Who could realistically lease capacity at 626–680 Lynnway and 0 Circle Ave, Lynn, MA: a radius screen combining external research with site information from the owner, organised by distance band and by fit against the Revenue Model's five tenant archetypes. Every claim is marked confirmed (sourced) or inferred (not yet evidenced); inferred items are a pipeline to validate, not contracted demand.
Six findings from site research, ordered by distance, then conclusions on the recommended action, offtaker potential and build size.
Findings
Cable adjacency (0.7km / 1.0km driving). The EXA Lynn Cable Landing Station, terminus of the Amitié transatlantic cable (400Tbps, commissioned 2023), is 2 minutes away. The defensible near-term product is a carrier-neutral meet-me room and network PoPs; there is no evidence that Meta or Microsoft would lease compute here directly.
GE Aerospace River Works (0.9km / 2.3km driving). A DoD-recognised jet engine manufacturing facility. Best fit is secure enterprise colocation, backup/DR and moderate inference (predictive maintenance, quality analytics), not a 250kW/rack training tenant.
Lynn Regional WWTP (~0.2km). Effectively co-located with the site and worth raising with the plant operator as a reclaimed or non-potable water source for cooling makeup. An adjacency to investigate; no agreement exists.
Local acute-care gap. Union Hospital (Lynn) closed September 2020; acute care now redirects to Salem Hospital (9.0km). A hospital is expected to return to the area, with no confirmed timeline, so the demand signal is real but its timing is uncertain.
Media & broadcast. No North Shore film soundstage is confirmed within the screen radius. Media stays opportunistic (post-production, broadcast workflows, content distribution), not a base-case revenue block.
Endicott College AMTI (~12.1km). A state-designated TechHub for AI, autonomy and robotics with a confirmed strategy-development grant: the strongest specific AI/robotics demand signal in range.
Conclusions
Recommended action: lease to an edge/micro-compute operator, not to local demand sources directly. GE, the hospitals and AMTI are end-customers, not data center operators. The realistic Lynnway Park counterparty is an edge operator such as Proxim8, Vapor IO or DataBank that takes the head lease and aggregates local demand itself. The underwriting question is whether an edge operator will find enough GE-and-hospital-scale demand nearby to justify anchoring here, and the findings above address it directly.
Offtaker potential is real and diversified. Within 12.1km sit four independent demand mechanisms: a transatlantic cable landing station, a large DoD-recognised industrial employer, a regional healthcare system in transition and a state-backed AI/robotics TechHub. None fills 6–12MW alone, so the underwriting case is a staged, multi-tenant, enterprise- and inference-led mix.
6–12MW is a realistic build size. The demand base maps onto 4–80kW/rack tenant profiles (cable-adjacent network tenants, GE-adjacent enterprise and edge workloads, regional healthcare colocation, AMTI-linked inference capacity), not a single 250kW/rack training anchor. 6–12MW matches both the surrounding demand and what the site's existing industrial footprint can accommodate.
Distances run from 626 Lynnway Rd, Lynn MA 01905 (42.453064, -70.961096): haversine straight-line figures, cross-checked against Google Maps driving routes where captured. They are screening distances, not surveyed parcel-centroid measurements.
Lynnway Park is an existing 252,000 sq ft industrial property across three buildings and 17 acres on Lynn's waterfront (recently sold for $40M), currently in mixed warehouse, food processing, cold storage and light-industrial use. The 6MW default / 12MW cap sizing reflects that: a smaller-format facility built within or alongside the existing industrial footprint, not a hyperscale greenfield campus. Any build timeline depends on how existing building use is transitioned, alongside the power, permitting and flood-mitigation items tracked in the cost assumptions. This page covers who the offtake demand is, not construction sequencing.
The tightest radius holds the strongest evidence: two confirmed, named, physically adjacent facilities and an on-parcel utility.
| Prospect | Straight-line | Driving | Best-fit archetype |
|---|---|---|---|
| EXA Lynn Cable Landing Station, 91 Commercial St | 0.7 km | 1.0 km / 2 min | Mail/web 4kW; carrier PoP; enterprise 8kW |
| GE Aerospace River Works, 1000 Western Ave | 0.9 km | 2.3 km / 5 min | Enterprise/healthcare 8kW; inference 80kW |
| Lynn Regional Wastewater Treatment Plant, 2 Circle Ave | ~0.2 km | n/a | Cooling-water partner, not a tenant |
EXA & Amitié. The Lynn Cable Landing Station backhauls EXA Atlantic, EXA Express, and Amitié. Amitié lands in Lynn, the UK, and France; 16 fibre pairs, 400Tbps total capacity, the highest-capacity transatlantic system deployed to date; commissioned and handed over to the Aqua Comms/Meta/Microsoft/Vodafone consortium in mid-2023. Meta (via its Edge USA subsidiary) is the landing party and holds majority ownership of the system, and separately leases colocation space in Lynn for the landing station's own power-feed equipment. That is a small-footprint precedent for cable-adjacent colo demand in Lynn today. Confirmed: proximity and network infrastructure. Inferred: a new Lynnway facility could attract EXA, carriers, cloud on-ramps or CDN nodes seeking diverse transatlantic connectivity; a landing station does not mean the cable owners will place GPUs next door. The base case is mail/web hosting and enterprise connectivity, with inference as a targeted growth product where customers need local processing.
GE Aerospace. The Lynn facility designs, produces, assembles, and tests military and commercial aircraft engines and components, spans 100+ acres with ~2,500 employees, and is a recognised DoD facility. GE's Predix industrial-cloud platform splits workloads between edge (real-time control, low latency) and cloud (bulk analytics), and a jet-engine manufacturing floor generating continuous sensor and quality data fits the local edge-compute profile. That supports a narrow, high-reliability inference/edge use case. It is not evidence that GE outsources its core classified or proprietary compute, nor a 250kW/rack training case. Sell compliance, physical security, data sovereignty and private fibre rather than generic AI capacity.
Water. The wastewater plant serves Lynn, Saugus, Swampscott, and Nahant, with 25.8 MGD dry-weather design flow and 110.8+ MGD wet-weather capacity. Because it sits on effectively the same Circle Ave parcel group as the site, it is worth raising reclaimed or non-potable water for cooling makeup directly with Lynn Water & Sewer, which could lower the water-related CapEx and OpEx lines in the cost model. An opportunity to evaluate, not a confirmed input: no agreement or capacity commitment exists.
The nearest confirmed healthcare facility is an outpatient centre; acute care is expected to return to the area on an unconfirmed timeline.
The Mass General Brigham Healthcare Center in Lynn (480 Lynnfield St) provides outpatient primary care, imaging, labs and specialties. It opened on the former Union Hospital site after Union Hospital closed in September 2020, with acute care reorganised toward Salem Hospital in the interim. A hospital is expected to return to the Lynn area, with no confirmed timeline, so the near-term prospect is the outpatient centre's parent health system, not a new local acute-care build: a credible 8kW/rack enterprise/healthcare colocation prospect (EHR, imaging/PACS, backup, DR, secure connectivity) through Mass General Brigham, not a stand-alone 6–12MW anchor. Sell a health-system service bundle (HIPAA-aligned controls, private connectivity, immutable backup, recovery testing) and revisit local acute-care capacity once a hospital timeline firms up.
The regional healthcare anchor and the most concrete robotics/AI signal both sit in this band; media does not.
| Prospect | Straight-line | Driving | Fit |
|---|---|---|---|
| Salem Hospital, 81 Highland Ave | 9.0 km | 9.2 km / 15 min | 8kW healthcare; backup & imaging |
| Beverly Hospital, 85 Herrick St | ~12.9 km | n/a | 8kW healthcare; regional DR |
| Endicott College / AMTI, 376 Hale St | ~12.1 km | n/a | AI inference 80kW; enterprise 8kW; testbed |
Healthcare. Salem Hospital is described as the North Shore's largest healthcare provider and is the current regional acute-care anchor while Lynn's own hospital capacity is in transition (see section 3). Beverly Hospital is a second regional health-system lead. Neither source proves a procurement need; both belong in a qualified pipeline sized around DR, imaging and backup workloads rather than a full colocation commitment.
AMTI (Advanced Mobility Technologies Institute). Led by MassAutonomy with MassDOT Aeronautics and housed at Endicott College in Beverly, AMTI is a designated Advanced Mobility TechHub (confirmed $99,990 strategy-development grant) covering AI, autonomy, robotics, microelectronics, aviation and uncrewed systems. It supports a testbed and research-cloud proposition (sensor data, simulation, computer vision, digital twins) on 8kW enterprise and 80kW inference racks. AMTI's roadmap is scheduled to inform infrastructure investment decisions by late 2026, so near-term demand should be marked prospective. 250kW/rack training is an option only if that roadmap produces a funded regional compute programme.
Media. No confirmed North Shore soundstage was found within this screen. The Ipswich "High Street Studios" lead is a dance/cultural-events centre, not film. Another Age Productions is in Newton (not North Shore); New England Studios is in Devens (outside this screen). A film-studio prospect raised in early discussions is unconfirmed. Media and broadcast stays opportunistic, not a base-case revenue block.
This band holds the competitive ceiling (established colo incumbents) and the clearest named robotics demand.
Competition. CoreSite BO1 (70 Inner Belt Rd, Somerville, 17.2 km straight-line / 18.7 km · 22 min driving) has 273,000+ sq ft, dense cloud/peering access, and an established interconnection ecosystem. Markley Group (One Summer St, Boston) is a 920,000 sq ft carrier hotel serving finance, healthcare, academia, government, entertainment, science, and technology. Expedient (1 Cabot Rd, Medford) is a smaller nearby competitor (38,000 sq ft, 25,775 sq ft raised floor, three 2.2MW generators, private 10Gbps interconnection). Lynnway cannot match this ecosystem depth on day one, so it has to differentiate on physical power, transatlantic cable adjacency, campus scale and a second-site/DR role rather than generic colocation density.
Robotics. Symbotic (200 Research Dr, Wilmington), Amazon Robotics (300 Riverpark Dr, North Reading), and Teradyne (600 Riverpark Dr, North Reading) are confirmed, named companies in warehouse automation, collaborative and autonomous mobile robots, and automated test equipment respectively. They establish local advanced-manufacturing and robotics demand, not that these companies will colocate production systems externally. The realistic fit is model inference, simulation, telemetry storage, CI/CD, demo environments and DR: 80kW inference or 8kW enterprise racks, with training as a speculative upside only.
Confirmed facilities and missions exist; a procurement need is not evidenced.
The City of Lynn IT Department provides telecom infrastructure and IT support to city agencies; MassCyberCenter provides cybersecurity guidance to municipalities generally. Essex County Sheriff's HQ (20 Manning Ave, Middleton) and MA State Police Danvers Barracks (485 Maple St, Danvers; Troop A HQ) are confirmed public-safety facilities in the wider region. Plausible workloads: backup, records, dispatch-adjacent systems, video retention, DR, continuity of operations and CJIS-sensitive data. Best fit is 4–8kW/rack enterprise/compliance colocation with documented controls, though public agencies may favour state-contracted centralised infrastructure over a private site. No procurement evidence exists for any of these: a category to explore, not a lead to underwrite.
The local demand sources in sections 2–6 are almost certainly end-customers of an operator, not direct lease counterparties. None of GE Aerospace, the hospital systems or Endicott/AMTI has disclosed a plan to lease and operate wholesale data center capacity. Their realistic route to Lynnway Park is as customers of a colocation or edge operator that takes the head lease, builds out the multi-tenant infrastructure and resells capacity in smaller increments.
The underwriting question is therefore whether an edge operator will find enough local enterprise, healthcare and industrial demand to justify anchoring here; sections 2–6 address it. The profiles below cover four real edge operators with this model and what Lynnway Park would earn as their landlord.
| Operator | Model | Typical site | Relevance to Lynnway Park |
|---|---|---|---|
| Proxim8 (Join Digital) | Micro Edge Data Centers (MEDC) in existing CBD office buildings, networked via Join's own fibre | 1,500–20,000 sqft, ~1MW/site | Named example; targets the “weeks not years” turnaround and existing-building reuse Lynnway offers |
| Vapor IO (Kinetic Grid) | Near-premises edge colocation + interconnection, multi-tenant, partners with wireless/cable networks | Micro data centers, multi-market rollout (100+ sites planned across major US metros) | Same "logical metro-wide data center via fiber-linked micro sites" structure as Proxim8; Crown Castle-backed, with tower and fibre access in most metros including Boston |
| DataBank | "Middle edge" positioning: national multi-market colocation, single- and multi-tenant options | Varies widely by site; e.g. 156,780 sqft Atlanta ATL4, adjacent utility substation | Established, credit-worthy national colo operator already active in the Boston metro market (see Market Rate Benchmark); a plausible anchor operator-tenant |
| EdgeConneX | Originally 1MW "lights-out" edge colocation in secondary markets; since EQT's 2024 majority stake, pivoting toward large single-tenant AI builds (23–30MW+) | Historical EDC: ~3,600 sqft, 1MW. Current: hyperscale build-to-suit | Historical model is the clearest edge/inference site-size comparable; the pivot shows even "edge" specialists consolidating into bigger single-tenant deals as AI inference demand scales |
Expected revenue. These operators need no new pricing: they are the AI/HPC Inference and Enterprise/Healthcare Colo archetypes already priced in Assumptions, section 1 (80kW/rack at $260/kW/mo, and 8kW/rack at $205/kW/mo respectively), both far closer to what an edge/micro-compute operator deploys than the 250kW/rack AI Training archetype. The Revenue Model's default blend (60% Inference / 30% Enterprise / 10% Media) is already an edge-operator revenue mix rather than a training-cluster one. The model's sourced rates applied to Lynnway Park's capacity range:
| Deal structure | Rate | At 6MW (start) | At 12MW (site cap) |
|---|---|---|---|
| Shell-Only (Lynnway as landlord to the operator) | $28.37/kW/mo (default rent) | ≈$2.04M/yr | ≈$4.09M/yr |
| Operator revenue (reference: if Lynnway ran the facility) | $237/kW/mo blended | ≈$17.1M/yr | ≈$34.1M/yr |
Every prospect above mapped against the Revenue Model's five tenant and workload scenarios.
| Archetype | Rack profile | Strongest local targets | Underwriting view |
|---|---|---|---|
| AI/HPC Training | 250kW/rack | AMTI, robotics cluster (long-run) | Upside only; no anchor proven |
| AI/HPC Inference | 80kW/rack | GE, robotics, AMTI, healthcare | Targeted growth product |
| Enterprise/healthcare colo | 8kW/rack | GE, Salem/Beverly/MGB, government, manufacturers | Most bankable initial segment |
| Media & broadcast | 30kW/rack | Regional producers, CDN/content customers | Opportunistic; local anchor unconfirmed |
| Mail/web hosting | 4kW/rack | Carriers, MSPs, public agencies, SMEs | Good fill / interconnection product |
Plausible categories outside the five scenarios, all inferred from Amitié's direct Europe connectivity, Boston's technology, research, finance and healthcare economy, and the North Shore robotics and mobility base: carrier and network PoPs, CDN/edge-cache operators, cloud on-ramps, financial-services connectivity customers and disaster-recovery operators. A carrier-neutral meet-me room is the most defensible of these: the landing station is a tangible network anchor, whereas financial-services and CDN demand already has an established home at Markley/CoreSite and would need a cost or power argument to move to Lynnway specifically.
How the site should be positioned and sequenced.
The prospects split into four demand groups: EXA/Amitié (network adjacency), GE (secure industrial demand), healthcare/government (compliance-driven continuity) and AMTI/robotics (future AI and simulation demand). On public evidence no single group supports the full 6–12MW on its own. Boston incumbents (CoreSite, Markley) are broader and farther from the landing station, so Lynnway has to compete on power, cable adjacency and a second-site role rather than ecosystem density. That fits the model's smaller-format 6MW/12MW-cap sizing rather than a large greenfield campus.
Recommended sequencing: (1) carrier-neutral meet-me room + 4–8kW/rack enterprise/DR product first; (2) 30–80kW/rack inference, robotics, industrial analytics, and healthcare modules second; (3) 250kW/rack AI/HPC halls only as an option, contingent on signed commitments and validated power/cooling design. This independently matches the Revenue Model's default blend (60% Inference / 30% Enterprise-Colo / 10% Media, Training deferred to the expansion case). The evidence supports a location thesis and a credible pipeline, not yet a fully committed 6–12MW demand case.
From section 8: Lynnway Park's lease is more likely signed by an edge/micro-compute operator such as Proxim8, Vapor IO or DataBank than by any local demand source, which are the operator's end-customers. The underwriting maths is unchanged, since the model already prices at the Enterprise-Colo/Inference rate such an operator would pay. What it changes: the leasing conversation should target operators first.
The full 50-reference list from the external research.