EQIX · Equinix, Inc. 30d old
conviction3 → 3.00 trust-weighted
coverage tierT1-FULL
horizon12 months
Return distribution probability-weighted scenarios, never collapsed to a point
EV +3.63%
why these probabilities: Returns are fixed at the falsifier-derived values and are not re-derived here. The probabilities are the judgement. p_bear 27% for a -28% drawdown, above both the historical (10.1% at -30%) and node-pooled (18.3% at -30%) anchors, because the specific mechanism is asymmetric: this is a 23.6x EV/EBITDA multiple on 4.4% ROIC where the entire holding case is delivery of a guided long-term AFFO path. A REIT whose equity story is a raised 2029 outlook re-rates violently when bookings decelerate, and the two things that could cause that — hyperscaler self-build substitution and churn breaching the 2.5% range — are both live in the evidence (EQIX/EQIX_2026Q2.json:33 on local vendors and sovereignty-dependent EMEA growth; 982c33d3d8:9 on churn guided to the low end, which is a guide, not a delivery). p_base 51% is the modal case because the bookings and backlog evidence is genuinely strong right now (982c33d3d8:7) and near-term revenue is contracted recurring billing that does not turn quickly. p_bull is held to 22% rather than higher because the bull requires the enterprise-AI colo acceleration to be evidenced, and the only support in this pack for it is management's own qualitative framing (EQIX/EQIX_2026Q2.json:45) — that is a Class A statement of opinion, not booked capacity. Evidence that would raise p_bull: disclosed enterprise-AI bookings as a named component of gross bookings,
not built — options-implied distribution on the same axis, with the KL number and where the disagreement sits
needs: desk-options implied bands (v1.4 Part 4a) — activates at 60% coverage; currently 33%
Visibility how far ahead this name can be seen
unscorable scored on 0/4 components
horizonabsent — no backlog/contract language found in the pack
dispersion_tightnessabsent — only 1 broker in unit $MM — dispersion undefined
delivery_reliabilityabsent — guidance_vs_delivery returned nothing
tape_stabilityabsent — revision tape EMPTY (0 up, 0 down) — not flat, ABSENT
Trust how much weight this name's own record can carry (credibility Z)
guidance_deliveryunavailable
composite0.251
Technology supply chain graph-backed context · read-only
chain / roledc-infra · infra
upstream suppliersVRT, ETN, PWR
downstream customersMSFT, GOOGL, AMZN, META, ORCL, NVDA
centrality / systemic importance0.8571
chokepointsnone declared
Evidence refs: EQIX->MSFT, EQIX->GOOGL, EQIX->AMZN, EQIX->META, EQIX->ORCL, EQIX->NVDA. Source: policies/supply_chain.yaml v2 — NOT Futu get_industrial_chain_*, which does not exist in this SDK (verified: only get_owner_plate / get_plate_list / get_plate_stock). Our own graph carries dated rationales per edge, which a vendor plate list does not.; generated 2026-09-16T18:49:43.027106+00:00.
Sizing survival-first size (half-Kelly) vs the enforced ladder
kelly not computable — no data
Payout durability does the distribution survive a downturn?
FAILDPS trajectoryrecovered_after_cut
interest cover now3.25x
interest cover at FORWARD rates1.74x
interest coverage at forward rates 1.74x < 2.0x (this is the test a static ratio cannot see); FCF <= 0: payout ratio undefined; treat as failing
Falsifiers what would prove this wrong
| Gross bookings growth remains positive year-on-year, sustaining the record backlog | | |
| MRR per cabinet continues to rise, evidencing pricing power rather than volume-only growth | | |
| Churn stays within the stated 2%-2.5% target range | | |
| Leverage stays at or below roughly 4x while funding the Build Bolder capex programme without equity issuance | | |
| Hyperscaler self-build does not displace leased retail colocation demand at EQIX | | |
| The raised 2027-29 outlook toward 2029E AFFO/share of $58 is maintained, not walked back | | |
| Utilisation does not fall as new capacity lands, i.e. the 52-project programme fills | | |
Conflicts surfaced, never netted out (T4)
I cannot check my view against the revision tape because there is no tape. eps_drift and rev_drift both return n_up 0, n_down 0, breadth 0.0, direction flat, median_move_pct null over a 365-day window, and outliers is em | high |
My base case of +9% to ~$1,201 is BELOW Morgan Stanley's $1,250 PT dated 2026-07-30, which carries an OW rating and is described as 3x risk/reward skew (f0a48e5516:3). I do not share the skew view: with the stock at 97.6 | medium |
No options view is possible. The pack reports no usable option chain for EQIX, which also nulls the options_implied probability anchor — the strongest anchor by the pack's own rule. My p_bear of 27% therefore has no mark | high |
Management is guiding churn toward the LOW end of the 2%-2.5% range for 2H26, which Bernstein attributes to renewal process execution and SOME DELAYED CHURN (982c33d3d8:9). Delayed churn is timing, not avoidance. Managem | medium |
The demand-cascade evidence is large and FACT-grade at the CUSTOMER — MSFT $368bn contracted backlog (MSFT/MSFT_2025Q4.json:39), GOOGL $514bn cloud backlog (GOOGL/GOOGL_2026Q2.json:31), AMZN capex laid out ahead of monet | high |
My quality score of 7 rests on interconnection switching costs and MRR/cabinet pricing power, but management itself describes competitive pressure — local vendors moving into the Southeast Asia market, and EMEA growth un | medium |
Bookings and backlog are accelerating (gross bookings +23% y/y, second-highest on record, record backlog — 982c33d3d8:7) while reported recurring revenue in the one region disclosed in this pack is DECELERATING (APAC rec | medium |
I scored management_delivery 6 on transcript evidence of honest disclosure and a delivered guidance raise, while Deutsche Bank names recent management turnover and strategy implementation uncertainty as a live thesis ris | medium |
The case against independent adversary, different model family (P6) weakened
Version history from the research record
archived versions1
| archived | hash | bytes |
|---|
| 2026-08-17T10:34:17 | fa3ab55edc39 | 36,083 |
One version only — the archive was created today. The diff view fills in on the next re-run.