Lucknow 01 is Sify’s northern edge facility, inside the HCL IT City campus on Sultanpur Road, and it is described by Indian trade press and state government communications as India’s first AI-ready edge data centre. Ground was broken on 23 February 2025 in the presence of the Uttar Pradesh Chief Secretary, with four phases planned and phase one targeted for June 2025. It uses edge caching — storing data close to users to cut response time — and publishes up to 4.8 MW of modular IT capacity. The number matters for a second reason: Sify’s three Uttar Pradesh facilities now total 45.76 MW, or 13.8% of the state’s 331.06 MW counted capacity, making Uttar Pradesh Sify’s second-largest state after Maharashtra.
1At a glance
2Four phases, one live
The delivery structure is the most concrete thing published about this site. Four phases, with the first targeted for June 2025 and the groundbreaking in February 2025. That is a four-month build to first phase — which is what a modular pod is designed to do, and it is the entire commercial argument for building at the edge instead of waiting three years for a campus.
| Item | Published | If each phase is one 4.8 MW pod |
|---|---|---|
| Phases planned | 4 | 4 |
| Capacity per phase | not stated | 4.8 MW assumed |
| Site ceiling | up to 4.8 MW on the card | 19.2 MW — contradicts the card |
| Phase one status | targeted June 2025 | Not confirmed in public |
| Investment | not stated | Roughly ₹27 crore at the replacement band |
There is a real tension in the record. The card says up to 4.8 MW for the whole facility; the announcement says four phases. Either the four phases together total 4.8 MW — about 1.2 MW a phase — or the 4.8 MW is the phase-one figure and the site grows to 19.2 MW.
Nothing published resolves it, and this page does not resolve it either. But it is a 15 MW question, and it is the kind of question that only gets answered by asking an operator directly. Uttar Pradesh planners, competing campuses and any buyer sizing the eastern NCR market all need that number.
Reading one: 4.8 MW is the site ceiling, split across four phases of about 1.2 MW. A cautious reading, and the one consistent with the published card.
Reading two: 4.8 MW is phase one, and the site is planned for 19.2 MW. Consistent with how Sify describes modular scalability and with a four-phase build.
The registry counts 4.8 MW — the conservative reading. If reading two is correct, Uttar Pradesh’s counted capacity is understated by 14.4 MW on this one site, and Sify’s stated edge plan of 10 to 12 sites is closer to 100 MW than to 57.6.
3What the first AI-ready edge site means
“First” claims deserve scrutiny and this one is worth taking seriously, because it is specific and it is dated. The claim is not that Lucknow has AI capacity. It is that this is the first edge facility in India built to an AI-inferencing specification — edge caching, modular pods, low-latency design — rather than a general-purpose colocation room in a tier-two city.
| Operator | Where | What is published |
|---|---|---|
| Sify | Lucknow and Chandigarh | Two live, up to 4.8 MW each, 10 to 12 planned |
| Sify | Visakhapatnam | AI Edge facility on 3.6 acres plus an open cable landing station, about ₹1,500 crore over two phases |
| Sify | Nagpur and Kolkata 02 | Announced, no capacity stated |
| Ctrls, NTT, STT GDC | Tier-one metros | No edge programme published at this scale |
The Visakhapatnam entry is the one to watch, and it is a different order of commitment from a 4.8 MW pod. Sify has taken 3.6 acres from the state government, committed about ₹1,500 crore over two phases, and is building an open cable landing station alongside the edge facility. That is a coastal landing point plus a data centre — the same combination that makes Chennai 02 strategic.
So the edge programme runs on two tracks. Modular 4.8 MW pods in tier-two cities, and anchored landing-station builds on the coast. Lucknow is the template for the first; Visakhapatnam will be the template for the second, and it has not started counting yet.
For a buyer the distinction is simple. Lucknow solves a latency problem. Visakhapatnam will solve a connectivity problem. The second is worth considerably more per megawatt.
4Sify in a 331 MW state
| Facility | MW counted | Role |
|---|---|---|
| Noida 02 | 30 | Hyperscale AI campus, northern India’s largest |
| Noida 01 | 10.96 | Flagship enterprise campus |
| Lucknow 01 | 4.8 | AI Edge, eastern UP and the NCR belt |
| Sify in the state | 45.76 | 13.8% of 331.06 MW |
| Uttar Pradesh total | 331.06 | 24 records, additive only |
Uttar Pradesh is the fourth-largest state in this registry at 331.06 MW, and Sify is one of its two largest operators. The three facilities are deliberately different products: a hyperscale AI campus at Noida 02, an enterprise campus next door at Noida 01, and an inference pod 500 kilometres east in Lucknow.
That geographic split is the strategy in one sentence. Noida serves the Delhi-NCR market, which is where the demand and the power are. Lucknow serves the state government, the eastern UP enterprise base and anyone who needs sub-10-millisecond response in the Awadh region — and it does so at 1.4% of the state’s capacity.
The investment signal sits with the state rather than the facility. Sify’s chairman has said publicly that the company plans to double its investment in Uttar Pradesh over three years, and has discussed AI cities in Lucknow and Noida with the Chief Minister, with a large hyperscale AI campus proposed near the Lucknow edge site. None of that is a record yet. It is the pipeline behind this 4.8 MW pod.
5The public-sector play
The groundbreaking was presided over by the Chief Secretary of Uttar Pradesh, not by a corporate guest. That is a signal about who the anchor tenant is expected to be.
Edge facilities in tier-two cities are rarely built for hyperscalers. They are built for state government departments, public-sector banks, regional enterprises and e-governance platforms — workloads that must stay in-state, must respond fast and cannot justify a 100 MW campus. Data-sovereignty rules and state IT procurement preferences make a certified in-state facility close to mandatory for some of that demand.
Which reframes the 4.8 MW figure. The measure of success for Lucknow 01 is not megawatts, it is contracted state and banking workloads. Those tenants pay far more per kilowatt than a hyperscale wholesale customer, and they sign longer. A 4.8 MW pod fully let to government and BFSI will out-earn 4.8 MW of wholesale capacity in Navi Mumbai by a wide margin.
No contract, tenant or rate is published for this facility. That absence is the single biggest gap in the record — because at this scale the tenancy is the whole business case, and every one of these sites stands or falls on whether a state department signs.
6Procurement
7Limits and open questions
- 4.8 MW is an “up to” figure identical to Chandigarh 01, and no installed capacity is published.
- The relationship between four phases and 4.8 MW is not stated. The site could be 4.8 MW or 19.2 MW.
- Whether phase one opened in June 2025 as targeted is not confirmed in any public source.
- No investment figure, tenant, rack count, PUE or floor area is published for this facility.
- The “India’s first AI-ready edge data centre” claim is a press and state communication, not an independent certification.
- The proposed Lucknow hyperscale campus discussed with the Chief Minister has no capacity, site or date and is not a record.
- What would close it. A phase schedule with megawatts, an installed-load figure and a named tenant.
8Sources
Sources: the Sify AI Edge data centre page; the Sify data centre portfolio page; Sify Infinit Spaces portfolio page; the Sify Infinit Spaces release of February 2025 on the Lucknow AI-Hub data centre; ETGovernment and Communications Today on the 23 February 2025 groundbreaking and the first AI-ready edge claim; DataCenterDynamics and W Media on the four-phase build and HCL IT City location; Times of India on the Uttar Pradesh investment plan; TimesTech on the Visakhapatnam edge facility and cable landing station; and RCR Tech on the edge programme.
