TL;DR

Delivery Hero is a contested M&A vehicle: Uber built a 25% economic stake, tabled bids at EUR 33 and EUR 38 — both rejected — and is weighing a third; DoorDash is reportedly circling. The stock doubled in 13 sessions on that catalyst and now sits at the 52-week high with RSI 91.9 and a final session close on 0.3× average volume. The bull won the directional thesis (two-path disjunction; controlled-subsidiary SOTP floor near EUR 26–28 from the Talabat stake). The bear won the trade structure at EUR 38.74 (the stop lives inside a documented low-volume node; no listed options exist to dampen a deal-break gap). The only honest expression of both findings is a staged long: no position today, two pre-defined triggers, an explicit hard-kill. Rating: HOLD, conviction 3, size 0.00% NAV now, up to 1.50% NAV on Trigger A or 0.75% NAV on Trigger B.

---

Verdict box

---

Why this verdict

The debate ran two full rounds and produced a split outcome that maps precisely onto the HOLD structure.

The bull won the directional argument on two points the bear could not refute. First, the deal probability is a disjunction, not a conjunction: the relevant calculation is 1 − P(Uber walks) × P(DoorDash never formalises), which sits above 80% under plausible inputs — not the 17–33% the bear computed by multiplying five independent conditionals as if they were independent coin flips. Second, Talabat is a controlled subsidiary (70–75% ownership, consolidated in DH accounts, distributable via German Abspaltung at NAV), not a trapped passive minority stake. The Naspers/Yahoo analogies the bear invoked apply to the latter; a 10–15% DFM illiquidity haircut is defensible, not 30–45%. That sets the SOTP-anchored deal-break floor at EUR 26–28 rather than the bear's opening EUR 20–25.

The bear won the trade structure argument on a point the bull never rebutted. The technical analyst documented a low-volume node (LVN) in the 34.81–35.78 EUR bucket: "any pullback into this zone is likely to be fast and deep — there is no 'soft landing' volume shelf in the 34–36 zone" (technical.md, support section). A stop placed at EUR 33–34, inside the LVN, will not fill near its level on a deal-break headline. A gap-open to EUR 30–32 is the realistic outcome, turning an implied -13% loss into a realised -22% to -28%. The absence of any listed options market (options-flow.md verdict: NO_DERIVATIVES_MARKET) removes the one mechanism that might have dampened the tape. The bull's round-1 "small, stopped at EUR 33–34" defence relied on a mechanic the data explicitly rules out.

The portfolio-manager honoured both findings: refused to express the long at EUR 38.74 (bear wins the entry), refused to rate UNDERWEIGHT or SELL (bull wins the direction). The staged long is not a fence-sit; it is the literal product of two separate half-wins.

---

What we believe

1. The catalyst is pure M&A, not an operating inflection. Macro factors contributed +9.4 percentage points to a +96% one-month move; the remaining +86.6 pp are idiosyncratic (macro-factor.md, factor decomposition). Uber's stake build — publicly disclosed via successive BaFin voting-rights thresholds from roughly 7% to 25% across six weeks — is the sole driver.

2. Uber is not walking cheaply. The stake build committed approximately EUR 2.5B to a 25% economic position. A buyer who has sunk that much capital does not abandon the process; they raise the bid or negotiate a structured exit. Either outcome is individually positive for DHER shareholders relative to a cold walk-away.

3. Standalone fundamentals do not support EUR 38.74 absent a deal. EV/EBITDA 46.8× versus peer median 22.1×, ROIC -4.8%, EBIT still negative at -EUR 226M (2024), interest coverage -0.71×, EUR 10.2B accumulated deficit, and gross margin compressed 550 basis points over two years from 29.9% to 24.4% (fundamentals.md). The 2024 FCF inflection to +EUR 359M is real but modest relative to the multiple the market has applied.

4. The analyst pack has not followed the price. Spot EUR 38.74 sits 16.3% above the mean analyst price target (EUR 33.30) and 23.0% above the median (EUR 31.50). Net analyst rating declined from +0.83 three months ago to +0.59 today — two downgrades, zero upgrades — while the price doubled (estimate-revisions.md, ERB 30d = -0.118). Jefferies' 26 May upgrade to EUR 42.50 is the sole voice above current spot.

5. The short-squeeze has exhausted its primary fuel. Volume peaked on 12 May at 4.5× average daily volume and tapered to 0.2× on 26 May, the day of the 52-week high. The gap-down-open / intraday-recovery pattern in 10 of 13 surge sessions identifies short-covering as the dominant volume mechanism (technical.md). Fresh institutional buyers at EUR 38.74 are risk-arb funds at par with a rejected bid — a structurally thin cohort.

6. The SOTP floor is real and bounded at EUR 26–28. Talabat's DFM market cap was AED 25.6B (~EUR 6.0B) at the time of analysis, with DHER holding 70–75%, implying EUR 4.2–4.5B before any illiquidity discount (news.md item #4). As a controlled, consolidated subsidiary distributable via Abspaltung, a 10–15% discount is appropriate, not the 30–45% holdco discount that applies to trapped passive minority stakes (debate, round-2 bull). The honest deal-break floor is EUR 26–28, not EUR 20.

7. EM internet is the most reliable macro factor and it is a headwind. EMQQ has returned -11.1% over three months, directly weighing on Talabat's GCC operating environment and Foodpanda's Asia footprint. Beta +2.18, R² = 0.333 — this factor explains more of DHER's non-M&A return variance than any other (macro-factor.md). If the deal stalls and EMQQ deteriorates further, Uber gains narrative cover to revise terms.

---

What would change the call

Trigger A fires — convert HOLD to LONG at 1.50% NAV: DHER.DE closes inside 34.44–36.11 EUR on daily volume of at least 1,470,899 shares (1× 50-day ADV) with at least one stabilising bar (hammer, doji, or two consecutive closes above 34.44 EUR). The volume requirement is non-negotiable: a low-volume drift into the LVN and a genuine tape-rebuilding pullback are indistinguishable on a closing-price chart but have opposite implications.

Trigger B fires — convert HOLD to LONG at 0.75% NAV: A news-confirmed Uber third bid at EUR 41 or above with Delivery Hero board engagement language (not rejection) — or a formal DoorDash approach above EUR 40 with board engagement — accompanied by a first daily close at or above EUR 39.00 on volume of at least 1.5× ADV (2,206,348 shares). Do not chase above EUR 39.50.

Hard-kill — stand down, escalate: All three of Uber files a formal withdrawal AND a DoorDash denial-of-interest wire is published AND the European Commission opens a Phase II review — within any 30-day window. If positioned: exit at next open, do not wait for the stop. If not positioned: remove from watchlist, escalate to portfolio-manager for a fresh /decide run that explicitly considers a tactical short above EUR 41.

Additional invalidation conditions (any one fires):

• Uber unwinds its 19.5% block back into the market.
• Talabat trading halt on DFM or material write-down of its food delivery operations.
• DHER management issues a profit warning or reduces FY 2026 guidance.
• Jefferies PT cut below EUR 35 within 30 days.
• DHER.DE daily close below EUR 31.68 on volume above 2× average (technical regime impairment; blueprints void).
• No resolution by 2026-07-27 (mandatory expiry; re-research from scratch).

---

Watchlist disposition and monitoring cadence

Disposition: ADD — staged long, both triggers pre-armed.

Primary monitoring window: 5 June to 15 July 2026. Daily review for any M&A headline, BaFin voting-rights notification, or Talabat DFM price action.

---

Limitations and data gaps

• No listed options on DHER.DE. Eurex returns zero expiries; DLVHF OTC carries none either. Stop-loss discipline is the sole risk-management tool, and the LVN gap-risk problem flows directly from this absence (options-flow.md).

• No same-ticker history in the journal. This is the first tracked decision under both the DHER.DE and DLVHF aliases. First-call cap of 4 applies; conviction is held at 3 by judgment. No same-ticker calibration exists.

• Estimate revision momentum (ERM) not computable. yfinance does not publish historical consensus time-series for DHER.DE. The revision trend verdict of FLAT carries conviction 2/5; ERM_30 and ERM_90 — the most important inputs — are missing (estimate-revisions.md).

• Short interest data unavailable from configured providers. yfinance returns null for all European short-interest fields; Bundesanzeiger/ESMA scraping was not performed. The squeeze-potential assessment of MEDIUM rests on volume behaviour and days-to-cover proxies only.

• German retail social channels are under-sampled. wallstreet-online.de and Comdirect community were not reachable via public API. Reddit is biased toward English-comfortable investors and likely undercounts domestic retail conviction (sentiment.md).

• FY2025 full-year figures were not extracted from the DHER IR page (SvelteKit SPA returns only JavaScript on server-side fetch). Fundamental analysis rests on FY2024 annual figures plus yfinance TTM snapshots.

• BaFin voting-rights notifications not directly accessed. The Uber stake-build timeline is inferred from price action and secondary news reporting, not from primary BaFin Stimmrechtsmitteilungen filings.

• Morgan Stanley equity-swap stake (~27%) cited from secondary sources and not independently verified from a primary filing.

---

Source files

All artefacts under data/reports/DHER.DE/2026-05-27/.