Built by Grok 4.6 HighEquinix reports Wednesday, October 28, 2026. Prepared Monday, October 26, 2026. This is fiscal 2026Q3, the quarter ended September 30, 2026. It has not been reported. The conference call is at 5:30 p.m. Eastern, after the close. The Wednesday session trades before the release. The reaction session is Thursday, October 29. Last reported quarter is 2026Q2, released after the close on July 29, 2026. Equinix is a calendar-year REIT. The homepage calendar labels October 28 as 2026Q4. That label is the announcement-month quarter, one quarter ahead. This page is 2026Q3.
EQIX | Earnings Preview
| Q3 revenue guide | $2.525–$2.575 billion · mid $2.550 billion · +10.1% vs Q3 2025 | Q3 MRR, normalized CC | 9–11% · midpoint 10% · 100 bps under the 11% Q2 print |
| Q3 adjusted EBITDA | $1.275–$1.315 billion · ~51% margin · +100 bps vs Q3 2025, not 53% | What is aggressive | FY capex $5–6 billion · 2027–29 revenue growth 10–13% · not a 2026 earnings item |
Equinix reports a confirmed Wednesday night print of a quarter that is already guided, and nothing since July 29 has moved the range. The live bar is revenue of $2.525 billion to $2.575 billion, adjusted EBITDA of $1.275 billion to $1.315 billion, and a margin of about 51%. The midpoint, $2.550 billion, is +10.1% versus Q3 2025 revenue of $2,316 million. Spoken growth is 9–11% as reported and 10–12% normalized and constant currency. FMP revenue of $2,563 million (17 analysts, pulled October 26, actuals null) sits inside the range, +0.5% versus the midpoint and about $12 million under the high. That is slightly conservative versus the street, and it is not a sandbag. Over the last seven settled quarters, revenue beat the guide midpoint only 3 times and cleared the high only once.
Growth trajectory, same quarter versus the prior year. The franchise KPI has been accelerating, and the guide stops it. Normalized constant-currency MRR went from 7% in Q2 2025 to 8%, 10%, 10%, and 11%. The Q3 band of 9–11% puts the midpoint back at 10%. That is a guided 100 bps deceleration off the exit rate, still +200 bps versus the year-ago 8%, and it matches the full-year line of “around 10%.” It is not a continuation of the acceleration.
Reported revenue did the opposite of that path in Q2, and the reason is a fee. As-reported growth was +16% on $2,625 million versus $2,256 million. Recurring revenue was $2,377 million versus $2,143 million, +10.9% on those two figures. Non-recurring revenue was $248 million versus $113 million. The call put about $120 million of that on 134 megawatts of xScale leases, including Hampton. The Q3 dollar midpoint is $75 million below the Q2 print because that fee is not in the guide, not because demand was cut.
As-reported recurring dollars already slowed before this guide. Recurring revenue growth was +7.6% in Q3 2025 ($2,215 million versus $2,059 million), +9.7% in Q4 2025 ($2,294 million versus $2,091 million), +11.7% in Q1 2026 ($2,331 million versus $2,087 million), then +10.9% in Q2. The normalized MRR rate still stepped up. The dollar rate did not. Do not carry +16% or +11% into Wednesday as if they were the same series.
Margin expanded, and the guide gives the fee back. Adjusted EBITDA margin printed 53% in Q2 versus 50% a year earlier, +300 bps. The deck said about half of that was xScale leasing fees. The Q3 point of 51% is about +100 bps versus Q3 2025’s 50%. The 53% print is the fee quarter. It is not the comparison. The full-year margin has been 51% at every version of the 2026 guide. They did not spend the revenue raise on a higher margin percent.
Diluted AFFO per share is the line management steers, and it is not what FMP scores. It accelerated from +8.6% in Q3 2025 ($9.83) to +12.5%, +11.6%, and +18.9% in Q2 2026 ($11.78 versus $9.91). Part of the last step is the fee. The annual hurdle is 10–12% normalized constant-currency AFFO per share, on a range of $42.69 to $43.29. There is no quarterly AFFO guide. A Q3 rate well below +18.9% is what the year already assumes.
What the street is underwriting. FMP is the only consensus source in this workspace. Visible Alpha and Bloomberg are not connected. FMP’s $4.11 is GAAP diluted EPS, not AFFO per share. Q2’s FMP actual was $4.83 against diluted AFFO per share of $11.78. FMP Q3 EBITDA of $1,205 million sits about $70 million under the company’s adjusted-EBITDA low. That is a definition gap, not a 7% miss versus the guide. FY2026 FMP revenue of $10,277 million is +0.3% versus the $10.245 billion midpoint and inside $10.205 billion–$10.285 billion, near the top. FY2027 FMP revenue of $11,376 million is +10.7% versus that FY2026 average, the bottom of the 2027–2029 band of 10% to 13%. The street has caught 2026. It has not underwritten the acceleration management says arrives toward 2029.
The contradiction worth listening for. Management’s tone is the most confident of the year, and on September 29 Olivier Leonetti said the company had undersold itself and that “the best is to come.” He also said the three-point raise in the through-2029 growth profile is demand already in hand, not the extra capex, and that the extra capex is a drag until sites are built and filled. The street’s FY2027 number sits on the low end of 10–13%. That is the management-versus-street gap. It is not a reason to expect Wednesday’s revenue to clear $2.575 billion. He has already said that capital does not earn its keep in 2026.
What to watch on October 28
- MRR versus 10%, not total revenue versus 16%. A hold at 11% says the Hampton fee was the only one-off. A print at 9% is inside the spoken band and breaks the “third straight double-digit quarter, and still there” claim. The year can still be “around 10%” either way.
- Non-recurring revenue without another $120 million. Q3 2025 non-recurring was $101 million. The long-term planning assumption on the call was about 5% of revenue. A revenue make that required another outsized fee is a lower-quality make.
- Adjusted EBITDA margin near 51%. Another 53% quarter means the fee did not drop out. A margin under 51% spends the cushion the full-year point does not have.
- Churn toward 2%, not another 1.8%. Q2 churn of 1.8% was helped by delayed churn. The back half was guided to the low end of 2–2.5%. A move through 2.5% is outside that sentence.
- Bookings against a harder comp, and interconnection adds. Q2 annualized gross bookings were $424 million, +23% versus a soft $345 million. The Q3 comp is $394 million. A flat $424 million is only about +8%. Net interconnections added 9,700 in Q2, a record on the call, versus about 7,100 in Q3 2025. There is no street number on either line.
- Cabinets are a Q4 item. Deliveries are supposed to double in the second half, and more than 7,000 cabinets were pulled from 2027 into Q4. New cabinets do not bill the quarter they are delivered. Q3 revenue does not collect that pull-forward.
- The capex caveat, as commentary. FY2026 capex of $5.0 billion to $6.0 billion replaced an opening plan of $3.655 billion to $4.155 billion. A raise that starts attributing 2029 growth to that extra spend would contradict the September 29 comment.
There is no tariff line in the guide. The cross-border item already booked is foreign exchange: about $49 million of negative FX in the full-year revenue bridge, and about $21 million of negative FX on the Q3 guide versus the prior quarter’s FX rates, after hedges (Q2 2026 earnings release). The operating constraint since the print is power — who pays for the grid — not a duty rate.
No U.S. retail-colocation peer prints first. Keppel DC REIT has already reported a supportive demand update on a different lease structure. Crown Castle has reported and does not speak to cabinets or xScale. American Tower reports Tuesday morning; use the data-center segment, not tower AFFO. Digital Realty, the closest U.S. comp, reports Thursday after the close, on Equinix’s reaction session, so it cannot preview this number and it does not contaminate that session.
Internal models, SharePoint, OneNote, Outlook, and Excel were unavailable. The +3.9% figure is the July 30 reaction session versus the July 29 close, from Daloopa adjusted prices. It has no fundamental id. It is not a multiple, and it is not this quarter’s setup.
Equinix guides the current quarter and the current year. Quarterly guidance issued on the Q2 call is the outlook for Q3. It is not compared with the Q2 actual. The July 29 ranges are still the bar. The company’s own practice, repeated on the last three calls, is not to update guidance during the quarter unless it does so in an explicit public disclosure. A search of calendar 2026Q3 filings found no revision. Classification of the Q3 revenue range: slightly conservative versus FMP, and not a sandbag. Slightly soft on the MRR rate, because the midpoint steps off 11%. The aggressive statement is the capex reset and the 2027–2029 algorithm, which management says will not show up in 2026 earnings.
The spoken 10–12% and the dollar range are not the same math. $2.525 billion–$2.575 billion versus $2,316 million is +9.0% to +11.2% as reported. The dollar midpoint is the low end of the spoken normalized band once FX and normalization are stripped out. Do not treat 12% as the as-reported bar. FMP’s $2,563 million is +10.7% versus that Q3 2025 base, about 76% of the way from the low to the high. A print at the midpoint is a small miss versus FMP and a hit versus the guide. A print at the low is a guide make and about a 1.5% consensus miss. A print above the high would be the exception. Q2 was that exception by $14 million, and the guide already contained the Hampton economics.
| Q3 2026 guide | Q3 2025 | Low | High | Mid | YoY at mid | Versus FMP |
|---|---|---|---|---|---|---|
| Total revenue | $2,316 | $2,525 | $2,575 | $2,550 | +10.1% | FMP $2,563 million, +0.5% vs mid, inside the range. |
| Adjusted EBITDA | $1,148 | $1,275 | $1,315 | $1,295 | +12.8% | FMP EBITDA $1,205 million is below the low. Different metric. Do not score it. |
| Adjusted EBITDA margin | 50% | 51% | 51% | 51% | +100 bps | Point, not a range. Gives back Q2’s 53% fee print. |
| MRR growth, normalized CC | 8% | 9% | 11% | 10% | +200 bps | Call language, not a dollar series. 100 bps under the Q2 print of 11%. No FMP line. |
| Recurring capex | — | $70 | $90 | $80 | — | 3%–4% of revenue. Not the growth capex. |
| GAAP diluted EPS | $3.81 | Not guided | Not guided | Not guided | +7.8% | FMP $4.11 vs the year-ago GAAP figure. Not AFFO per share of $9.83. |
| AFFO per share | $9.83 | Not guided | Not guided | Not guided | — | Annual only. No FMP AFFO book. Do not diff $4.11 against $9.83. |
Guidance credibility is informative, not performative. They are not chronic sandbaggers, and they are not chronic missers of their own outlook. Revenue versus the prior-quarter range, seven settled quarters: three beats of the midpoint, one clear of the high (only Q2 2026), two misses beyond 1% (Q4 2025 inside a wide range near the low, Q1 2026 below the low). Average versus the midpoint is about −0.6%. The last four averaged about −1.0%. A mechanical “whisper” of midpoint minus 0.6% would sit near the low half of the Q3 range. That average is distorted by xScale timing and should not be traded as a number. The pattern to take in is simpler. Clean quarters have finished at or below the midpoint. The only recent clear of the high was the quarter the slipped fee arrived. They beat the opening annual guide because they raise through the year, then finish the last guide near the low end on revenue and EBITDA and on the midpoint for diluted AFFO per share. FY2025 revenue of $9,217 million was $9 million above the last-guide low. FY2025 adjusted EBITDA of $4,530 million was $1 million under the last-guide low. FY2025 diluted AFFO per share of $38.33 was on the midpoint.
The Q3 range was widened to $50 million, from $20 million on the year-ago quarter’s guide. That is the right response if they will not forecast fee timing more tightly. It also means an in-line quarter can still miss FMP. Do not fade them for sandbagging. Do not pay up for a habitual beat the data does not show.
What the year leaves for Q4, if Q3 prints at the midpoint and the year is not raised, is arithmetic, not a guide. First-half revenue was $2,444 million plus $2,625 million, or $5,069 million. The FY midpoint of $10.245 billion minus that half minus $2.550 billion leaves about $2.63 billion for Q4, about +8.5% versus Q4 2025 revenue of $2,420 million. The joint-midpoint path has Q3 at +10.1% and Q4 at +8.5%. 2026 does not require a second-half acceleration inside the year. The acceleration already happened: FY2025 revenue growth was about +5.4% on $9,217 million versus $8,748 million, and the FY2026 midpoint is +11.2%. FMP’s Q4 revenue of $2,638.5 million is about $12 million above that residual. The street is not modeling a Q4 breakout above the year that is already guided.
Equinix is a retail colocation and interconnection REIT. The stack above total revenue is recurring revenue (colocation, interconnection, managed infrastructure, other) and non-recurring revenue, which is where xScale fees land. The KPI management discusses is normalized constant-currency MRR, not GAAP gross margin. Gross margin and operating margin are not guided. Internal is n/a on every line. Adjusted EBITDA and AFFO are the steered results. GAAP diluted EPS is shown because FMP scores it. It is the wrong surprise metric. YoY is the same quarter a year earlier.
Eight reported quarters
| Metric | Q3 24 | Q4 24 | Q1 25 | Q2 25 | Q3 25 | Q4 25 | Q1 26 | Q2 26 |
|---|---|---|---|---|---|---|---|---|
| Total revenue ($M) | $2,201 | $2,261 | $2,225 | $2,256 | $2,316 | $2,420 | $2,444 | $2,625 |
| Revenue YoY | +6.8% | +7.1% | +4.6% | +4.5% | +5.2% | +7.0% | +9.8% | +16.4% |
| Recurring revenue ($M) | $2,059 | $2,091 | $2,087 | $2,143 | $2,215 | $2,294 | $2,331 | $2,377 |
| Recurring YoY | — | — | — | — | +7.6% | +9.7% | +11.7% | +10.9% |
| Non-recurring ($M) | $142 | $170 | $138 | $113 | $101 | $126 | $113 | $248 |
| MRR growth, normalized CC | 5% | 5% | 6% | 7% | 8% | 10% | 10% | 11% |
| Interconnection ($M) | $384 | $393 | $393 | $407 | $422 | $433 | $446 | $453 |
| Adj. EBITDA margin | 48% | 45% | 48% | 50% | 50% | 49% | 51% | 53% |
| Diluted AFFO/share | $9.05 | $7.92 | $9.67 | $9.91 | $9.83 | $8.91 | $10.79 | $11.78 |
| AFFO/share YoY | +10.5% | +8.5% | +9.1% | +7.5% | +8.6% | +12.5% | +11.6% | +18.9% |
| MRR churn | 2.0% | 2.5% | 2.4% | 2.6% | 2.3% | 2.2% | 1.7% | 1.8% |
| Gross bookings, ann. ($M) | — | — | — | $345 | $394 | $474 | $378 | $424 |
The +16.4% is the fee. Recurring growth slowed 80 bps from Q1 to Q2 while normalized MRR still accelerated from 10% to 11%. Margin at 53% is the high of this window and is not the run-rate. AFFO per share growth reaccelerated to +18.9% after a Q4 seasonal trough of $8.91. Churn improved 80 bps year over year and is partly deferred. Bookings of $424 million are the second-highest print in the window, under the Q4 2025 record of $474 million. Interconnection revenue of $453 million versus $407 million is +11.3%. The call’s “around 9%” was an oral characterization. Use the dollars.
Regional normalized MRR in Q2: Americas 13%, flat versus Q1’s 13% and +700 bps versus Q2 2025’s 6%. EMEA 7%, still the slow book, −100 bps versus Q2 2025’s 8%. Asia-Pacific 11%, +300 bps versus Q2 2025’s 8%. Americas has stopped accelerating. It has not rolled over. Q2 Americas revenue of $1,251 million was +24.6% versus $1,004 million with the fee in it. The call said ex-Hampton it was still low-teens. That was a Q2 statement, not a Q3 guide.
Current quarter — Q3 2026e
Not reported. Guide columns are the Q3 outlook stored on the Q2 period. A full-year range is not pasted in as if it were a quarterly range. Consensus is FMP only, and only where the definition matches. Internal is n/a.
| Q3 2026e | Guide low | Guide high | Guide mid | Consensus | Internal | What would move it |
|---|---|---|---|---|---|---|
| Total revenue | $2.525B | $2.575B | $2.550B | $2,563M | n/a | +0.5% vs mid. Bullish is above $2.575B without another fee. Bearish is under $2.525B. |
| MRR growth, nCC | 9% | 11% | 10% | n/a | n/a | Hold of 11% is the quality beat. 9% is inside the band and a guided decel. |
| Adjusted EBITDA | $1.275B | $1.315B | $1.295B | Not comparable | n/a | +12.8% vs Q3 2025 at the mid. FMP $1,205M is the wrong definition. |
| Adj. EBITDA margin | 51% | 51% | 51% | n/a | n/a | +100 bps YoY is the hurdle. 53% would mean the fee stayed. |
| MRR churn | ~2% | 2.5% | ~2% | n/a | n/a | Back half, not a Q3 point. Still better than Q3 2025’s 2.3% if it stays near 2%. |
| GAAP diluted EPS | Not guided | Not guided | Not guided | $4.11 | n/a | +7.8% vs $3.81. A Q4-style charge can miss this without touching AFFO. |
| AFFO / AFFO per share | Not guided | Not guided | Not guided | n/a | n/a | Year-ago $965M and $9.83. Do not hurdle Q3 at Q2’s +18.9%. |
Colocation, interconnection, the three regions, bookings, and cabinet billings are not guided for Q3. The checks, against the last print rather than a range: colocation $1,772 million was +11.8% in Q2; interconnection $453 million was +11.3%; EMEA revenue $845 million was +10.2%, with Frankfurt and Amsterdam still called constrained; Asia-Pacific $529 million was +9.1% after a year-ago decline. Cabinets billed were 307,600, +5.2% versus 292,500. Capacity is not the same thing as billed cabinets.
FQ+1 — Q4 2026e
Equinix guides one quarter ahead. Q4 revenue, margin, MRR, and AFFO are not guided today. They will be set on October 28. The only Q4-specific operating comment is capacity: cabinet deliveries double in the second half, and more than 7,000 cabinets move from 2027 into Q4. That is not a revenue guide. The $2.63 billion residual above is not in the Guide columns.
| Q4 2026e | Guide | FMP | Internal | Year-ago anchor |
|---|---|---|---|---|
| Total revenue | Not guided | $2,638.5M | n/a | +9.0% vs $2,420M. No company mid to diff. |
| MRR growth | Not a Q4 range. Year is “around 10%.” | n/a | n/a | Q4 2025 was 10%. |
| Adjusted EBITDA | Not guided. FY margin stays 51%, not a Q4 point. | Not comparable | n/a | Q4 2025 $1,186M at 49%. |
| GAAP diluted EPS | Not guided | $3.97 | n/a | +47% vs $2.69. That base was the depressed quarter. Not an inflection. |
| AFFO per share | Not guided | n/a | n/a | Q4 2025 $8.91. Q4 is the seasonal AFFO trough. |
| Churn | Back half near the low end of 2–2.5% | n/a | n/a | Q4 2025 was 2.2%. |
If Q3 revenue comes in at the high end and the year is not raised, the Q4 residual falls toward $2.60 billion, about +7.5% versus Q4 2025. That mechanical tension is what the new Q4 guide has to resolve. FMP is not positioned for a third revenue raise.
FY2026 standing guide, and the only FY2027 framework
This is the guide going into October 28. It was raised on July 29 and has not been updated. There is no FY2027 dollar range. The next-year framework is the 2027–2029 outlook at the bottom. FY2026 FMP revenue is no longer under the range. After two raises, the street low is about the company midpoint and the street high ($10,337 million) is about $52 million above the company high. That is a guide the street has caught. It is not a guide the street still has to walk up to, and it is not a guide above the entire street.
| FY2026 guide | FY2025 | Low | High | Mid | YoY | Read |
|---|---|---|---|---|---|---|
| Revenue | $9,217 | $10,205 | $10,285 | $10,245 | +11.2% | FMP $10,277M is +0.3% vs mid. Spoken growth 11–12% normalized CC. |
| MRR growth | 8% | — | — | ~10% | +200 bps | Call: around 10%, the end of the original 8–10% band. Not a new range. |
| Adjusted EBITDA | $4,530 | $5,210 | $5,270 | $5,240 | +15.7% | FMP EBITDA $4,954M sits below the entire range. Wrong definition. |
| Adj. EBITDA margin | 49% | 51% | 51% | 51% | +200 bps | Held at every 2026 vintage. The 2029 line is the one they lifted. |
| AFFO | $3,761 | $4,240 | $4,300 | $4,270 | +13.5% | No FMP AFFO. Spoken 13–14% as reported, 12–13% normalized CC. |
| AFFO per share | $38.33 | $42.69 | $43.29 | $42.99 | +12.2% | As-reported math is the top of the 10–12% normalized band. FMP $17.16 is GAAP. |
| Capex, ex acquisitions and xScale | — | $5.0B | $6.0B | $5.5B | Step-up | Opening plan was $3.655–$4.155B. Largest raise on the release. Not a P&L beat. |
| Dividend per share | — | — | — | $20.64 | Unchanged | Same dollar at the Q4, Q1, and Q2 guides. Payout about 50%. |
| 2027–2029 outlook | Low | High | Where 2027 starts | Street |
|---|---|---|---|---|
| Revenue growth per year | 10% | 13% | Low end, then higher by 2029. Prior outlook was 7–10%. | FY2027 FMP revenue +10.7% vs FY2026e. The bottom of the band. |
| Adj. EBITDA margin by 2029 | — | — | 53% or higher. +200 bps vs the FY2026 point, over three years. | No FMP margin on this definition. |
| AFFO per share growth | 9% | 12% | Follows revenue, with noise from non-recurring fees. Starts low. | FY2027 FMP EPS $18.92 is +10.3% and is not AFFO per share. |
| Annual capex | $5,000M | $7,000M | Above the FY2026 capex midpoint. Prior plan was about $3–4B. | No street capex line used here. |
The 2026 P&L guides extend a recurring trajectory that is already in the prints and assume the Q2 fee does not repeat. The 2027–2029 growth rate and the capex band are aggressive relative to the prior 7–10% and roughly $4 billion plan. Management’s own words say that aggression does not have to show up on Wednesday.
The guide that is still in force was set on July 29 and has not been revised. Tone is the most confident it has been since the initial 2026 outlook in February: a second consecutive raise, described on the call as the largest single guidance raise in the company’s history, plus a reset of the 2027–2029 framework. The confidence is about demand, pricing, bookings, and the ability to build. It is not a promise that Q3 will look like Q2.
Keith Taylor presented the February outlook. Olivier Leonetti, on April 29, said he was nearly two months in. The Q1 miss versus the printed range was against a guide the prior CFO set, and Leonetti explained it as xScale timing. The Q2 beat and the long-term raise are his. Two quarters is not a track record. The raises under him have been delivered as language. The capital has not yet had to earn the 25% cash-on-cash he is still quoting.
| Topic | Q4 2025 call · Feb 11 | Q1 2026 call · Apr 29 | Q2 2026 call · Jul 29 |
|---|---|---|---|
| Who owned the numbers | Keith Taylor, then CFO. | Leonetti, about two months in. | Leonetti, with the long-term outlook in his voice. |
| Opening posture | Plan confidence “has grown.” Initial guide. FX and churn in the footnotes. | “Pleased,” then “we take nothing for granted.” | “Exceptionally strong.” Largest guidance raise in company history. |
| Revenue / MRR | Normalized revenue 9–10%. MRR 8–10%. | Raised $21 million. Growth to 10–11%. Hampton still not in the quarter. | +$100 million before FX, growth to 11–12%. MRR only to about 10%. |
| AFFO per share | 8–10% normalized. | 9–11%. | 10–12% normalized. 11–13% as reported. |
| xScale | Hampton assumed in the Q4 guide, then slipped. Timing, not demand. | Slipped again. About $80 million of revenue and $0.65 of AFFO per share moved to Q2. Full year unchanged. | Closed. About $120 million of non-recurring fees on 134 MW. The slip is over. |
| Capex | $3.655–$4.155 billion. | Top of that range, about $4.1 billion. Not a new plan. | $5–6 billion in 2026 and $5–7 billion a year through 2029. |
| Churn | Q4 at 2.2%. Full-year average 2.4%, inside 2–2.5%. | 1.7%, helped by delayed churn. Year tracking the low end of 2–2.5%. | 1.8%, same caveat. Second half near the low end of 2–2.5%. |
| What they leaned on | Bookings, pre-sales, powered land. | Cost discipline and a robust xScale pipeline. Fee timing was the public risk. | Firm price, record backlog, inference on top of enterprise, and a balance sheet they will lever by about a turn. |
The direction of travel is a guidance raise, not a wider set of excuses under an unchanged headline. February and April were “the plan is intact, Hampton is a timing item, capex stays inside the original envelope.” July kept the margin percent and the dividend and changed the growth and capital envelopes. The new hesitation is execution and financing, not whether the demand is there. That is a better problem than April’s, and it is a real one. The 2029 P&L does not automatically get the capex.
Fox-Martin tied the July raise to bookings, presales, and execution, not to one deal. Leonetti’s bridge before FX was revenue +$100 million, adjusted EBITDA +$62 million, AFFO about +$50 million, and AFFO per share +$0.46. FX then took $49 million, $27 million, $18 million, and $0.18 back off. The published ranges are the net. He also said a significant portion of planned capacity additions for the rest of 2026 is already committed through bookings and presales. The Q2 deck says about 30% of 2026 retail expansion cabinets are already pre-sold. That is visibility, not a blank check. The full-year revenue range is only $80 million wide, about 0.8% of the midpoint. The Q3 band is $50 million wide, and it steps down from Q2 for a reason they already named.
Assumptions under the guide, from the call and the release, not from a model. Constant currency, with the FX haircut already taken. Non-recurring revenue stays lumpy, and the moderate long-term assumption is about 5% of revenue. The cabinet acceleration is a Q4 and 2027 story. Pricing is firm, and the mark-to-market is multi-year, so Q3 does not need a price step-up. It needs price not to reverse. Demand is described as broad, with AI as an accelerant rather than a single-customer book. Projects are announced after power and permit gates, and they said they can slow deployment if realized demand misses the forecast. The balance sheet stays investment grade while leverage rises about one turn, funded with retained cash and debt, not a new equity program. Quarter-end net leverage was 3.6x on the company’s definition, with liquidity about $7.7 billion. Ratings cited on the deck were Baa1 / BBB+ / BBB+. No verified rating action after the capex raise was found. atNorth is not a consolidated-revenue assumption.
Since July 29, the ranges did not move. Conference appearances on August 10, August 11, September 14, September 15, and September 29 did not issue a new number. The one post-guide transcript with the CFO on the record is RBC on September 29 (document 27204756). He said the team had undersold the company, that demand is strong enough that they are turning some of it away to keep the ecosystem mixed, and that within three years he expects agentic AI to be about half of AI demand and 30–40% of data-center activity. On a hyperscale capex slowdown, he said he sees no slowdown yet, that it would show up in training before inference, and that the P&L can delay capex. He separated xScale from the core: retail colocation is the focus, and xScale is getting crowded with lower economics. He did not pre-announce Q3. He did make the capex caveat explicit. The three-point long-term raise is demand. The extra capex is a drag into 2029. Capacity language was about 1 gigawatt deployed by the end of 2029, with about 2 gigawatts of land and power still available after that. That is the physical version of the $5–7 billion band, not a Q3 revenue tell.
The July guide survives Q3 if four things are intact and one is merely on schedule. Intact: full-year revenue still 11–12% and AFFO per share still 10–12% normalized, with no new FX haircut beyond the $49 million already in the range. The 51% margin still has no low-end bias. MRR is still described around 10% for the year, which can coexist with another quarter at 11% without forcing a raise, and cannot coexist with a slide toward the old 8% floor without forcing a cut. Churn is still headed for the low end of 2–2.5%. On schedule, not delivered: the second-half cabinet doubling. A revenue print inside the $2.525–$2.575 billion band at a margin near 51% would confirm that the xScale fee did not have to repeat. A miss of that band, or a push of those cabinets back out of Q4, is the execution caveat they attached to the capex raise.
Q3 is being set up as an in-range quarter, not as a repeat of the Q2 non-recurring spike. FMP revenue sits inside the lowered post-xScale guide. The live debate is whether 11% MRR and record interconnection adds survive once the Hampton fee drops out. The $5–6 billion capex plan is a 2026–2029 delivery test, and most of the cabinet pull-forward was aimed at Q4. Thresholds below are the company’s range or the last print. There is no internal model, and there is no street line for MRR, churn, bookings, or interconnection adds.
| Catalyst | What is expected | Bullish / bearish | Source |
|---|---|---|---|
| Revenue after the fee | Guide $2.525–$2.575B. FMP $2.563B, +10.7% vs Q3 2025. A step down from the Q2 range. Consensus is not asking for another $2.6B print. | Above $2.575B without a new fee is the exception. Under $2.525B breaks a range that, outside the Q1 slip, has held. | Daloopa guide ids 174908927 and 174908928. FMP 2026-10-26. |
| MRR holds double digits | 9–11%, midpoint 10%. Last print 11%. Recurring dollars were +10.9% in Q2 and had already slowed from +11.7% in Q1. | 11% says the fee was the only one-off. Under 10% breaks the run of double-digit quarters even if revenue is in range. | July 29 call. Last print 11%. |
| Margin gives the fee back | About 51%. Midpoint EBITDA $1.295B is +12.8% vs Q3 2025. Q2 was 53%, and about half of the +300 bps was fees. | Another 53% means the fee did not drop. Under 51% spends a full-year point that has not moved all year. | 51% vs Q3 2025 50%. Q2 53% is the fee print, not the comp. |
| Interconnection adds | Q2 stocks sum to 522,700. The change versus Q1 is 9,700, matching the call. Q3 2025 flow was about 7,100. Revenue grew +11.3%, ahead of the “around 9%” oral line. | A hold near 9,700 says the ecosystem is still widening. A fall back toward 7,100 says Q2 was the spike. No street add number. | Regional stocks 237,400, 170,900, 114,400. |
| Bookings vs a harder comp | Q2 was $424M, +23% vs a soft $345M. The Q3 comp is $394M. On July 29, more than 45% of the Q3 bookings target was already closed. That comment is now stale. | Holding $424M is only about +8% YoY. A break under the $394M comp would say the pre-sale cushion did not convert. | $424M vs $394M. |
| Churn ticks toward 2% | Q2 was 1.8%, and management said delayed churn helped. Back half near the low end of 2–2.5%. | 2.0% is worse than Q2 and still better than Q3 2025’s 2.3%. Through 2.5% is outside the sentence. Staying at 1.7–1.8% is better than they telegraphed. | 1.8% vs 2.3%. |
| Cabinets are a Q4 test | Capacity was 395,500, +5.2% YoY. Deliveries double in the second half. More than 7,000 cabinets pulled into Q4. Istanbul IL3 is the named Q3 opening. | Missing the Q4 pull-forward is the execution caveat. A few hundred extra cabinets in September would not prove the 2029 plan. | 395,500. Pull-forward is the call. |
| Capex and the cost of capital | FY capex $5–6B. Through 2029, $5–7B a year, 10–13% revenue growth, about 53% margin, 9–12% AFFO per share. Leverage 3.6x, about one more turn, cost of capital up about 150 bps, still investment grade. | Already in the guide. Not a Q3 beat item. It becomes the argument if bookings or pre-sales soften. Q3 cash flow cannot prove a 2029 return. | $5.0B–$6.0B. RBC, Sep 29. |
| atNorth, first quarter as a stake | Closed Sep 2. About 34% for $895 million. Not consolidated. Called immediately accretive to AFFO per share, with no number and no range change. | Look for it in AFFO, not in the $2.563B revenue line. Do not treat the close as a revenue beat. | Company close release, Sep 2. Not a Daloopa revenue series. |
| Peers and hyperscale demand | Keppel DC already printed, supportive and not the same lease. AMT Tuesday morning. DLR Thursday night. Meta has reported a large 2026 capex plan. MSFT and GOOGL report the same day. AMZN the next day. | The backdrop supports the demand sentence. It does not book September Equinix revenue. A tower beat is not an MRR beat. | Company IR dates. Meta financing release, Oct 26. Not used as an Equinix estimate. |
Utilization is not tight, which is why the capex argument is about future deliveries rather than a full building today. Americas cabinet utilization was 81%, unchanged year over year. EMEA was 78%, +100 bps year over year. Asia-Pacific was 73%, −100 bps year over year. Price per cabinet is doing more of the work than the cross-connect count: North America MRR per cabinet $2,743 was +6.7% versus $2,570. Management said pricing was firm even as footprints got larger. A Q3 in which price reverses would matter more than a few hundred cabinets.
AFFO per share should look much slower than Q2 without being a miss. First-half diluted AFFO per share is $10.79 plus $11.78, or $22.57. The low end of the full-year guide, $42.69, leaves $20.12 for the second half. That does not require another Hampton, and it does not require Q3 to grow 19%. Q4 is the seasonal trough. There is no quarterly AFFO consensus to beat.
The window is July 30 through October 26, after the Q2 report and before this print. The July 29 release, the guidance raise, the call, and the regular dividend are excluded. Nothing in the window is a revised guide, a new bookings figure, or a disclosed Q3 billing update. The pattern is funding and positioning. Equinix raised $3 billion of notes, closed an $895 million minority stake in atNorth, and signed take-or-pay terms so a Georgia cooperative’s members do not carry the Hampton grid upgrade. 2027 AI products and open PJM and Texas power fights do not change unreported Q3.
| Date | What happened | Why it does not change Q3 | Source |
|---|---|---|---|
| Jul 29 | Cape Town tribunal approved two data centres, about 170 MW, over activist objections. No site plan yet. An appeal was still open. | Permit step in a new market. Not capacity that can bill in Q3. | Reuters, same morning as the earnings release. Not part of the release. |
| Aug 6 | $3.0 billion of senior notes settled, 2029 through 2036. Coupons 5.000% to 5.800%. Ratings on the term sheet were the existing Baa1 / BBB+ / BBB+, stable. Use of proceeds is general, including development and refinancing. | Funds the build. Does not create Q3 revenue. The coupon stack is the cost of keeping that build going at the current rating. Against Q2 revenue of $2,625 million, $3 billion is a bit more than one quarter of sales in new unsecured debt. | Free writing prospectus, Jul 30. August 6 8-K. |
| Aug 6 | Hampton, Georgia: 20-year take-or-pay with Central Georgia EMC. Equinix pays the grid upgrade, including unforeseen costs. | Right side of the ratepayer fight, and a long-dated cost if the hall is late. No megawatts and no rent-commencement date. No Q3 revenue. | Equinix press release, Aug 6. |
| Aug 11 | Equinix endorsed Governor Abbott’s Texas data-center standards. Nine Texas IBX sites. A later September halt on permits is a separate state action. Equinix was not named, and the company did not comment in the sources checked. | Texas development risk is higher than it was on the Q2 call. It is not a disclosed slip in a named project. | Equinix release, Aug 11. Inside Climate News, Oct 19, and Houston Chronicle, Oct 20, on the later halt. |
| Sep 2 | atNorth closed. Equinix about 34% and $895 million. CPP Investments about 51%. Partners Group about 10%. The platform keeps operating independently. Eight live Nordic sites. | Not consolidated IBX revenue. Accretion to AFFO per share was not quantified and the July ranges were not updated. Smaller than the $3 billion raised. | Equinix / CPP close release. |
| Sep 2 | Horizon: Fabric One with AWS and Google Cloud, beta later in 2026 and general availability planned for 2027. Inference Exchange with NVIDIA and Together AI, available starting Q1 2027. | The product answer to a wholesale-power narrative. A customer who waits for 2027 is not in the September quarter. | Equinix, Sep 2. |
| Sep 10 | DD Dasgupta named chief marketing officer, reporting to the chief customer and revenue officer, not the CEO. | Not a CEO or CFO change. Leonetti was still the CFO on the Sep 29 schedule. Does not change who signs the Q3 numbers. | Equinix press release, Sep 10. |
| Sep 29 | RBC conference. Leonetti did not walk the ranges up or down. Tone stepped up. The capex-versus-growth split was made explicit. | Confirms July. Does not raise the Q3 bar. See Set Up. | Daloopa document 27204756. |
| Oct 20 | Equinix moved to intervene in FERC docket EL26-108, the PJM reliability-backstop paper hearing. It owns both colocation sites and larger limited-customer data centers inside PJM. | A seat on who pays for backstop capacity. No megawatts and no dollars in the intervention summary. Not a Q3 revenue item. | S&P Capital IQ, Oct 21. FERC order of Sep 29. |
| Oct 21 | EPIC liquid-cooling labs opened inside live IBX halls in Frankfurt, London, Osaka, Paris, and Sydney. More sites expected in 2027. | Can pull forward a deployment decision. Does not add colonized capacity in Q3. No price and no named dollar commitment. | Equinix, Oct 21. |
Left out on purpose. Sell-side initiations and target changes are sentiment, not a change in the guide, and are not used. Smaller brand items with no disclosed revenue — a cloud hub, a Bogotá expansion, a foundation gift, a quantum hosting note — are not in the table. The October 1 media alert only sets the October 28 call. Conference webcasts that did not revise a number are in Set Up, not here as news.
What matters for October 28 is what the news does not do. No announcement since July 29 adds another disclosed one-time fee or a new bookings print. Cash has been raised at 5.0% to 5.8% and partly spoken for by a minority stake that stays off the consolidated IBX count. Hampton shows management will sign a 20-year take-or-pay to keep a project politically viable. Texas permitting is tighter than the August 11 release admits, and PJM is still deciding who pays for capacity built for large loads. Fabric One, Inference Exchange, and EPIC are 2027 products. The Q3 test is still MRR, the fee, and the 51% margin.
Mixed. On the line FMP scores, Equinix is not a consistent beater. Revenue missed FMP in 8 of the last 12 reported quarters (33% beat rate) and in 3 of the last 4. GAAP diluted EPS beat in 8 of 12 (67%) and in 2 of the last 4. Both lines beat together in 4 of 12 and 1 of the last 4. Revenue-surprise magnitude has deteriorated. The four-quarter average went from about −0.1% to −0.2% to −0.9%. The GAAP EPS beat has compressed to about a dime outside two Q4 air-pockets. The operating trajectory is the opposite of that tape. Diluted AFFO per share, which management calls its lighthouse metric and which FMP does not score, accelerated from +8.6% to +18.9% year over year over the last four quarters. Q3 2026 has not printed and is not on the grid.
A beat is any positive surprise and a miss is any negative surprise. There is no zero in this window. Surprise is the Daloopa actual minus the FMP estimate, divided by the FMP estimate. FMP stored basic EPS in Q1–Q3 2025, one or two cents above diluted. Using diluted does not flip the sign in any quarter. Two of the revenue “beats” are economically in line: Q2 2024 was +$0.5 million, and Q1 2025 was +$1.3 million. They are still counted as beats. Six of the eight revenue misses were inside 0.6%. The two that were not are Q4 2025 (−1.6%, −$38 million) and Q1 2026 (−2.8%, −$71 million). Q2 2026 then beat by +1.4% (+$35 million). Those three quarters are one Hampton fee, counted as it slipped and then as it landed.
GAAP is the scored EPS line, and it is the wrong line for a REIT. Q4 2024 diluted EPS of $(0.14) against diluted AFFO per share of $7.92 is the cleanest illustration. Scoring AFFO against the FMP GAAP estimate would manufacture absurd beats in every quarter. That comparison is not used.
| Vs FMP | Q3 24 | Q4 24 | Q1 25 | Q2 25 | Q3 25 | Q4 25 | Q1 26 | Q2 26 |
|---|---|---|---|---|---|---|---|---|
| Revenue ($M) | $2,201 | $2,261 | $2,225 | $2,256 | $2,316 | $2,420 | $2,444 | $2,625 |
| Revenue surprise | −0.1% | −0.5% | +0.1% | −0.1% | −0.4% | −1.6% | −2.8% | +1.4% |
| GAAP diluted EPS | $3.10 | $(0.14) | $3.50 | $3.75 | $3.81 | $2.69 | $4.20 | $4.83 |
| EPS surprise | +5.4% | −105% | +2.6% | +7.8% | +6.1% | −27.5% | −2.3% | +2.1% |
| Diluted AFFO/share | $9.05 | $7.92 | $9.67 | $9.91 | $9.83 | $8.91 | $10.79 | $11.78 |
| AFFO/share YoY | +10.5% | +8.5% | +9.1% | +7.5% | +8.6% | +12.5% | +11.6% | +18.9% |
Heatmap — last 12 reported quarters
Quarters run left to right, oldest to newest. One scale is used for both rows, capped at 8 points, so a revenue miss inside 1% stays pale and a GAAP EPS miss larger than 8% is the same dark red whether it is −8% or −105%. Green is a beat. Red is a miss. Depth is magnitude. Brand green is #1E8449 and brand red is #C0392B at the 5-to-8-point band. Darker cells are beyond that cap. Q3 2026 is not on the grid. AFFO is not a row, because there is no consensus to color.
| Metric | Q3 23 | Q4 23 | Q1 24 | Q2 24 | Q3 24 | Q4 24 | Q1 25 | Q2 25 | Q3 25 | Q4 25 | Q1 26 | Q2 26 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Revenue vs FMP | +0.2% | −0.1% | −0.6% | +0.0% | −0.1% | −0.5% | +0.1% | −0.1% | −0.4% | −1.6% | −2.8% | +1.4% |
| GAAP diluted EPS vs FMP | +36.3% | +11.1% | −8.3% | +18.8% | +5.4% | −105% | +2.6% | +7.8% | +6.1% | −27.5% | −2.3% | +2.1% |
Surprise versus FMP. Shared scale: under 1 point pale, 1 to under 2 light, 2 to under 5 medium, 5 to under 8 brand green #1E8449 or brand red #C0392B, 8 points and over the darkest green or red. The −105% cell is capped. It is not eight times an −8% miss in operating terms. Q2 2024 revenue prints as +0.0% and is still a +$0.5 million beat. Not a stock-price reaction. Q3 2026 has not printed.
Read left to right. The revenue row is pale for two years, then two darker red cells as the Hampton fee slips, then one green cell when it lands. The EPS row is the GAAP tape: dark green when a small revenue print still produced a large earnings beat in 2023 and early 2024, dark red in both Q4s, and pale again in 2026. The last cell on each row is a small beat. That is not a company that has started missing the operating plan. It is a company whose FMP scorecard is GAAP EPS plus a revenue estimate that periodically includes a fee the company has not yet signed.
Against the next-quarter revenue range they issued, the count in the transcribed window is different from the FMP count. Q4 2024 was a dollar under the low before a $22 million FX bridge Taylor described. Q1 2026 was under the as-issued range only because about $80 million of Hampton slipped. Q2 2026 was $14 million above a high end that already included those economics. Through 2025 they described revenue, adjusted EBITDA, and AFFO as in line or better while FMP revenue was a scratch or a small miss. The FMP miss rate and the company-guide record are different facts. The second one is the one they manage.
Into October 28. FMP revenue of $2,563 million sits inside $2.525 billion to $2.575 billion, about $13 million above the midpoint. That is the same seat FMP occupied in several of the pale-red quarters. A print at the midpoint is a small miss versus FMP and a make versus the guide. A print at the low end is about a −1.5% FMP miss, the size of Q4 2025, without needing another slipped fee. A print at the high end is a small beat. FMP EPS of $4.11 is about +8% versus Q3 2025 diluted EPS of $3.81. Do not compare that figure with Q2’s $4.83. Q2 is a different quarter, and it included the fee. The base rate from the heatmap is a small revenue miss or an in-line print versus FMP, and a GAAP EPS result within a few percent of the estimate, unless a one-time item hits diluted EPS the way Q4 did. The base rate from the company is a print inside the revenue range, with AFFO the line they will call a beat or a raise. A Q3 that merely hits the $42.69–$43.29 full-year framework will not look like the +36% GAAP cells on this grid, and it does not need to.