widens:
reverse its promotion. XPEV can outrank it only with reconciled
durable consolidated improvement, not by inheriting first place as
though its own gate passed. LI moves ahead of still-unresolved XPEV
when margin exceeds 9.4% and operating cash covers actual investment
with the bridge explained. Deteriorating LI margin plus renewed cash
pressure defeats early repair.

No new data: retain absolute status and last actual financial period.
Return the missing input and future retrieval checkpoint. Repeated
social claims and elapsed review dates are not failed financial tests.

## ATTENTION_TRADE_BOARD

Attention Trade Board

Object | Attention Stage | Attention Source | Why Today | Hard
Evidence | Narrative Gap | Crowding Risk | Likely Window | Fade Signal
NIO | Mix scrutiny | @NioToday | Quarter complete | Brand counts; Q2
financials | Group units versus brand contribution | Unmeasured;
holder bias | Days to next disclosure | Launch rhetoric without
persistence
XPEV | Attribution demand | @Matteo_Inv | Volume known | Two Q2
gross-profit pools | Services versus robotics value | Unmeasured;
technology halo | Days to financial disclosure | Vocabulary without
separate economics
LI | Guidance reassessment | @Jas0nYu | September dip meets quarterly
ceiling | Original range reconciled | Execution versus funded repair |
Unmeasured; rebound extrapolation | Days to results | Repeating
upper-end result without cash bridge

WKAP Attention View

Strongest new closure: LI's near-ceiling result corrects the
monthly-miss shortcut. Cleanest evidence-to-attention asymmetry: LI's
period comparison. Largest optionality/evidence gap: XPEV's
unsegmented technology discussion. Most crowded: unknown without
positioning data. Highest qualitative fade risk: XPEV if disclosure
demand is replaced by promotional labels. Best durable-thesis
candidate: NIO conditional on margin/expense conversion and valuation,
not because the highest vehicle margin defines total company quality.

## RADAR_OBJECT_INDEX

THESIS_OBJECT_1: NIO
THEME: China EV / brand mix and operating leverage
STATUS: active_trackable — rank 1; quantity closed, earnings open
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $3.4205 USD (Yahoo Finance, Nasdaq Real Time Price,
U.S. premarket, October 2, 08:58:56 EDT / 19:58:56 ICT;
https://finance.yahoo.com/quote/NIO/)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-01
SETUP_TYPE: Delivery closure → mix/expense conversion
ATTENTION_STAGE: Brand-persistence interpretation
ATTENTION_WINDOW: 1–7 days for discussion; financial disclosure for conversion
KEY_QUESTION: Does fulfilled volume preserve vehicle margin and narrow
operating loss?

THESIS_OBJECT_2: XPEV
THEME: China EV / two-engine profit attribution
STATUS: active_trackable — rank 2; quantity closed, attribution open
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $9.32 USD (Yahoo Finance, Nasdaq Real Time Price,
U.S. premarket, October 2, 08:55:17 EDT / 19:55:17 ICT;
https://finance.yahoo.com/quote/XPEV/)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-01
SETUP_TYPE: Delivered units → vehicle/services bridge
ATTENTION_STAGE: Separate-disclosure demand
ATTENTION_WINDOW: 1–7 days for attention; later financial/contract evidence
KEY_QUESTION: Which profit engine improves, and is its contribution durable?

THESIS_OBJECT_3: LI
THEME: China EV / recovery and funded investment
STATUS: active_trackable — rank 3; quantity closed, funded repair open
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $10.86 USD (Yahoo Finance, delayed quote, U.S.
premarket, October 2, 08:45:00 EDT / 19:45:00 ICT;
https://finance.yahoo.com/quote/LI/)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-01
SETUP_TYPE: Guidance attainment → margin/cash repair
ATTENTION_STAGE: Period-corrected execution comparison
ATTENTION_WINDOW: 1–7 days for reassessment; later financial confirmation
KEY_QUESTION: Can improving vehicle margin fund actual investment?

Quotes are timestamped reference observations, not entry
recommendations. All first-added dates are preserved from the
canonical September 1 cards:
https://wkap.ai/radar/wkap-radar-feed-2026-09-01.html . A follow-up
does not create a new inception date.

## THESIS OBJECTS

### THESIS_OBJECT_1 — NIO

CARD_ID: WKAP-RADAR-2026-10-02-NIO
CARD_TITLE: Quantity Passed; Mix Must Earn the Operating Upgrade
TYPE: trackable_wow
THEME: Brand persistence / operating leverage
STATUS: active_trackable — first economic-diligence priority
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $3.4205 USD (Yahoo Finance, Nasdaq Real Time Price,
U.S. premarket, October 2, 08:58:56 EDT / 19:58:56 ICT;
https://finance.yahoo.com/quote/NIO/)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-01
ATTENTION_STAGE: Post-delivery mix scrutiny
ATTENTION_WINDOW: Days for discussion; next financial disclosure for
the earnings verdict

WHY_A_READER_CARES_NOW: September closes the predicted quantity hurdle
while changing the composition question.
PREVIOUS_STATE: Delivery attainment and earnings conversion unresolved.
CURRENT_STATE: Original range met; earnings confirmation still absent.
DECISION_OR_MONITORING_PAYOFF: Replace a volume forecast with the
brand/margin/expense worksheet.

THESIS_SUMMARY

The vehicle-margin lead earns research priority, not proof of overall
safety. Q2 adjusted net profit RMB26.1 million coexisted with GAAP net
loss RMB528.0 million; share-based compensation explains the
adjustment. A profitable vehicle gross margin is not corporate
break-even.

WKAP_ANGLE

The surface-level frame: group delivery growth proves premium scaling.
The alternative frame: period-specific brand mix and corporate costs
determine conversion.
The key research question: does mix preserve contribution after
selling and development expenses?

CORE_THESIS

REPORTED_FACT: the matrix separates September's FIREFLY-heavy
increment from Q3's slightly higher NIO-brand share. UNCHANGED FACT:
Q2 vehicle margin 18.5%, GAAP operating loss RMB347.2 million.
INTERPRETATION: the combination is the next test, not either number
alone. The RMB56.7 billion broad cash measure includes restricted
cash/investments; June current liabilities RMB87.619 billion were
close to current assets RMB89.250 billion. These constrain a
simplistic cash-minus-market-cap argument without establishing
imminent distress.

ATTENTION_TRADE_FRAME

Attention Source: Karl / @NioToday;
https://x.com/NioToday/status/2105991833775984977
Why Today: A disclosed holder questions ONVO's sustained appeal after
launch attention.
Attention Stage: Quantity recognition → brand-persistence debate.
Attention vs Evidence — Hard evidence: monthly brand counts and Q2
financial tables.
Primary sources:
https://ir.nio.com/news-releases/news-release-details/nio-inc-provides-august-2026-delivery-update
; https://ir.nio.com/news-releases/news-release-details/nio-inc-reports-unaudited-second-quarter-2026-financial-results
Attention / interpretation: Long NIO since 2020; the causal brand
critique is opinion, not measured consumer preference.
Attention Path: Brand persistence → mix → vehicle contribution →
operating leverage.
Attention Asymmetry: A holder can raise a useful disconfirming
question without proving the answer.
Crowding Risk: Unmeasured; selected-post reach is not market positioning.
What Could Sustain Attention: Brand persistence accompanied by
preserved margins and lower operating loss.
What Could Make Attention Fade: Repeated launch milestones without
economic follow-through.
Attention-to-Thesis Conversion: Vehicle margin at least 18.5% and GAAP
operating loss narrower than RMB347.2 million; analytical rule, not
guidance.

WEAKEST_ASSUMPTION: Additional units retain contribution after mix and
brand spending.
MOST_IMPORTANT_DATA_POINT: Q3 vehicle margin paired with GAAP
operating loss on the same period.
NEXT_DATA_POINT: October brand counts and next issuer financial
release, dates not established here. Joint margin decline/loss
widening reverses promotion; one-sided change stays under review.

### THESIS_OBJECT_2 — XPEV

CARD_ID: WKAP-RADAR-2026-10-02-XPEV
CARD_TITLE: Attribute Both Profit Engines Before Promoting Either
TYPE: trackable_wow
THEME: Vehicle contribution / services persistence
STATUS: active_trackable — second priority; automotive and
consolidated verdicts separate
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $9.32 USD (Yahoo Finance, Nasdaq Real Time Price,
U.S. premarket, October 2, 08:55:17 EDT / 19:55:17 ICT;
https://finance.yahoo.com/quote/XPEV/)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-01
ATTENTION_STAGE: Post-delivery economic attribution
ATTENTION_WINDOW: Days for questions; later financial/contract
disclosure for answers

WHY_A_READER_CARES_NOW: Completed volume permits a precise
contribution bridge, not a technology-value shortcut.
PREVIOUS_STATE: September quantity hurdle and earnings attribution open.
CURRENT_STATE: Quantity closed; vehicle/services persistence unresolved.
DECISION_OR_MONITORING_PAYOFF: Maintain two profit columns and a third
consolidated verdict.

THESIS_SUMMARY

Services generated real value in Q2. Recognizing that value does not
justify treating 20.7% group margin as the margin on each vehicle, nor
assigning all services profits to robotics. Attribution protects both
the automotive and non-automotive cases.

WKAP_ANGLE

The surface-level frame: higher deliveries plus AI optionality settle
the earnings case.
The alternative frame: vehicle contribution, service milestones and
disclosed new-business economics have different evidence.
The key research question: which incremental gross profit persists
after costs and recognition timing?

CORE_THESIS

REPORTED_FACT: Q3 units rose 14.6% sequentially, calculated against
103,295; the company rounded to 15%. Q2 vehicle margin was 12.1%,
while services/other margin was 75.1%. MANAGEMENT_EXPLANATION:
technical-R&D milestones helped services. INTERPRETATION: neither
delete those earnings nor annualize their timing. UNRESOLVED_QUESTION:
the quarter's increase in units may change auto contribution, but Q3
vehicle revenue/cost and service persistence are still needed.

ATTENTION_TRADE_FRAME

Attention Source: Matteo / @Matteo_Inv;
https://x.com/Matteo_Inv/status/2105650021614186825
Why Today: The author asks what separately disclosed economics would
justify non-car value after delivery news.
Attention Stage: Volume recognition → attribution demand.
Attention vs Evidence — Hard evidence: delivery release and Q2
vehicle/services tables.
Primary sources:
https://ir.xiaopeng.com/news-releases/news-release-details/xpeng-reports-second-quarter-2026-unaudited-financial-results
Attention / interpretation: Licensing, robotics and robotaxi
disclosure requests; no holding disclosure visible. No asserted
technology valuation is adopted.
Attention Path: Separate disclosure → attributable contribution →
durability → valuation.
Attention Asymmetry: A smaller revenue category can produce large
gross profit without proving a recurring AI business.
Crowding Risk: Unmeasured; generic technology premiums can outrun evidence.
What Could Sustain Attention: Higher vehicle contribution plus
transparent, durable services economics.
What Could Make Attention Fade: Repeated AI labels without new attribution.
Attention-to-Thesis Conversion: Automotive promotion requires vehicle
margin above 12.1% and higher vehicle gross profit. Assess services
and consolidated improvement independently.

WEAKEST_ASSUMPTION: Incremental deliveries and service milestones
produce durable profit after associated expenses.
MOST_IMPORTANT_DATA_POINT: Q3 vehicle/services revenue-minus-cost
bridge with milestone commentary.
NEXT_DATA_POINT: Next issuer financial and material contract
disclosure. Vehicle deterioration defeats an automotive upgrade even
if services help group results; durable consolidated improvement can
still change relative rank.

### THESIS_OBJECT_3 — LI

CARD_ID: WKAP-RADAR-2026-10-02-LI
CARD_TITLE: Guidance Closure Is Not Yet Funded Repair
TYPE: trackable_wow
THEME: Product recovery / cash after investment
STATUS: active_trackable — third priority; cash-confirmation required
POSITION_CONTEXT: [not provided]
PRICE_AT_PUBLISH: $10.86 USD (Yahoo Finance, delayed quote, U.S.
premarket, October 2, 08:45:00 EDT / 19:45:00 ICT;
https://finance.yahoo.com/quote/LI/)
DATE_FIRST_ADDED_TO_RADAR: 2026-09-01
ATTENTION_STAGE: Period-corrected execution reassessment
ATTENTION_WINDOW: Days for delivery debate; later financial evidence for repair

WHY_A_READER_CARES_NOW: The near-ceiling quarter prevents a false miss
classification while preserving the cash hurdle.
PREVIOUS_STATE: September attainment uncertain; early margin recovery
unproven as self-funding.
CURRENT_STATE: Quantity closed; economic and cash repair unresolved.
DECISION_OR_MONITORING_PAYOFF: Stop litigating the completed range;
demand the margin/CFO/capex bridge.

THESIS_SUMMARY

Two favorable comparisons need limits. Q3 attained guidance, while Q2
margin improved from a low base. Neither establishes sustained demand
momentum or funded investment. Larger reported liquidity is a buffer,
not a substitute for operating recovery.

WKAP_ANGLE

The surface-level frame: September's decline means failure, or
near-ceiling Q3 means recovery completed.
The alternative frame: volume execution and durable cash repair are
separate stages.
The key research question: does better product mix fund actual
investment without temporary collection/payment support?

CORE_THESIS

REPORTED_FACT: Q2 vehicle margin 9.4% exceeded Q1's 6.1% but remained
below prior-year 19.4%. The matrix reconciles slightly positive CFO to
negative FCF. MANAGEMENT_EXPLANATION:
collection-versus-inventory-payment timing helped cash.
INTERPRETATION: test actual capital requirements and working-capital
quality together. UNRESOLVED_QUESTION: the new quarter's mix and cash
conversion cannot be inferred from its delivery-range finish.

ATTENTION_TRADE_FRAME

Attention Source: Jason / @Jas0nYu;
https://x.com/Jas0nYu/status/2105532543240221122
Why Today: The original three-company comparison highlights the
36-unit upper-end gap.
Attention Stage: Monthly-headline reaction → quarterly context.
Attention vs Evidence — Hard evidence: monthly counts, original
guidance and Q2 cash-flow statement.
Primary sources:
https://ir.lixiang.com/news-releases/news-release-details/li-auto-inc-august-2026-delivery-update
; https://ir.lixiang.com/news-releases/news-release-details/li-auto-inc-announces-unaudited-second-quarter-2026-financial
Attention / interpretation: Useful execution framing, not margin or
FCF proof; no holding disclosure visible.
Attention Path: Correct period → executed volume → margin recovery →
investment coverage.
Attention Asymmetry: A declining month can coexist with strong
original-range attainment.
Crowding Risk: Unmeasured; rebound narratives may overread one completed test.
What Could Sustain Attention: Vehicle-margin gains and cash funding
actual capex.
What Could Make Attention Fade: Repeating guidance attainment while
economics stay unresolved.
Attention-to-Thesis Conversion: Vehicle margin above 9.4% and CFO
covering actual capex with working-capital changes explained.

WEAKEST_ASSUMPTION: Product recovery yields durable contribution and
cash rather than temporary timing benefits.
MOST_IMPORTANT_DATA_POINT: Same-quarter vehicle margin, CFO, capex and
working-capital bridge.
NEXT_DATA_POINT: Next issuer financial disclosure; no date assumed.
Another delivery rebound alone leaves status unchanged; deteriorating
margin plus renewed cash pressure defeats early repair.

## CROSS_OBJECT_ATTENTION_COMPARISON

Cross-Object Attention Comparison

Rank | Object | Attention Asymmetry | Evidence Quality | Catalyst
Clarity | Crowding Risk | Attention Window | Conversion Potential
1 | NIO | Group growth versus period-specific mix | Primary counts;
older financial base | Margin/expense rule defined | Unmeasured | Days
to results | Preserved contribution and operating leverage
2 | XPEV | Small revenue share versus large profit share | Primary
segment lines; future durability open | Attribution request defined |
Unmeasured | Days to disclosures | Two credible engines, separately
reconciled
3 | LI | Monthly decline versus quarterly attainment | Primary
original-range closure | Margin/cash rule defined | Unmeasured | Days
to results | Investment funded by improved operations

Cleanest Attention Trade: LI's period-correction research setup. Most
Evidence-Backed Attention Trade: all three quantity closures; none has
new Q3 financial proof. Most Crowded Attention Trade: unknown. Highest
Fade Risk: XPEV's generic technology narrative. Best Candidate to
Become a Durable Thesis: NIO conditional on its joint test and
valuation. These labels describe research readiness, not transactions.

## 7_DAY_RESEARCH_WORKFLOW

REVIEW_AT: 2026-10-09, Vietnam date. Review dates are not asserted
earnings dates. Retrieve newly dated issuer releases through NIO:
https://ir.nio.com/news-events/news-releases ; XPEV:
https://ir.xiaopeng.com/news-events/news-releases ; LI:
https://ir.lixiang.com/news . Follow exact new disclosure links;
unrelated launches do not refresh financial actuals.

### NIO — 7-Day Checks

- Reconcile July/August/September brand totals and both monthly and
quarterly shares. Preserve the 93.8% net-increment denominator; do not
label it a sales share.
- Queue Q3 vehicle margin and GAAP operating loss together. Record
adjusted earnings separately.
- Request brand/model economics and cash restrictions; if absent,
retain unquantified rather than infer brand margins from price
positioning.

### XPEV — 7-Day Checks

- Preserve unrounded Q2 revenue/cost inputs and recompute both
gross-profit pools.
- Request Q3 vehicle revenue, vehicle costs and service
milestone/contract details. Do not substitute robotaxi headlines for
those lines.
- Maintain automotive, services and consolidated verdicts; record
which one new evidence changes.

### LI — 7-Day Checks

- Close the three-month sum and original September hurdle with source links.
- Prepare CFO−actual-capex and collection/payable/inventory
reconciliation fields for the next disclosure.
- Compare new vehicle margin with Q2 and prior-year bases; a low-base
improvement is not restoration to historical economics.

### Cross-Object — 7-Day Checks

Forward earnings worksheet: one issuer/quarter per row, with vehicle
revenue, vehicle cost, deliveries, services/other revenue and cost,
GAAP operating loss, CFO, capex and working-capital explanation.
Calculate vehicle gross profit from matched revenue/cost; derive
vehicle revenue per delivery only as a within-company proxy, not
reported sticker-price ASP. Reject annual/quarter mixing, a missing
cost line and use of group margin as vehicle margin. XPEV's automotive
comparison uses Q2 gross profit RMB2,058.886 million, while its
consolidated verdict also preserves services value. A missing input
yields UNQUANTIFIED, never zero or automatic failure.

Return ticker | observation_date | retrieved_at | financial_period |
metric/unit | definition | old→new | evidence_class |
calculation_inputs | model_cell | absolute_status_effect |
relative_rank_effect | missing_dependency. Recompute completed
quantities; leave undisclosed price/mix, margin and cash fields
unquantified. If no relevant update exists, return NO_NEW_DATA and
preserve financial period/status. Before October 9 retain that
checkpoint; on/after October 9 set next_review_at seven calendar days
after the actual Vietnam review date, or an earlier newly confirmed
future disclosure. Never reschedule to the current or a past date.

Primary-source retrieval registry: NIO:
https://ir.nio.com/news-releases/news-release-details/nio-inc-reports-unaudited-second-quarter-2026-financial-results
; XPEV: https://ir.xiaopeng.com/news-releases/news-release-details/xpeng-reports-second-quarter-2026-unaudited-financial-results
; LI: https://ir.lixiang.com/news-releases/news-release-details/li-auto-inc-announces-unaudited-second-quarter-2026-financial
; original financial/guidance anchors, supplemented by delivery links
above and forward issuer indexes. Monthly source completion for the
closed ledger: NIO July
https://ir.nio.com/news-releases/news-release-details/nio-inc-provides-july-2026-delivery-update
; XPEV July https://ir.xiaopeng.com/news-releases/news-release-details/xpeng-announces-vehicle-delivery-results-july-2026
; XPEV August https://ir.xiaopeng.com/news-releases/news-release-details/xpeng-announces-vehicle-delivery-results-august-2026
.
Attention-source retrieval registry: NIO:
https://x.com/NioToday/status/2105991833775984977 ; XPEV:
https://x.com/Matteo_Inv/status/2105650021614186825 ; LI:
https://x.com/Jas0nYu/status/2105532543240221122 ; interpretation
sources, never substitutes for issuer data.
Evidence-state snapshot: NIO=quantity closed/mix-expense open;
XPEV=quantity closed/attribution open; LI=quantity closed/funded
repair open; classification=trackable_wow; rule=close completed volume
tests without promoting undisclosed economics; source-rule=primary
facts change business status, attention changes research questions.

## 30_DAY_RESEARCH_WORKFLOW

REVIEW_AT: 2026-11-01, Vietnam date, or earlier on relevant
disclosure. Do not presume results arrive within thirty days. Without
new financials, retain Q2 as the actual baseline and return
NO_NEW_DATA. Before November 1 retain that checkpoint; on/after
November 1 set next_review_at seven calendar days after the actual
Vietnam review date, or an earlier confirmed future event. Review
expiry is not a business miss.

### NIO — 30-Day Checks

Validate: reconcile actual mix, vehicle margin and operating loss.
Upgrade: margin at least 18.5% with loss narrower than RMB347.2
million. Downgrade: one condition deteriorates, keeping the joint
verdict under review. Invalidate promotion: margin declines and loss
widens together. Do not count financing or adjusted profit as the
missing GAAP operating improvement.

### XPEV — 30-Day Checks

Validate: reconstruct both profit pools and their recognition drivers.
Upgrade automotive: margin above 12.1% and vehicle gross profit rises.
Upgrade consolidated diligence separately if durable reconciled
economics improve. Downgrade automotive if margins/contribution
weaken, while preserving any valid services improvement. Invalidate an
automotive-upgrade claim when its required conditions fail;
unquantified robotics cannot rescue it.

### LI — 30-Day Checks

Validate: reconcile margin, CFO, actual capex and working-capital
changes. Upgrade: margin above 9.4% and cash covers investment with
the bridge explained. Downgrade: one leg improves while the other
weakens, leaving funded repair unproven. Invalidate early repair:
deteriorating margin and renewed cash pressure. A lower stock price or
another delivery rebound is not the specified test.

### Cross-Object — 30-Day Checks

Return prior→current rank, decisive source, triggered rule and
strongest alternative explanation. XPEV can precede NIO on durable
consolidated improvement while NIO fails its joint test; LI can
precede unresolved XPEV on funded margin repair. A relative move need
not mean every absolute gate passed. Only then build normalized
earnings/cash valuation cases with current liabilities and dilution.
End with one missing input capable of reversing each verdict, not
three generic requests to watch earnings.

## WKAP DAILY TOP 3

Three market sources worth feeding into today’s market chat. Not
required reading — WKAP has already extracted the signal.

### 1. @NioToday — Launch Attention Must Become Brand Persistence

URL: https://x.com/NioToday/status/2105991833775984977
WKAP signal: Karl's October 2 ONVO persistence critique; explicitly
long NIO since 2020. It is a holder's interpretation, not independent
measurement of customer preferences.
Why it matters today: Fulfilled quantity changes the question to mix
and contribution; the monthly/quarterly distinction constrains both
bullish and bearish readings.
Themes/tickers: Brand demand, model mix, operating leverage / NIO.
Question to ask: Which brand-persistence observation and same-quarter
margin/expense result would change the earnings verdict without
confusing FIREFLY's net monthly increment with total sales or
full-quarter mix?

### 2. @Matteo_Inv — Separate Economics Before Assigning Technology Value

URL: https://x.com/Matteo_Inv/status/2105650021614186825
WKAP signal: October 1 demand for licensing, robotics and robotaxi
economic disclosure; no holding disclosure visible. No technology
valuation or commercialization assertion is adopted.
Why it matters today: The completed vehicle quarter allows a
disciplined earnings bridge, while the services category cannot be
relabeled robotics or spread across every car.
Themes/tickers: Vehicle/services attribution, milestone timing,
durable contribution / XPEV.
Question to ask: What reported revenue/cost and contract evidence
separates an automotive improvement, valuable services milestones and
still-unquantified technology options?

### 3. @Jas0nYu — A Weak Month Can Still Complete a Strong Range Test

URL: https://x.com/Jas0nYu/status/2105532543240221122
WKAP signal: October 1 original three-company guidance comparison
highlights LI's 36-unit upper-end gap; no holding disclosure visible.
Why it matters today: Correctly classifying quarterly execution
prevents a false miss while leaving margin and investment funding
unresolved.
Themes/tickers: Guidance closure, period discipline, funded recovery / LI.
Question to ask: After closing the original delivery hurdle, what
vehicle-margin and same-period CFO-minus-actual-capex evidence would
justify funded repair rather than merely better headline framing?