Pricing Strengthens the Growth Case, but Transaction Income and Funding Costs Keep Us at Hold

Results vs. Consensus

Equinix is demonstrating pricing power before the largest wave of new capacity opens. That strengthens the reinvestment case we carried into this quarter. It does not yet resolve our two reservations: whether interconnection can regain share of recurring revenue, and whether the earnings growth reaches shareholders quickly enough to justify the funding commitment. Q3 improves the first-year earnings outlook without removing either constraint.

Revenue finished $21M above the $2,550M company-guidance midpoint but $4M below the high end. Adjusted EBITDA exceeded the $1,295M midpoint by $26M. This is a better operating result than the tiny revenue surprise suggests, although it is not another revenue beat above the company range. Normalized monthly recurring revenue growth of 11% met the upper end of the prior 9–11% expectation.

Year-over-year comparison

Sequential comparison

Quality of the print

Revenue: Recurring revenue rose $65M sequentially while non-recurring revenue fell $119M as Q2’s xScale leasing fees rolled off. The resulting $54M decline in total revenue therefore masks expansion in the core franchise. Non-recurring revenue of $129M was 5.0% of sales, matching the steady-state framework introduced last quarter. That commitment has been met; the xScale risk remains contained.

Margins: Adjusted EBITDA margin was 51.4%, up about 180 basis points year over year and down from Q2’s fee-assisted 53.2%. Cash operating expenses rose just $4M year over year to $420M while revenue rose $255M. This is meaningful operating leverage, with cash G&A falling $9M. GAAP gross margin improved only 42 basis points to 51.1%; much of the profitability improvement therefore comes from expense discipline below gross profit, rather than a dramatic change in delivery economics.

Per-share earnings: AFFO grew 19.8% and diluted AFFO per share rose 18.4%, with a modest dilution drag. The atNorth closing generated non-recurring income below the operating line, helping lift GAAP EPS as well. The normalized, constant-currency AFFO-per-share increase excluding that benefit was 12%. We give more weight to that rate than the headline: it supports improved recurring earnings, but does not justify extrapolating $11.64 for four quarters. Net interest expense rose to $117M from $75M a year ago, a $42M increase that absorbed nearly a quarter of the $173M increase in adjusted EBITDA.

Segment Performance

Americas: fee normalization exposes a solid recurring engine

Revenue of $1,148M fell 8.2% sequentially as non-recurring revenue returned to $55M from $184M. Recurring revenue nevertheless increased to $1,093M from $1,067M, while normalized recurring growth was 12% year over year. Adjusted EBITDA of $560M compares with the fee-inflated $641M in Q2 and $489M a year ago. The 48.8% margin is therefore a cleaner indication of the regional operating run rate than last quarter’s 51.2%.

Assessment: The region remains a strong contributor to the capacity-expansion case. We view the sequential profit decline primarily through the known fee comparison, rather than as evidence of weaker enterprise demand. The important conversion test shifts to the cabinets scheduled for delivery in Q4.

EMEA: growth resumes, but incremental margin does not

Revenue rose to $878M from $845M, and recurring revenue rose to $831M from $806M. That extends the sequential recovery after the flat recurring result flagged in Q1. Normalized recurring growth of 9% year over year remained below the Americas and Asia-Pacific. Adjusted EBITDA was unchanged at $456M, taking margin to 51.9% from 54.0%.

Assessment: Revenue continuity is improving, but the region did not convert sequential growth into additional profit. We retain a constructive demand view while allowing for costs to absorb part of the near-term pricing benefit. Portfolio-wide operating leverage should not be assumed to arrive evenly across regions.

Asia-Pacific: profit recovers, margin remains a watch item

Revenue reached $545M and recurring revenue $518M, both higher sequentially. Adjusted EBITDA recovered to $305M from $299M, reversing the absolute profit decline we flagged last quarter. Its 56.0% margin still edged down from 56.5%, because revenue grew faster than profit. Normalized recurring growth was 12% year over year.

Assessment: The worst interpretation of Q2’s profit decline is weakened by the recovery, but the margin question remains open. The region continues to earn the highest margin in the portfolio; maintaining that advantage as capacity expands matters more than one quarter’s return to the prior dollar-profit level.

Service Mix and Key KPIs

Interconnection’s share of recurring revenue slipped to 19.04% from 19.06% in Q2 and 19.05% a year ago. The sequential change is small; the more important fact is that interconnection again failed to outgrow colocation. Managed infrastructure declined 8.5% year over year, consistent with the ongoing Metal wind-down being a drag on the broader mix.

Key Topics & Management Commentary

Overall Management Tone: Management remained confident but spent more time defending the durability of demand, the deliverability of capacity and its financing advantages. Compared with Q2’s expansive long-term reset, this call offered more concrete evidence on pricing and customer concentration, while acknowledging project delays and the possibility of deliberately higher churn.

1. Pricing now provides an observable growth mechanism

Pricing on new deals was up 12% year-over-year in the quarter.
— Adaire Fox-Martin, CEO and President

Last quarter’s open question was whether the higher revenue algorithm depended chiefly on more cabinets or higher revenue per cabinet. This quarter provides evidence for both: net cabinet billings increased by 5,700 and new-deal pricing increased 12%. New-contract pricing does not immediately reprice the installed base; it establishes a benchmark that can work through renewals over time. The opportunity is more attractive than volume alone because existing assets already carry much of their operating cost.

Assessment: Pricing strengthens our confidence in the earnings trajectory and supports a modest FY2027 estimate increase. We do not apply 12% to all recurring revenue: renewal dates, customer density and product mix determine how much is realized. The disclosed price signal closes part of the mechanism gap left in Q2 without supplying a complete price-volume bridge.

2. The mature interconnection moat holds; the portfolio test still fails

Our 193 stabilized assets are collectively 83% utilized and interconnection revenue in this pool increased 10% year-over-year, outpacing colocation revenue performance at a strong 7% growth year-over-year.
— Olivier Leonetti, CFO

The stabilized estate is important counterevidence to the bearish interpretation of two quarters of portfolio-level interconnection underperformance. Mature buildings still show interconnection growing faster than colocation. New capacity and the changing asset mix can therefore influence the consolidated comparison. The data do not establish that virtual connections are cannibalizing physical cross-connect economics, nor do they establish that the consolidated mix trend will reverse.

Assessment: We keep the flywheel pillar AT RISK rather than declaring it broken or restored. Our existing upgrade condition required two quarters of interconnection growth above colocation across the business. A favorable relationship in a narrower stabilized cohort cannot satisfy that condition. Fabric revenue growth of 24% is encouraging, but packaging and pricing of the new products remain unfinished.

3. Record bookings improve the visibility behind the build

Annualized gross bookings, are those bookings that occur within the quarter where we anticipate that we will be able to convert them to revenue within a 90-day window.
— Adaire Fox-Martin, CEO and President

The $533M bookings result grew 35% year over year and excludes atNorth. Management had already closed 60% of its Q4 target by the call, compared with more than 45% of the following-quarter target at the Q2 call. Those percentages indicate momentum against internal targets, not identical dollar hurdles. The current definition also suggests a shorter conversion horizon than the two-to-three-quarter characterization in our prior recap.

More than 20% of the planned Q4 capacity and 15% of the 2027 retail capacity were presold. These are different delivery cohorts; they are not evidence that the prior roughly 30% presale figure deteriorated. Total sales activity, including presales, grew 38%, but the quarter’s presales dollars were not separately specified on the call.

Assessment: Nearer-term bookings and contracted future capacity improve the case for delivering the revenue guide. They reduce demand uncertainty without eliminating it: most of next year’s capacity is not yet presold, and bookings are annualized contracted revenue, not revenue recognized in Q3.

4. New AI services extend the franchise before they prove monetization

The official launch is next year. We will be entering beta for both of these products over the course of this quarter, in fact, in the next few days, and with a number of customers for both of these. And then I think once we've come out of that process and gone into GA, then we will have defined pricing and packaging, and we'll be able to share what that might look like going forward in terms of impact on our longer-term guide.
— Adaire Fox-Martin, CEO and President

Fabric One aims to automate connectivity across cloud, network and AI environments, with AWS and Google Cloud as lead integration partners. Inference Exchange brings together compute, inference services and Equinix’s space, power and connectivity. Both enter beta before a launch next year; Fabric One’s prepared-remarks timetable points to early 2027. This broadens the potential use of the neutral network without requiring Equinix to own the entire computing stack.

Assessment: These services reinforce the distributed-inference pillar, but we assign no separate new-product revenue increment in the base forecast. Pricing and packaging are prerequisites to a defensible contribution estimate. Liquid-cooling bookings growing more than tenfold signal customer interest but lack an absolute base, so they do not substitute for that earnings bridge.

5. Capacity delivery advances without another capex raise

We expect more than 13,000 cabinets in Q4, our highest quarter of delivery on record and more than 20% of that Q4 capacity has been presold.
— Adaire Fox-Martin, CEO and President

Full-year retail deliveries are now expected to exceed the original plan by nearly 40%, while the $5–6B capex range remains unchanged. The 48 projects across 32 markets create options to move equipment and construction effort when one site slows. Management acknowledged a delayed Dubai project and some power delays, an important qualification to the confident delivery narrative.

Assessment: More delivery within the existing spending envelope is preferable to another capital reset. It supports the 2027 growth entry point without activating our prior downgrade trigger of annual 2027 capex above $7B. Moving projects around is nevertheless different from being able to cancel aggregate spending if demand weakens; that downside flexibility remains unquantified.

6. The funding burden is arriving before the full return

We have 4.1x net leverage, $9.1 billion of available liquidity, a 3.2 percentage blended in-place borrowing rate primarily fixed with an average maturity over 6 years, access to lower cost capital globally to fund growth.
— Olivier Leonetti, CFO

Nine-month operating cash flow of $3,027M less $4,392M of purchases of other PP&E leaves a $1,365M deficit, versus $108M a year earlier. Cash dividends of $1,547M widen the gap to $2,912M before real estate acquisitions and net equity investments. Reported free cash flow of negative $801M includes other investing flows and excludes net equity investment purchases; it is not the same measure of the organic funding requirement. The $869M atNorth investment adds a separate use of capital.

The balance sheet is liquid, but liquidity has been replenished with borrowing. Total debt principal rose to $25,217M from $22,153M in Q2, while cash and short-term investments rose to $3,597M from $2,224M. Leverage increased by half a turn. The 3.2% after-tax borrowing rate remains low, but it is a historical blended rate, not the marginal funding cost of the next project.

Assessment: The funding risk remains EMERGING, as anticipated last quarter, rather than becoming an unexpected solvency problem. The 27% stabilized return is cash gross profit relative to gross PP&E; it is not an equity IRR or a return after all corporate costs. Its comparison with borrowing cost is directionally encouraging, but subtracting the two does not produce an investable economic spread. The next stage of value creation depends on new assets reaching utilization while refinancing and spending remain controlled.

7. atNorth closes, with clearer ownership but an incomplete earnings run rate

The September 2 transaction resolves the closing uncertainty carried for two quarters. Equinix contributed $869M for approximately 34% of the common and preference shares. The preference shares carry a cumulative 12% annual dividend and rank ahead of common equity. The investment is not consolidated into revenue or bookings; common-equity earnings are recognized with a three-month reporting lag. Equinix also committed to lease a minimum amount of capacity by the end of 2030, with locations and terms still to be finalized.

This structure makes the transaction more than a simple addition of Nordic cabinets to the consolidated model. The release identifies a non-recurring closing fee within net income, while the full-year AFFO bridge adds $31M from atNorth. That full-year net increment is not a disclosed recurring quarterly run rate and should not be confused with the gross fee or with revenue.

Assessment: Closing and capital contribution are now observable, and the preference economics add useful protection. We remove the $0.31 per-share full-year atNorth increment from the base used to extrapolate FY2027 growth rather than capitalize transaction income indefinitely. Future preference returns, common-equity results, funding costs and capacity commitments need to be assessed together before giving the investment a separate valuation premium.

8. Low direct neocloud exposure limits one form of AI risk

And even if I take the totality of the neo presence at Equinix today, in total, that cohort as a class represents less than 2% of our MRR.
— Adaire Fox-Martin, CEO and President

Management disclosed 150 neocloud network nodes, including 31 added in the quarter, and described strict credit protections. Their role is partly to connect enterprise customers to specialized compute providers, rather than to fill giant campuses with a single tenant. No individual customer represents more than 2.5% of revenue.

Assessment: Less than 2% direct MRR exposure makes a single neocloud failure less threatening to the consolidated recurring base than the sector narrative might suggest. It does not insulate the ecosystem from a broader reduction in AI capital spending or provide a measure of exposures inside joint ventures. We retain the AI pillar ON TRACK while resisting the stronger claim that diversified customers remove cyclical risk.

Guidance & Outlook

The $100M revenue midpoint increase comprises $84M underlying and $16M currency. The $70M EBITDA increase comprises $59M underlying and $11M currency. AFFO rises $90M: $52M underlying, $31M atNorth and $7M currency. At the per-share level, $0.54 of the $0.92 increase is underlying, $0.31 atNorth and $0.07 currency. The recurring improvement is substantial, but it is smaller than the headline raise.

Implied ramp: Subtracting nine-month revenue of $7,640M from the $10,345M full-year midpoint gives $2,705M for Q4, within $1M of the explicit quarterly midpoint due to rounding. The EBITDA bridge is exact: $5,310M less $3,962M equals $1,348M. Full-year AFFO of $4,360M less nine-month AFFO of $3,389M implies $971M in Q4, or roughly $9.78 per share using the guided 99.3M diluted shares. Recurring capex rising $83M sequentially explains approximately $0.84 per share of the decline before other movements.

Guidance style: This is the third consecutive full-year raise, but the company still expects a Q4 margin trough. Management attributed the sequential decline to seasonal spending, not a quantified new-capacity drag. We accept the disclosed cadence rather than assume another beat. The unchanged 2027–2029 framework still calls for 10–13% annual revenue growth, 9–12% AFFO-per-share growth, 53%+ exit EBITDA margin and $5–7B annual capex, with growth beginning toward the lower end.

Analyst Q&A Highlights

Bookings exclude acquired growth

Q: But was this an outlier quarter? Or do you think the business has the potential to push north of $0.5 billion going forward as a sustainable pace? And then could you just remind us, too, if the atNorth portfolio was included in the 3Q figures or if that's excluded?
— David Guarino, Green Street

A: atNorth was not part of those statistics. This is not going to be a consolidated entity. We will use the equity method to recognize our share of ownership of atNorth, and you will see that in the earnings, other incomes. But no impact in the numbers we have just discussed.
— Olivier Leonetti, CFO

Assessment: The accounting answer excludes an acquisition explanation for the record bookings. Management also described strong pipeline and 60% of the next-quarter target closed, but did not commit to a permanent $500M quarterly bookings floor. Our forecast relies on conversion and pricing, not a new guaranteed bookings run rate.

Renewal pricing may deliberately raise churn

Q: So first, curious if you could share more detail as to what's been driving down your MRR churn on a quarterly basis? And when you think about the guidance that you gave for annual revenue growth over the long term, so the annual rate, I think you said it at 10% to 13%. I'm curious what churn assumption you thought about there because if it's the 2% to 2.5% that you historically used, so 2.25% at the midpoint, you've been averaging like 1.7% to 1.8% last 3 quarters. That's a 50 basis point difference from an annualized basis, 2%. I'm just trying to wonder what that means for your revenue growth opportunities over time.
— Michael Rollins, Citi

A: However, as we do look at our installed base and we look at the opportunity to mark-to-market in our installed base, it is possible that we may choose to do selective churns where we may choose to churn out with intention a customer that's operating at a particular price point and perhaps low interconnection or network density and churn in a customer that has some of the opposite parameters.
— Adaire Fox-Martin, CEO and President

Assessment: The response qualifies the benefit of 1.7% churn. Management credited renewal execution and constrained supply, while retaining a 2–2.5% guide. A higher future churn figure would need to be tested against replacement pricing, density and interconnection attachment; it would not automatically mean retention had deteriorated. We do not annualize the current retention advantage as permanent incremental growth.

The Q4 margin trough is attributed to seasonal spending

Q: So I think the guide calls for 49% to 50%. Next quarter, that's relative to 51.4%. Is the dip mainly tied to some of the new capacity that's coming online? Or are there other maybe unique factors to consider?
— Matthew Niknam, Truist

A: On your EBITDA margin, traditionally, you have a Q4 EBITDA margin declining relative to the prior quarter. This year will be no exception. So it's due to seasonality of spend.
— Olivier Leonetti, CFO

Assessment: This directly addresses the trough flagged in our prior recap. Seasonal expense is the stated explanation; management does not supply a dollar bridge separating that from capacity-opening costs. We therefore model the 49.9% quarterly midpoint and continue to regard the full-year roughly 51% result as compatible with operating progress.

Power is largely secured, with a contracting qualification

Q: If you can give us an idea of sort of how far out you've got the resources, labor, materials and so forth. And then what if any of those resources are most at risk, particularly with some of the pledges you made for power and so forth. How should we think about that?
— Frank Louthan, Raymond James

A: When we announced the project, so when you see it in our expansion tracker, you should be aware that at that point, we have cleared permitting hurdles and that we are largely secured in terms of power. If it's not contracted, it's very near to contract by the time that we put it visible to you in our expansion tracker.
— Adaire Fox-Martin, CEO and President

Assessment: The answer provides useful process evidence and goes on to describe equipment pre-orders, fungible designs and contractor relationships. It does not establish that all power is already under contract. That qualification preserves the open question from Q2: portfolio flexibility mitigates individual delays, but the proportion of future capacity backed by executed power agreements remains undisclosed.

What They’re NOT Saying

1. A comparable portfolio pricing bridge: New-deal pricing is now measurable, but renewal spreads, the share of contracts repriced and the contribution of density versus pure price are not. Those determine how much growth can be earned without proportionate new capital.
2. Interconnection product economics: The company provides growth and unit additions, but not enough product-level revenue or margin data to establish whether Fabric shifts dilute or enhance unit economics. The mature-asset comparison mitigates concern without resolving it.
3. Aggregate spending flexibility: Reallocating equipment among sites addresses timing risk. It does not specify how much of the multi-year capex plan can be deferred without cost if bookings slow.
4. Recurring atNorth accretion: Closing terms are clearer, but the sustainable earnings contribution after financing costs and the minimum capacity lease commitment are not quantified. The full-year net contribution is not a substitute for that run rate.
5. Presales dollars and contracted power coverage: Future-capacity percentages offer useful visibility, but no comparable quarterly presales-dollar series or percentage of power under executed contract was provided on the call.

Market Reaction

• Pre-print: Shares closed October 28 at $1,007.66, up 31.5% year to date versus 13.1% for the S&P 500, but down 0.3% over the preceding thirty days.
• October 29 reaction session: The stock opened at $1,038.74, traded between $1,023.00 and $1,068.33 and closed at $1,036.67, up 2.9%. Volume was 0.8M shares, 1.6 times the thirty-day average.
• Index comparison: The S&P 500 declined 0.3%, leaving roughly 3.2 percentage points of relative outperformance.

Our interpretation is that bookings, explicit pricing gains and the third guidance raise outweighed concerns about the transaction-assisted AFFO result. The stock had largely moved sideways over the preceding month, leaving room for improved operating expectations to be rewarded. The reaction supports a positive reading of execution; it does not show how much investors assigned to any individual driver.

The rally also leaves the valuation decision close to where it was after Q2: a high-quality growth franchise with a large funding commitment. We would not infer from a positive day that investors have resolved the cost-of-capital issue. That test requires several quarters of delivery and financing, not one trading session.

Street Perspective

Pricing power versus capital intensity

Bull view: Scarce retail capacity and shorter renewal cycles let Equinix reprice its installed base while filling new buildings.

Bear view: Rising leverage and multi-year development spending can absorb the improvement before it reaches distributable cash.

Our take: The operating evidence favors the bulls, but the funding objection remains material. Pricing is improving now; the cash return on the new build comes later. We raise earnings modestly without expanding the valuation multiple.

An AI platform or an AI credit risk

Bull view: A neutral network benefits when enterprises need access to multiple compute and model providers.

Bear view: Weakness among heavily financed AI providers can reduce ecosystem activity and delay investment even without large direct tenant exposure.

Our take: The less-than-2% neocloud MRR disclosure limits the direct concentration concern. We remain constructive on the platform while treating new-product revenue and broader AI sensitivity conservatively.

A clean beat or transaction-supported earnings

Bull view: The underlying AFFO guide increase and stronger bookings demonstrate recurring progress.

Bear view: The atNorth fee increases the headline earnings result, while the Q4 cadence is materially lower.

Our take: Both observations are true. The $0.54 underlying per-share guide increase supports a higher recurring base; the $0.31 atNorth increment does not deserve automatic capitalization. The useful debate is the normalized earnings path, not whether the headline beat happened.

Our Estimates and Valuation

Our FY2027 estimate begins with $43.91 less the $0.31 atNorth net guide increment, or $43.60. Applying 9% growth, the low end of the existing long-term framework, gives $47.52, rounded to $47.50. This deliberately avoids separately adding a full-year atNorth preference return before its net financing and accounting effects are clear. Renewal pricing, bookings conversion and cost discipline support that entry-point growth; higher debt costs and opening expenses argue against moving immediately to the upper end of the range.

At the $1,036.67 reaction close, EQIX trades at 23.6x FY2026 guided AFFO and 21.8x our FY2027 estimate. The former compares with 24.4x in the prior recap. Our retained 24x FY2027 base multiple rewards the mature interconnection franchise and 27% stabilized asset yield, but assumes no additional premium for product launches or the 2029 growth endpoint. The quarterly dividend of $5.16 provides $20.64 over four quarters at an unchanged rate, a 2.0% yield; dividend growth would be upside to this assumption.

The bull case assumes stronger renewal pricing and faster capacity conversion allow earnings above our base while interconnection regains portfolio-level momentum. The bear case assumes earnings stay near the current-year level as delivery or financing costs absorb growth, and the market compresses the multiple to 20x. That combination matters: a two-year capital commitment can lose value through both lower earnings and a higher required return.

We use a 10% twelve-month S&P 500 total-return assumption as an investment hurdle, not an observed forecast. The base case’s roughly two-point excess return is too narrow to justify Outperform against a credible 13% total-return downside and an unresolved interconnection mix test. Each 1x change in the multiple moves base value by $47.50, or 4.6% of the share price; that sensitivity exceeds the benefit of our $0.50 earnings upgrade.

Thesis Scorecard Post-Earnings

Overall: The operating thesis strengthens on pricing, demand conversion and mature-asset evidence, while the cash cost of growth becomes more visible. No pillar changes status. Q3 meets the recurring-growth commitment and the normalized non-recurring revenue framework; it does not meet the portfolio interconnection recovery condition. The atNorth closing resolves an uncertainty without establishing repeatable accretion.

Action: Maintain Hold. An upgrade would require a de-rating toward 21x FY2026 guided AFFO (23.6x today) with the earnings case intact, or two quarters of interconnection revenue outgrowing colocation accompanied by an attractive twelve-month return. Downgrade tests remain 2027 capex above the $7B ceiling without commensurate returns, borrowing costs outpacing the previously guided roughly 150-basis-point increase through 2029, or interconnection growth below 9% while unit additions set records. For Q4, the immediate tests are the $2,704M revenue midpoint, 49–50% EBITDA margin, delivery of more than 13,000 cabinets and conversion of strong bookings without an adverse shift in funding needs.