Summary

  • Equinix’s June 2024 CN1 plan described US$65 million of initial investment and 850 first-phase cabinets. Its September 2025 opening announcement described US$69 million and 800 cabinets, while the eventual site total moved from 4,950 to 4,250.
  • These are changes in company disclosures, not measurements of cost overruns, lost computing capacity or current availability. A purchasing decision needs a dated account of power, cooling, connectivity and capacity available for customer activation.

For a customer considering Equinix’s CN1 data centre in Chennai, the meaningful offer is not a cabinet count alone. It is space with specified power, cooling and network access, available on a date that fits the deployment. The company’s historical announcements provide investment and capacity descriptions. They do not supply that complete delivery test.

The distinction matters because three published numbers moved in different directions between the earlier plan and the opening announcement. Investment increased; the first-phase and eventual cabinet figures decreased. That is a concrete finding worth examining without turning it into an unsupported verdict on construction or operations.

Two disclosures, three numerical changes

In its 24 June 2024 announcement, Equinix described an initial US$65 million investment in CN1, 850 cabinets in the first phase and 4,950 cabinets when fully built. It expected the facility to open in the fourth quarter of 2024.

On 19 September 2025, Equinix announced the opening of CN1, its first International Business Exchange data centre in Chennai. That disclosure described an initial US$69 million investment, 800 first-phase cabinets and eventual support for 4,250 cabinets.

Published measure June 2024 plan September 2025 opening disclosure Calculated change from the earlier figure
Initial investment US$65 million US$69 million US$4 million higher; about 6.2%
First-phase cabinets 850 800 50 fewer; about 5.9%
Eventual full-build cabinets 4,950 4,250 700 fewer; about 14.1%

The changes above are arithmetic comparisons of the two disclosures. In particular, both full-build totals describe eventual scale. The difference of 700 cabinets is not evidence that 700 operating cabinets were removed from service.

This is a retrospective comparison of the 2024 and 2025 announcements, not a report of inventory available in September 2026. Neither historical headline should be used as a current availability quotation.

A higher investment figure is not a cost-overrun calculation

The investment figures are associated with the first phase, but the two announcements do not reconcile their underlying cost components. A common phase label does not establish an identical accounting perimeter, work scope or spending timetable. The earlier plan and opening disclosure therefore support a US$4 million increase in the published figure, not a like-for-like budget-overrun finding.

Dividing each investment number by its corresponding cabinet count would not resolve that problem. It would assume that the costs and the units being compared were equivalent. The disclosures do not establish those assumptions. The useful follow-up is a reconciliation of what each investment figure covered, rather than a more precise-looking ratio built on an uncertain denominator.

The same discipline applies to the calendar. An expected Q4 2024 opening and a September 2025 opening announcement warrant a schedule question. But an announcement date is not, by itself, the date of first customer service. These records do not establish equivalent service-commencement milestones from which to calculate an operational delay.

Cabinets do not measure workload capacity

A cabinet count describes one dimension of a facility’s capacity. To assess a deployment, a buyer also needs the power and cooling available to the proposed equipment, the required connectivity and the date on which that combination can be used.

Fewer cabinets could coexist with unchanged or greater workload capacity if the power and equipment density changed. Conversely, an unchanged cabinet count would not prove that the same workload could be supported. These are possible relationships, not explanations established for CN1. The two announcements do not provide a normalized comparison of commissioned IT load or computing output.

Nor does a first-phase total distinguish capacity already contracted, equipment already installed by customers and capacity immediately available to another buyer. Those categories answer different commercial questions. A cabinet can be physically ready without being available for a new order; an expansion ambition can be large without giving a particular customer a delivery entitlement.

The most useful unresolved question is therefore specific: at the September 2025 opening announcement, how many cabinets and how much sellable IT load were commissioned, powered, cooled, network-enabled and capable of customer activation—and how much was contracted, occupied or immediately available?

The absence of that reconciliation in these disclosures is not evidence of a failed commissioning process. It also does not establish that power delivery, cooling or network readiness constrained CN1. For a purchase now, the same questions should be answered as of the proposed delivery date, not by relying on either historical announcement.

CN1’s published figures changed. What those changes meant for customer-usable supply remains a separate evidential question. Keeping the two apart makes the comparison useful without pretending it is an operating audit.