LULU › analyze
LULU · Analyze from before
2026-09-10 · $96.88 · Mkt cap $10.73B · EV $11.48B · Consumer Cyclical / Apparel Retail
Updates: Output/Stocks/Consumer/LULU/analyze-2026-08-11.md (WATCH, conviction 5.0, FV $95–150, entry $100–110, trim 16x fwd)
1. What this updates
The baseline (2026-08-11) was written explicitly to wait for one dateable event: the Q2 print and the arrival of new CEO Heidi O'Neill, calling it "the highest-probability negative catalyst in today's entire batch." That event has now happened.
Event list since baseline (2026-08-11 → 2026-09-10): 1. Q2 FY print, Sept 3-4 2026 (after-close Sept 3, reaction Sept 4) — delivered by interim co-CEOs Meghan Frank and Andre Maestrini, not by O'Neill, who had not yet started. 2. Second FY guidance cut of the year — revenue to $10.35–10.50B (from $11.00–11.15B, itself already a cut from the original $11.35–11.50B); EPS to $9.48–9.73 (from $10.95–11.15). 3. Stock fell ~17.8–20% on the print, then continued lower to $96.88 — through the old $100–110 entry band, through the $95.67 fifty-two-week low's neighborhood, and to an 8-year low. 4. Heidi O'Neill officially started as CEO on 2026-09-08 — four days after the print she did not author. 5. A wave of analyst price-target cuts: JPMorgan −38% to $95 (from $154), UBS to $106 (from $120), at least one outright downgrade to Sell. 6. Michael Burry reaffirmed and extended his position rhetoric — LULU is now his single largest holding (~17.4% of his book, down >52% YTD on it), and he has stated intent to "aggressively buy" more below $100. 7. No re-escalation of the Chip Wilson matter; the Oct 1 2026 third board seat deadline from the May settlement has not yet occurred as of this report's date.
The central finding of this pass: the Q2 print did not resolve the baseline's coin-flip. It moved both sides of the ledger at once — the trend evidence got worse (a sixth consecutive negative Americas quarter, China's comp flipping negative, the deceleration accelerating into Q3 guidance) while the valuation got cheaper enough to roughly re-base to fair, not cheap. The baseline pre-registered an explicit decision rule for exactly this scenario, and it fires below.
2. The delta ledger
Lead rows are the ones that moved status. Full ledger follows.
🔄 SUPERSEDED / 📉 DRIFTED (lead with these)
| # | Claim (baseline) | Status | What changed |
|---|---|---|---|
| 1 | "Five consecutive quarters of declining Americas comps" — recoverable merchandising stumble, not yet structural | 📉 DRIFTED, at the baseline's own stated breakpoint | Q2 Americas comp −12% (from −6% in Q1) — the sixth consecutive negative quarter, on top of a second FY guidance cut. Baseline's own pre-committed rule: "a sixth negative Americas quarter with further guidance cuts → downgrade toward 3.5 and treat as structural." Both conditions are now met. See §3. |
| 2 | "China remains the offset, but de-risked only on paper" — underlying comp ~+5% ex-CNY-timing, decelerating | 🔄 SUPERSEDED — the offset broke, not just decelerated | China Mainland comp swung from +13% headline / ~+5% underlying (Q1) to −8% (Q2). Revenue still grew +4% (new-store openings), but same-store sales in the one market carrying the growth story are now negative. This is not the gradual deceleration baseline modeled — it's a sign flip. |
| 3 | "Management expects markdowns to improve modestly in FY26, year-over-year decreases not appearing until 2H26" — i.e., the worst is roughly known | 🔄 SUPERSEDED — the trajectory is deteriorating, not stabilizing | Q3 guide: revenue −10% to −11% YoY (vs Q2's −4%, Q1's +4%). The rate of decline is accelerating two guidance cuts into the year, not decelerating as the Q1-era commentary implied. |
| 4 | "Q1 gross margin −410bps YoY, largest compression in the company's public history" — margin damage is tariffs + deliberate markdowns | 🔄 SUPERSEDED by a bigger, disguised number | Q2 gross margin reported +200bps YoY (60.5%) — but this includes a 560bps benefit from a one-time $134.5M IEEPA tariff refund. Stripped of the refund, organic gross margin fell ~360bps YoY, and organic operating margin fell ~750–760bps YoY (reported op margin −190bps despite the same 560bps refund benefit). The underlying compression worsened materially versus Q1's already-record −410bps; the headline "margin improved" print is the exact inverse of what happened economically. |
| 5 | Graham IV $102–108 (trailing/guided-Aug), stock "16–23% above fair value" | 🔄 SUPERSEDED — Graham now sits at spot, not above it | On the new FY guide midpoint ($9.605 EPS) and current BVPS ($42.90): Graham = √(22.5×9.605×42.90) = $96.29 — within 1% of the $96.88 price. See §6. The prior "overvalued" read is gone; it has re-based to "fair," not "cheap." |
| 6 | Q2 headline EPS "beat" narrative (implicit in a WATCH-not-AVOID stance) | 🆕 NEW finding, files as a force against trusting the print at face value | Reported adjusted EPS $2.92 beat consensus $1.82 by $1.10 — but $0.86 of the $2.92 came from the tariff refund. Ex-refund EPS ≈ $2.06, a ~34% YoY decline (vs $3.10 py), not the ~6% the headline implies. Per Knowledge/Playbook/pitfall-forward-pe-above-trailing-pe-flags-an-inflated-base and its sibling on stale forward EPS: a non-repeating credit sitting in a reported "beat" is exactly the trap this repo has caught before (NVO, HON). Same shape, different ticker. |
✅ CARRIED
A portfolio-specific passage was removed from the public build.
⏳ UNTESTED
| # | Claim | Why untested this pass |
|---|---|---|
| 17 | "New-CEO kitchen sink" risk — baseline's single largest near-term risk | Did not fire as predicted, but is not resolved either. The Sept 3-4 guide cut was delivered by interim leadership before O'Neill's Sept 8 start. Her actual strategic reset — the one analysts are waiting for and one source describes as having "failed to impress" so far in preliminary form — has not yet happened. This is the same risk, deferred one quarter, not retired. Flag for the next pass: if this stays UNTESTED a second consecutive time, that itself is informative (a company that keeps deferring its own reset). |
| 18 | Precise TTM P/FCF on the new data | Yahoo's quarterly cash-flow feed has a data gap on the year-ago Q1 (Apr 2025) row (nulled fields), the same class of gap baseline flagged for FY22 annual data. Q2-vs-Q2 is clean and reported (below); a reconciled trailing-twelve-month figure is not. |
| 19 | Magnitude of DTC/foot-traffic share loss to Alo/Vuori (baseline flagged as "contested," direction established, magnitude not) | No new dataset found this pass. Direction (share loss real) remains asserted from the baseline's single tracked-panel source; not independently re-verified. |
🆕 NEW
| # | Finding | Detail |
|---|---|---|
| 20 | Q2 cash generation actually improved YoY | FCF $225.2M vs $150.8M py (+49%), OCF $374.8M vs $328.7M py (+14%) — same-quarter, apples-to-apples comparison. Driven by the tariff-refund cash inflow, lower capex ($149.7M vs $177.9M py), and working-capital discipline from the inventory cuts (#9). This tempers the "cash flow is structurally breaking" read from the FY26 annual figure (still true as a historical fact — see #21) with a genuinely positive in-quarter data point. |
| 21 | FY26 (year ended Jan 2026) FCF −42% remains true as history | This is an already-closed fiscal year; nothing in Q2 changes it. It is the base rate the market re-rated off in the June cut, still valid as a historical fact distinct from #20's in-progress-year signal. |
| 22 | O'Neill started Sept 8, four days after, not before, the guide cut | Timeline detail baseline could not have had — it assumed a Q2 print landing "within days of" the CEO start and treated the ordering as ambiguous. It is now clear the interim team owns this cut, not O'Neill. |
| 23 | Analyst target compression has converged toward spot, not stayed well above it | JPMorgan $95, UBS $106, Yahoo mean $107.43 — targets now cluster close to or below the current price, versus baseline's wide $115–148 scatter. Read two ways: (a) capitulation / most bad news now priced by the sell-side, or (b) the sell-side has simply caught down to the stock rather than the stock catching up to a floor. Genuinely ambiguous; not a strong signal either way. |
3. How the close calls were decided
The central judgment: does the sixth negative Americas quarter plus a second guidance cut make this structural, per the baseline's own pre-registered rule?
The baseline did not leave this to post-hoc interpretation. It wrote, in its own verdict section: "a sixth negative Americas quarter with further guidance cuts → downgrade toward 3.5 and treat as structural." Both conditions are satisfied by the Q2 print: Americas comp −12% is the sixth consecutive negative quarter (Q1 was stated as the fifth), and the FY guide was cut a second time in the same year. This is not a new judgment call — it is honoring a decision the analysis already made in advance of the evidence, which is the entire point of pre-registering a trigger. The alternative — re-litigating the threshold now that the number has actually printed — is exactly the kind of post-hoc rationalization the framework is built to avoid.
Weighed against this: the two strongest counter-forces are inventory (#9) and cash generation (#20), both independently verified, both pointing the same direction (operational discipline intact, no demand-collapse signature of bloated stock or cash burn). These are not enough to overturn the structural read outright — they answer "is management executing competently through the decline," not "has the decline stopped" — but they are the reason this lands at WATCH, conviction 3.5 rather than AVOID. A business with this Americas trend and deteriorating inventory and deteriorating cash generation would be a clean AVOID. This one only has the first.
Second close call: does the tariff refund's one-time nature retract the Q2 "beat," or just annotate it? Resolved as an annotation with real weight, not a full retraction of the quarter — revenue still missed estimates independent of any EPS adjustment, and the guide cut (the actual market-moving event) was not affected by the refund accounting at all. The refund matters for one thing specifically: don't let the "$2.92 beat" headline anchor the fair-value math, which is why Graham below is run off the FY guide, not off trailing EPS.
Third close call: the tariff regime. Initial read of "Supreme Court invalidated the tariffs that were crushing margins" looked like a structural tailwind reversal. Checking the company's own guidance assumptions (20% tariff rate embedded for H2) rejected that reading — the refund is retroactive and one-time; the forward tariff burden is comparable or larger under new legal authority. Recorded as CARRIED (#15), not SUPERSEDED, because the underlying economic force (tariffs pressure margins) is unchanged even though its specific legal vehicle changed.
4. Thesis persistence and conviction delta
Persistence: of the ~16 Structural/Trend/State claims tracked, 9 CARRIED, 3 SUPERSEDED, 1 DRIFTED at the trigger point, 3 UNTESTED, 2 NEW. The Structural rows that carried (switching costs, competitive positioning, balance sheet strength, inventory discipline) are the "is management executing" layer. The rows that broke or drifted (Americas trend, China offset, margin trajectory, guide credibility) are the "is the top line still declining" layer — and that is the layer the verdict is actually about. This is moderate-to-low persistence concentrated in exactly the claims that matter most, which is the honest reading, not a good one.
Conviction: 5.0 → 3.5. Driven specifically by: - Row #1 (sixth negative Americas quarter, baseline's own pre-registered trigger) — the primary mover. - Row #2 (China offset flipped negative, not just decelerating) — removes the "second engine" bull argument almost entirely. - Row #3 (guide deceleration accelerating into Q3, not moderating) — directly contradicts the "modest 2H26 improvement" framing the baseline carried forward from management. - Row #4/#6 (organic margin and EPS both meaningfully worse than the reported headline) — the "beat" was optical.
Offset partially by rows #9 and #20 (inventory, cash generation), which is why this is 3.5, not lower — a full structural-collapse read would put this in AVOID territory, and the evidence does not support that yet.
5. What is genuinely new
- The $134.5M IEEPA tariff refund and its accounting mechanics (rows #6, #15) — a data-quality trap of the same shape this repo has already priced twice (NVO, HON): a non-repeating credit inflating a reported "beat."
- Q2's cash-flow improvement (#20) despite the demand deterioration — a genuinely two-sided fact that complicates a simple bear thesis.
- The CEO-transition timeline detail (#22): the guide cut was delivered by interim leadership, not O'Neill, meaning her actual reset is still ahead and untested (#17).
- The newly-disclosed $600M undrawn revolver (#10) — incremental balance-sheet cushion baseline did not have.
6. Updated verdict
WATCH — conviction 3.5/10 (down from 5.0). Do not buy into an unresolved, worsening trend at a price that has only re-based to fair, not cheap.
Valuation, recomputed from scratch per protocol:
Multiples off the current, non-stale numbers (verified: implied forward EPS from the vendor's displayed PE(fwd) of 10.69x is $9.06 — already below the company's own new guide low end of $9.48, meaning the vendor feed has already caught down and is not exhibiting the stale-post-crash lag this repo has previously caught on other names; safe to cross-check against the company's own guide rather than discard):
- P/E(ttm) 8.21× on EPS $11.80 — inflated by the one-time tariff refund; clean trailing P/E ≈ 8.86× on an EPS of ~$10.94 with the refund's $0.86 stripped out.
- P/E(fwd) 10.0–10.2× computed directly off the company's own new FY guide ($9.48–9.73, midpoint $9.605) — consistent with, and very close to, the vendor's displayed 10.69×.
- P/B 2.26×, BVPS $42.90.
Graham IV — the single most important number in this pass. On the new FY guide midpoint: √(22.5 × 9.605 × 42.90) = $96.29, essentially identical to the $96.88 price (within 1%). Across the full guide range ($9.48–$9.73), Graham lands $95.66–$96.91 — the stock has re-based from "16–23% above Graham" (baseline) to "at Graham," which is a genuine and material re-rating, but it lands at fair, not at a discount. The framework's own test — great company + cheap = value, poor company + cheap = value trap — needs the second half of that sentence answered before the first half matters, and "fairly priced" doesn't answer either half.
Fair value range: $74.00 – $145.00, center ≈ $96 (essentially at spot — a genuine change from baseline, where spot sat above the range's center).
- Bear $74 — the guide proves too optimistic a third time (it has already been cut twice this year); FY EPS resets toward ~$8.25; Americas comps stay double-digit negative through FY28; China comps turn structurally negative rather than one bad quarter; full-price selling never returns; multiple compresses to 9× on a credibility discount.
- Base $96 — the company hits its own $9.48–9.73 guide; multiple holds at Graham parity (~10×); the market keeps pricing a stabilized, no-further-growth Lululemon, same shape as baseline's reverse-DCF conclusion, just re-based lower.
- Bull $130–145 — O'Neill's product reset (newness 23%→35%) gains real traction in FY28, Americas comps inflect toward flat, China comps recover to mid-single-digit growth, markdown intensity declines, EPS recovers toward $11–12 on a 12–13× multiple as the brand question starts to resolve favorably.
Entry: $75.00 – $85.00 — below Graham and below the base case, inside the bear case. This is a deliberately tighter and lower band than baseline's $100–110, because the new information is net negative on trend even after accounting for the price decline: buying at "fair" on a name with an accelerating negative trend and a two-time-broken guidance record is not a margin of safety, it's a bet that the trend inflects from here with no evidence yet that it has.
Trim: 13× forward earnings (down from 16×) — reflecting reduced confidence that the brand still commands a premium multiple even in a successful recovery scenario; re-assess upward if Americas comps actually inflect.
Break triggers (would justify a further downgrade toward AVOID): - A seventh consecutive negative Americas comp quarter. - A third FY guidance cut, especially one delivered by O'Neill herself after her "deep dive." - China comps staying negative for a second consecutive quarter (confirming a sign-flip, not a blip). - Inventory units turning up while comps stay negative (the demand-collapse signature this report currently lacks).
Upgrade conditions (would justify moving back toward 5.0+): - Americas comps inflecting less negative sequentially (e.g., −12% → high single digits) in Q3. - China comps returning to positive. - O'Neill's strategy update landing with specifics analysts credit, without a further guidance cut. - The stock trading into the $75–85 entry band with the above intact.
7. What this pass did NOT test
- Row #17 (new-CEO kitchen sink) — genuinely deferred, not resolved. This is the second time a leadership-driven reset has been anticipated and not yet delivered (first at the April CEO announcement, now at the Sept 8 start); worth flagging explicitly per protocol that a twice-deferred question is itself informative about how this company communicates bad news.
- Row #18 (precise TTM P/FCF) — a Yahoo data gap on the year-ago Q1 quarter blocks a clean reconciliation; the Q2-vs-Q2 comparison (#20) is solid, the full trailing-twelve-month figure is not and should not be quoted without flagging the gap.
- Row #19 (DTC/foot-traffic share-loss magnitude) — still resting on the single baseline source; no independent re-verification attempted this pass given time allocation went to the print itself.
- Chip Wilson's Oct 1 board seat — has not yet occurred as of this report; the next pass should check it explicitly rather than assume the settlement holds.
- Single-source items carried at face value this pass: the "20% tariff rate for H2" guidance assumption and the "$105M excluded from guidance" figure both come from one aggregated search synthesis, not a direct primary-source read of the 10-Q or earnings call transcript. Recommend the next pass pull the 10-Q (filed, per SEC EDGAR search results) directly if the tariff assumption becomes decision-relevant.
Next check: Q3 FY print, estimated ~2026-12-10 based on last year's Dec 11 2025 cadence — verify exact date on IR. The real test is whether O'Neill's first full quarter shows any inflection at all, or a third guidance cut.