Arelion’s AS1299: Tier-1 Access Needs Continuous Renewal

Summary

• Arelion is the private carrier that was once called Telia Carrier; Twelve99 remains a technical and historical identity, while AS1299 represents its global routing identity.
• AS1299 reaches the entire internet through customer routes and settlement-free peering, but Tier-1 status does not guarantee the shortest path, best support, or immunity from incidents.
• IP transit and dedicated internet access sit alongside Ethernet services, wavelengths, managed optical networks, cloud connectivity, peering exchanges, DDoS mitigation, mobile, voice, and AI Direct.
• Routes, fibre, points of presence, peers, and accumulated experience grant advantage, yet funding capacity, security, and expansion continuously remains a burden without published independent financial data.

Four names describe four different layers

The activity began within the Telia Group and operated for years as Telia International Carrier then Telia Carrier. Polhem Infra completed its acquisition of Telia Carrier on 1 June 2021, moving control outside Telia Company. The carrier adopted the Arelion name in January 2022. These dates separate stages of corporate history: referring to Telia ownership is correct before the deal was completed, and incorrect if used to describe the current company.

Twelve99 did not disappear with the rebranding. The name is still visible in the technical domain twelve99.net, including the public looking glass tool, and remains associated with AS1299. This continuity is useful for engineers because autonomous system numbers, hostnames, route filters, customer settings, and operational references often need stability longer than a brand’s lifespan. But it does not create a second company. Twelve99 represents technical continuity, while Arelion is the current commercial and corporate identity.

AS1299 is yet another layer. It is the autonomous system identity through which the backbone originates, receives, selects, and announces routes. An ASN is neither a legal entity nor a physical cable. It defines a routing scope whose policies are expressed through routers, PoPs, and interconnection points. Arelion operates this scope, while the underlying routes may use fibre under various ownership and contractual arrangements.

The four-part separation—former brand, current company, technical hostname, and routing system—prevents several common mistakes. It stops historical continuity from being turned into a false ownership claim, prevents treating a technical identifier as a subsidiary, and forces network-size numbers to be attributed to the company that publishes them. It also reveals the central theme in Arelion’s history: corporate control changed faster than the backbone’s operational identity.

A Tier-1 network is a relationship, not a medal

The phrase ‘Tier-1 backbone’ compresses several facts into one label. At the routing layer, a Tier-1 autonomous system can reach the entire internet via routes learned from customers and through settlement-free peering with other large networks. It does not need to buy transit from an upstream provider to obtain full global access. This position matters because it removes a class of supplier dependency and allows the network to sell full-table transit to others.

It is a maintained position, not a permanent grant: peers can change policies, traffic can shift, and the carrier must retain enough capacity and direct relationships to sustain payment-free exchange acceptable to both sides.

These definitions do not certify latency, support quality, DDoS protection performance, enterprise access, or financial strength. BGP selects paths by policy and attributes, not by a universal shortest-distance rule. A Tier-1 network can still depend on colocation providers, submarine systems, access carriers, router vendors, optical equipment suppliers, and customers whose traffic gives it economic weight. The label describes routing independence under a particular set of interconnection relationships. It should be the start of study, not the end.

Arelion’s current position illustrates this distinction strikingly. Its market claim is not merely that AS1299 has global reach, but that this reach can be translated into IP transit, packet private transport, optical services, cloud and exchange access, network-layer security, and operational support. Buyers therefore test Tier-1 status through a service contract whose value depends on hand-off point, geography, capacity, routing policy, and fault-repair capability.

From incumbent national carrier roots to international carrier

Arelion traces the organic growth of its backbone from 1993. The network grew within the Telia lineage, but its operating logic was international. A domestic incumbent serves retail, mobile, and enterprise customers inside a national market. An international carrier must connect other networks across borders, place routers in neutral facilities, obtain long-distance paths, maintain optical capacity, and negotiate interconnection with organisations that may also be competitors.

This difference helped make Telia Carrier separable from its former parent. By 2020 the carrier had customers and infrastructure relationships that extended far beyond Telia’s national business. Telia Company agreed to sell it to Polhem Infra in October 2020, and the acquisition closed the following June. Completion of the deal, not its announcement, defines the change of control. The Polhem Infra statement also noted a continuing strategic network relationship with Telia, showing that corporate independence did not cancel mutual commercial dependence.

The 2022 rebrand gave the independent carrier a name unlinked to the former parent. But the core asset was not a newly assembled network—it was a mature backbone whose fibre paths, PoPs, customer links, and peering relationships had accrued over decades. This matters because interconnection cannot be cloned simply by installing a software-defined control layer. A newcomer can lease capacity and activate ports quickly in specific markets, but it cannot instantly create the history of direct paths, operational trust, and fault-recovery practices embedded in a long-lived global network.

Independence also changed the investment question. Inside Telia, the carrier competed for capital within a diversified telecoms group. Under Polhem Infra it became a focused infrastructure asset backed by Swedish pension capital. That structure may favour long-term investment, but it does not remove return targets or make capital unlimited. It changes who decides, how the asset is framed, and what outsiders can see.

The backbone is a chain of control, not a uniform asset

Arelion says it has more than 80,000 km of fibre and more than 350 PoPs and delivers service in 129 countries. It also says it has more than 2,000 customers and about 450 access partners. These numbers describe different layers and should be kept attributed to the company. Fibre kilometres relate to physical reach; PoPs to interconnection and service locations; countries to commercial availability; access partners to extending delivery beyond direct facilities. Adding the numbers together does not yield a meaningful metric.

The physical chain may include fibre owned by Arelion, fibre controlled through long-term arrangements, or fibre whose capacity is leased; optical systems placed on that fibre; routing and switching hardware; space and power in colocation facilities; cross-connects to customers and peers; capacity in submarine systems; cloud on-ramps; and local access circuits provided by other carriers. Arelion can control the service and routing policy while sharing physical failure domains with suppliers.

‘Global backbone’ describes an operational system assembled across several forms of control, not proof that one company owns every duct, cable, and building.

This separation is operational, not merely linguistic. Two logical paths that look diverse on a network diagram may run through the same duct or the same submarine cable. Two services sold separately may share a router, an optical line system, or a facility’s power feed. Conversely, leased infrastructure can be highly resilient when the contract, monitoring, and physical diversity are well-designed. Ownership alone does not answer the reliability question; the important evidence is failure domain and the provider’s ability to detect and recover from it.

Arelion’s commercial platform therefore starts from control of a chain. It must know which piece belongs to its own backbone, which relies on a facility, which is delivered by an access partner, which end is under a cloud provider’s control, and where the SLA changes contractually. The customer sees a single order and a single support relationship, but a root-cause incident may cross several organisations. Part of the carrier’s value lies in making that chain manageable without pretending it is a single entity.

Fibre provides the path, optics provide usable capacity

Fibre length is a geographical metric, not a capacity number. A single strand can carry varying numbers of wavelengths, and each wavelength can be upgraded through coherent optics and new line systems. So the same physical path can carry far more traffic after an equipment refresh without digging a new route. Arelion’s ongoing programme of 400G and above capacities falls within this optical layer, where transponders, pluggable coherent optics, amplifiers, and spectrum engineering turn glass into efficient transport.

A wavelength service gives the customer a dedicated optical channel over a qualified path. It suits large predictable traffic such as data centre interconnect, replication, or carrier traffic aggregation. The customer gets a service with a clearer capacity ceiling than public IP transit, but the wavelength remains dependent on the physical fibre, the optical gear, and path protection. ‘Dedicated’ refers to the channel; it should not be expanded into a claim that the customer owns a physically isolated cable.

Arelion also sells dark fibre and managed optical fibre network services. The managed service transfers more design, equipment, and operational responsibility to the carrier. It may spare a large enterprise, cloud operator, or service provider from building an in-house optical engineering team for each route. In return, dependency increases on the geography Arelion supports, its supplier choices, its restoration procedures, and its lifecycle decisions.

The optical services also reveal why a carrier cannot treat an existing backbone as a completed asset. Traffic grows, interfaces move to higher speeds, and customers expect new paths to clouds and data centres. Each upgrade consumes capital before demand is certain. Under-building can lead to congestion and poor service; over-building can leave idle optics, ports, and contractual commitments. So the carrier’s financial discipline is hidden inside capacity planning long before a customer notices a speed change.

PoPs turn long-haul capacity into a marketplace

A point of presence is the location where backbone capacity becomes available for interconnection. It may house routers, optical systems, switches, customer ports, and cross-connects inside a data centre or carrier hotel. A dense PoP map can reduce the distance between Arelion and its customers, peers, clouds, and exchanges. But it also creates operational exposure to facility power, cooling, building access, and delivery procedures.

The total of more than 350 PoPs is a scale claim, not a statement that every product and every port speed is available in every location. An IP transit site may not support a particular wavelength path; 400 Gb/s Ethernet requires technical qualification; cloud connectivity depends on the provider’s on-ramp; a remote enterprise site may need a partner link. So the useful buying question is “What service is available at this hand-off point?” rather than “Does Arelion have a presence in this country?”.

PoP density also affects interconnection. Direct interconnection reduces the number of intermediate networks a path may cross, and more direct customers and peers can improve path selection. Arelion claims that 95% of end users in the United States and Europe can be reached within one network hop, and describes AS1299 as the world’s most connected backbone by its chosen metric. These statements should be kept attributed because the result depends on the dataset, definition, and date. One-hop reach does not equal lowest latency for every flow, and interconnection is not traffic market share.

The packet layer carries several products over shared infrastructure

Above the optical layer, routers and packet systems turn capacity into routable and private services. The BGP control plane on AS1299 exchanges internet routes. MPLS and segment-routing capabilities support traffic engineering and packet private transport. Ethernet provides Layer-2 connectivity; Smart IP-VPN provides a routed, managed WAN; Cloud Connect and IX Connect extend these paths into specific ecosystems. These products share parts of the underlying infrastructure while presenting different service boundaries.

Arelion’s Ethernet portfolio includes point-to-point EVPL and multi-point ELAN arrangements at speeds from 10 Mb/s to 400 Gb/s, subject to path availability and technical qualification. The carrier’s documentation describes segment-routing and Flex-Algo mechanisms for selected low-latency paths within the MPLS backbone. These controls can improve predictability inside the provider’s scope, but they cannot cancel an access link or an external end-point that sits beyond Arelion’s full control.

SLA numbers make this boundary visible. Arelion publishes a 99.999% availability claim for MPLS backbone services, 99.99% for basic Ethernet between PoPs, and 99.5% for an end-to-end service that includes a managed NID and an access link. The figures are not interchangeable. The lower commitment for the broader service reflects additional equipment and third-party dependencies. Actual terms remain contract- and path-specific.

A customer can also place multiple services on a single port through logical separation. That may reduce cross-connect and interface costs, but it concentrates services onto one physical failure domain. A port failure could affect several logically separate circuits at once. Effective multiplexing and concentrated risk must be explained together.

IP transit is the specific commercial expression of Tier-1 status

An IP transit customer typically runs an autonomous system, announces prefixes via BGP, and receives internet routes from Arelion. The carrier carries outbound traffic to other networks and returns traffic to the customer’s prefixes. The product is priced and contracted as access, capacity, and service, while the core value comes from AS1299’s customer and peer relationships.

The transaction looks simple because BGP hides the physical path behind route announcements. Operationally, however, Arelion must maintain full tables, filters for incorrect or unauthorised announcements, load-balancing across links, capacity provisioning, session protection, community management, and fault recovery across many PoPs. A customer using more than one carrier gains diversity and path choice but also acquires a harder routing problem. It must understand preferences, inbound traffic engineering, and how failures propagate across each provider.

Tier-1 economics do not mean every interconnection is free. Settlement-free peering eliminates payments for a specific exchange with qualifying networks. But Arelion still pays for fibre, facilities, equipment, power, access, labour, maintenance, and other commercial relationships. It may buy local services or access from a partner where that is more efficient than building directly. The absence of an upstream transit cost is one element of the cost structure, not a costless network.

The carrier’s position in the paths must also be preserved. If the traffic balance, geographic reach, or commercial policy changes materially, a peer may request different terms or end the relationship. Therefore the practical moat is cumulative and continuously defended. The network needs enough customers, direct reach, capacity, and operational credibility that other major networks keep exchanging traffic with it without settlement.

Dedicated internet access moves more routing responsibility to the provider

Dedicated internet access serves enterprises that want managed connectivity without running a full BGP relationship and global routing table. A customer may receive a default route or a managed edge, while Arelion takes on more responsibility for the internet-facing design. The physical access and service quality can resemble IP transit at some sites, but the operational contract is different.

This difference matters during faults and changes. A transit customer can use its own ASN, prefixes, communities, and multi-carrier policy. A DIA customer often has less control and relies more on the provider’s routing and third-party support. Neither model is inherently superior. Transit suits organisations with routing expertise and a need for policy control; DIA suits buyers who prefer a managed hand-off point.

SecureConnect bundles internet access or transit with automated DDoS protection. This bundling reduces procurement and configuration boundaries for customers who would have contracted transport and mitigation separately. But the protection remains scoped. It does not protect credentials, endpoints, application logic, or attacks outside the chosen network-layer service.

BGP communities make route policy part of the product

BGP communities are tags attached to route announcements. Arelion can interpret customer tags and apply provider-defined actions—changing local preference, restricting propagation, adding AS-path prepending, or triggering blackholing. They give the customer influence inside the provider’s network without direct router access.

This is a powerful form of delegated control. An operator can make a prefix less attractive in a region, limit where it is announced, or sacrifice reachability to protect the rest of the network during an attack. The meanings are AS1299-specific; a community value that performs one action at another carrier may carry a different meaning. Customers need current documentation, change control, and a way to verify the outcome through routing tools.

Communities also carry risk of error. A misplaced tag could withdraw reachability or send traffic down an unintended path. Remotely triggered blackholing intentionally drops traffic destined to a specific target at the upstream layer, so that an attack does not congest wider links. This is a useful emergency action because it trades one target’s availability for network stability. But it is not costless mitigation.

Looking glass tools reveal part of the routing reality

Arelion’s Twelve99 looking glass lets users inspect routes, pings, and traceroutes from selected network locations. It can answer practical questions: how AS1299 sees a particular prefix, which path is chosen from a specific observation point, and where latency appears from that perspective. It also preserves the Twelve99 name as a technical identity after the corporate rebranding.

A looking glass is a single viewpoint’s evidence, not an audit of the entire network. It does not show every alternative path, historical incident, internal preference, or physical fibre route. BGP policy can differ by location, and a traceroute may hide devices or respond differently under load. The tool is most valuable when used together with the customer’s own measurements, route collectors, service tickets, and provider performance data.

Arelion also publishes monthly IP network performance metrics. These reports help buyers track provider-chosen metrics over time, but the methodology and scope remain company-determined. Transparency improves when a provider discloses operational data; independent assurance still needs clear definitions and external monitoring.

RPKI reduces one routing risk without fixing BGP completely

Route Origin Authorisation lets a prefix owner specify which AS is allowed to originate the route. Route origin validation can classify an announcement as valid, invalid, or not found. Blocking invalid origins reduces some hijacks and misconfigurations. Arelion’s educational materials present RPKI as part of routing security.

This technique checks origin authority for covered prefixes. It does not verify the full AS path, does not guarantee that a peer exports routes correctly, and does not prevent every route leak. A valid origin can still propagate via an unintended path, and operational errors can still occur in filters, route objects, or customer setups. RPKI narrows the trust problem, but it does not replace BGP policy, monitoring, and incident response.

This limit matters for a Tier-1 carrier because path security and reachability can conflict. Strict filters may block legitimate traffic when records are wrong; loose filters may let invalid announcements propagate. An operator needs a staged policy, customer communication, exception handling, and current data. Saying “RPKI secures BGP” hides the operational judgement still required.

IX Connect sells access to interconnection, not full internet access

Internet exchange points provide places where networks can establish peering. IX Connect transports a customer from a qualifying Arelion site to a port at an exchange, enabling remote participation without building a separate local network presence. The product can lower the cost and time needed to reach multiple peers, especially for a network expanding into a new market.

But the transport does not create the peering policy. The customer still needs exchange membership where required, compatible ports, bilateral or route-server arrangements, and its own route filters. The remote path can also add extra dependency compared with placing a router physically at the exchange. Therefore IX Connect is an access service to a marketplace of paths, not a substitute for the customer’s own interconnection strategy.

This product illustrates why a Tier-1 carrier simultaneously competes with and cooperates with exchange points. Arelion peers at exchange points while selling transport to them. A customer may replace some paid transit with direct peering but still buy Arelion capacity to reach the exchange or to cover the rest of the internet. The boundaries among transit, peering, and transport are commercial decisions built on top of the same physical network.

Cloud Connect joins the backbone to a hand-off point controlled by the cloud provider

Arelion lists private connectivity with AWS, Microsoft Azure, Google Cloud, Oracle, and IBM. Cloud Connect carries customer traffic to supported cloud on-ramps, reducing reliance on public internet paths for that segment. This can provide more predictable routing, capacity, and security than sending all cloud traffic over public internet transit.

The service ends at a shared boundary. The cloud provider controls its virtual interface, region availability, quotas, and internal fabric. The customer controls the account, routes, and workloads. Arelion controls the carrier path it sells. A fault can lie on any side, and configurations must be coordinated. Describing the entire path as “private” can mislead if it suggests one provider owns every component or that there is no shared infrastructure.

Cloud connectivity also changes the carrier’s competitive position. Hyperscale providers operate extensive private backbones and carry more traffic between their own regions. Arelion’s opportunity lies in connecting enterprises, data centres, multiple clouds, and networks that lie outside a single administrative scope. Its limit is the same: it cannot dictate what happens inside the cloud nor replace the cloud provider’s own fabric.

DDoS defence became part of backbone capacity planning

Volumetric attacks consume links, router resources, and scrubbing capacity. A global carrier sees enough traffic to detect shifts and can divert attacked flows before congestion reaches a customer access circuit. Arelion’s DDoS service combines detection, route-based diversion, and scrubbing, then re-injects allowed traffic towards the destination.

In its report published on 15 July 2026, Arelion said the Aisuru botnet represented about one-third of observed attack traffic in its dataset, and that the largest attack reached 6.1 Tb/s. These are provider observations based on Arelion’s network and methodology. They are useful because they show the attack scale a large backbone sees, but they are not a comprehensive global census of DDoS activity.

The 6.1 Tb/s observation should not be turned into a guarantee of mitigation capacity. An attack may be observed across the backbone without a single customer receiving its full size, and service commitments depend on infrastructure and contract. Mitigation effectiveness also depends on detection time, BGP convergence, scrubbing locations, clean-path capacity, and false-positive control. Application-layer attacks may pass volumetric filters because the packets themselves look legitimate.

Remotely triggered blackholing represents the emergency design limit. A BGP community can cause traffic heading to an attacked prefix to be dropped upstream. That protects shared capacity but makes the target unavailable. Scrubbing aims to keep service alive; blackholing accepts an outage to contain damage. A mature DDoS product needs both mechanisms, clear triggers, and customer authority to decide when to use the most aggressive response.

SecureConnect bundles mitigation into the connectivity purchase rather than treating it as an optional add-on. That may improve adoption because protection is active before an incident. The strategic test is not the bundle name but the attach rate to services, response evidence, scope of protected services, and ability to scale as attack peaks rise.

AI Direct bundles wide-area transport for distributed AI systems

AI Direct is Arelion’s connectivity portfolio for moving data between AI clusters, data centres, and clouds. It groups familiar carrier services—Ethernet, wavelengths, internet access, managed optical networks, and security—under an AI infrastructure proposition. The portfolio does not provide GPU units, storage systems, or model-training software. Its role begins when data must leave a single facility or a single administrative domain.

This boundary matters economically. Training inside one campus relies on local high-performance fabrics whose latency and collective behaviour a wide-area carrier cannot replace. Wide-area demand appears when datasets, checkpoints, replicas, inference traffic, or entire workloads move between facilities. Arelion’s advantage is access between those locations; its service must connect to the internal AI fabric, not become that fabric.

The company added 400G EVPL to AI Direct in May 2026. The announcement proved a high-capacity packet option on qualified routes, not 400G availability everywhere. Arelion had also announced Danish network upgrades and a cable landing station environment in April, describing them as part of a northern AI corridor. Earlier additions included North American expansion such as an Oklahoma City PoP, while a 2026 Mexico channel programme extended distribution reach. These events signal an active path-and-product strategy, but the company did not disclose activated utilisation, customer concentration, or per-project investment value.

The strongest evidence that AI Direct is more than a label would be named customers with significant capacities, activated ports, recurring traffic, and path-specific service commitments. “AI superhighway” announcements express strategic intent but reveal little about how much revenue, usage, or new capacity is actually driven by AI workloads. Current evidence supports a plausible transport portfolio and an investment direction, but it does not yet support a claim that AI has already transformed the company’s economics.

Enterprise services earn backbone revenue beyond transit

Transit prices face long-term pressure as capacity improves and buyers have more alternatives. Arelion can respond by selling services that bundle the same routes with greater control, hand-off points, and support. Ethernet, Smart IP-VPN, Cloud Connect, IX Connect, managed optical networks, and DDoS protection each widen the set of problems the carrier can solve for a single customer.

This portfolio does not mean every service has the same margin or the same buyer. A wholesale carrier may buy transit and wavelengths; a multinational enterprise may buy managed access and IP-VPN; a cloud operator may buy optical capacity; a regional provider may buy transit and exchange access. The commercial organisation must sell shared infrastructure through different contracts and support models.

A broad portfolio also creates internal operational complexity. Capacity reserved for private services, public transit, and mitigation must be planned together. An optical path change can affect several packet products. Local access can delay an otherwise automated global order. Asset reuse improves economics only when service isolation, change management, and capacity accounting remain precise.

Arelion’s mobile data, IoT, voice, and messaging services widen the portfolio further. The research evidence is stronger for the backbone and connectivity products; the adjacent services should therefore be treated as part of the current catalogue rather than analysed with unsupported market-share claims. Their strategic significance is that Arelion is not purely a wholesale IP transit seller, even though AS1299 remains its core technical identity.

Access partners widen reach and change SLA boundaries

Arelion says it has about 450 access partners. These carriers can connect a customer building to an Arelion PoP where the backbone’s own fibre does not reach the premises directly. The model extends coverage without requiring Arelion to build every last mile. It also gives the customer a single commercial relationship for a path assembled from several providers.

The access link is often the least standardised part of the service. Delivery lead time, repair procedures, available bandwidth, jumbo-frame support, demarcation equipment, and local regulation all vary. Arelion may monitor and manage the circuit, but the physical repair may be another carrier’s responsibility. The lower published availability for end-to-end Ethernet versus PoP-to-PoP reflects this extra exposure.

Global service reach should therefore be read as delivery capability, not a map of owned facilities. Enterprise buyer due diligence includes knowing the local supplier, confirming path diversity, understanding escalation rights, and verifying whether a second access path shares the same duct. A single global contract can simplify governance while the underlying local physics stay the same.

Operations turn network assets into a service customers can use

A backbone has no value unless it can be provisioned, monitored, and repaired. Arelion’s customer model includes the MyArelion portal, service management, and network operations support. These systems translate ports, routes, tickets, planned maintenance, and performance into a customer experience.

Automation can shorten provisioning and reveal status, but a global carrier cannot reduce every event to an API call. Fibre repair needs field crews; a cloud endpoint may reject a configuration; a peer may change policy; a colocation facility may need a cross-connect; a local carrier may miss a delivery date. A carrier’s automation quality includes clarity about how exceptions and responsibilities are represented, not just the speed of processing a normal order.

Arelion says it has won customer-experience recognition and recorded a high Net Promoter Score. These claims can indicate an intended service culture, but the methodology and respondent sample have not been independently audited in the evidence provided. Awards and surveys belong alongside operational detail, not in place of it.

The business model reuses a single infrastructure base across multiple hand-off points

Arelion earns revenue by selling access to capacity, routes, service assurance, and operational responsibility. IP transit monetises global route access. Ethernet and IP-VPN monetise private packet paths. Wavelengths and managed optical networks monetise optical capacity and engineering. Cloud and exchange products monetise ecosystem access. DDoS services monetise visibility and protection. Each product starts at a different hand-off point even when several share the same backbone.

Reuse can improve asset utilisation. A fibre path and a PoP support more revenue when they serve multiple products and customer groups. Nevertheless the carrier needs separation and headroom. If every service is planned against the same optimistic utilisation level, an attack or a traffic shift can reveal hidden contention. Capacity accounting must distinguish reserved ports, normal traffic, protected traffic, and failure scenarios.

The commercial leverage comes from reducing the customer’s coordination work. A multinational buyer could separately assemble local carriers, exchange ports, cloud links, transit, optical paths, and security. Arelion offers to integrate a larger portion of this chain. The customer pays not just for bits, but for fewer contracts, a single support path, and the ability to shift responsibility. The provider earns this premium only when the integrated service works better than the piecemeal alternative.

Pension-backed ownership changes the time horizon, not the laws of economics

Arelion is owned by Polhem Infra. According to its current description, Sweden’s Third and Fourth AP Funds jointly own Polhem Infra. That links the carrier to national pension capital through an infrastructure investment vehicle. The structure suits an asset whose routes and interconnection position are built over a long arc.

Patient capital can absorb investments whose payback spans years—fibre rights, optical upgrades, new PoPs. It may favour steady cash generation over short-term public-market signals. But the owner remains subject to fiduciary duties and return requirements. The pension-fund link does not mean routes are funded regardless of demand, nor does it reveal how much debt, operating cash, or owner equity supports Arelion.

The governance change after 2021 was clearer than the technical change. AS1299 kept routing while board control, funding, and strategic priority moved from Telia to Polhem Infra. Daniel Kurgan became CEO in October 2023, and Charles Gill joined as CFO in March 2024. The current management page identifies a wider leadership team in strategy, commercial, legal, people, technology, and operations, but it is not a complete legal board register.

Not every architectural decision can be attributed personally to the executives. The backbone is the accumulated work of engineering, operations, procurement, sales, and partner teams across decades. Leadership sets capital allocation, risk appetite, product direction, and disclosure. The day-to-day network quality depends on distributed expertise and institutional memory.

Private ownership leaves core financial questions unanswered

The provided evidence does not establish Arelion’s current standalone revenue, profit, debt, capital expenditure, valuation, traffic volume, contract durations, or customer concentration. The absence matters because a global carrier must keep funding equipment, rights of way, colocation, access, and security capacity. Product announcements reveal activity but do not reveal returns.

Historical deal information describes the 2020–2021 sale, not the company’s value in 2026. Carrying an old deal number into the present ignores traffic, asset, capital-structure, and market-condition changes. Similarly, an interconnection ranking cannot substitute for revenue or margin. A network can be highly connected and still operate in a market where the price per bit falls.

This opacity changes how performance is assessed. Public BGP data and looking glasses can show reachability and path behaviour. Press releases can show path and product activity. Customer announcements can show selected deployments. But none reveals whether capital is earning an adequate return, whether a single customer dominates revenue, or whether indebtedness constrains the next upgrade.

The correct conclusion is bounded: Arelion appears to be an active global carrier with a broad portfolio and continuing investment. The available evidence does not support a profitability estimate or a claim that pension-fund ownership guarantees the next capital cycle.

The economics of renewal sit between physical ownership and contractual control

The investment problem is more complex than choosing whether Arelion owns or leases a particular route. A Tier-1 carrier must maintain several forms of control simultaneously. Fibre ownership gives direct authority over a physical path but leaves dependencies on permits, power, repair crews, landing stations, and equipment vendors. Long-term fibre rights can resemble ownership for operational planning, while contracted commitments lock in years. Buying capacity under shorter contracts preserves flexibility but exposes the carrier to repricing and supplier availability.

The right structure depends on the route, market, and failure consequences.

This point matters because network economics are shaped before the service is sold. Arelion may need to install line systems, routing capacity, cross-connects, and headroom before a new customer commits traffic. A new PoP may require facility contracts and equipment while the commercial case rests on future demand. A high-capacity AI corridor can look attractive at market level while each individual customer remains uncertain about where its workloads will sit. Capital therefore arrives before some of the revenue evidence it is meant to justify.

The same timing problem applies to resilience. A protection path has value precisely when the primary path fails, meaning part of its capacity can appear under-used during normal conditions. Finance teams can see idle margin where operations teams see protection. The network must decide how much apparently idle capacity is economically justified by the contracts and failure scenarios it protects. Too little headroom can turn a fibre cut or equipment failure into customer loss; too much compresses return if pricing does not recognise the reliability value.

Arelion’s broad portfolio makes the calculation harder because the same physical infrastructure and packet layer serve products with different demand patterns. Transit traffic can be volatile and price-sensitive. Ethernet and wavelength services may carry large contracted flows. DDoS protection needs capacity that may remain under-used until an attack. Cloud links depend on placements and the growth of external platforms.

A capacity planner cannot therefore simply sum average traffic and buy the result; instead they must model correlated peaks, protection-driven shifts, maintenance windows, and the possibility that several services move onto the same surviving path.

The AI Direct expansion in 2026 adds another layer of uncertainty. The company has a logical reason to position high-capacity Ethernet and optical services around distributed AI infrastructure, because moving data between facilities is a genuine wide-area need. But AI-linked demand does not cancel the carrier’s ordinary investment discipline. A path still needs to connect physical buildings, ports still need activation, a customer still must pay for capacity, and traffic must persist long enough to justify the committed equipment and rights. A trending workload does not alter the accounting of unused capacity.

Supplier strategy also affects the renewal cycle. Router and optical platforms are not infinitely interchangeable. A major upgrade can create years of operational habituation, sparing plans, software dependencies, and vendor support. Open optical systems and pluggable coherent optics can reduce some lock-in by disaggregating parts of the optical stack, but interoperability still needs engineering and testing. Flexibility is only gained if the carrier maintains the skills and processes that let it be used.

Facilities create a similar form of embedded commitment. A PoP becomes more valuable as customers, peers, and clouds connect to it, but that density makes it harder to move. Cross-connects, customer equipment, maintenance procedures, and commercial relationships accumulate around the location. A facility can therefore gain bargaining power over the carrier even when the carrier owns the routers inside it. The practical question is whether Arelion has alternative sites, path diversity, and enough commercial leverage to prevent a single building or supplier from becoming a hidden strategic constraint.

This is why the private financial opacity matters in an infrastructure file. Without current capex, debt, lease commitments, and cash flow, an outsider cannot see whether the network renewal is funded from operating cash, owner capital, or borrowing—nor how much headroom remains for the next equipment cycle. Technical announcements prove that spending is happening; they do not prove the company can sustain the same pace if prices weaken, demand slows, or resilience requirements rise higher than expected.

The strongest external evidence is therefore cumulative: routes actually completed, not just announced; PoPs attracting customers; modern interfaces becoming orderable; direct paths remaining stable; security systems absorbing larger attack peaks. No single element proves profitability. Together they show whether capital is being turned into an operating system that keeps AS1299 economically relevant.

For Arelion, the duty to renew is not separate from the competitive moat. The advantage comes partly from a network position accumulated over more than three decades. That position saves customers the task of assembling the same access themselves. But it also creates a compound base that must be continuously refreshed. The more valuable AS1299 is as a neutral global path, the larger the potential damage from postponed investment. Tier-1 status therefore confers bargaining power, but it also imposes a permanent capital obligation.

Geography is a service map, a paths map, and a dependencies map

Arelion’s roots and ownership base are Nordic, while its backbone stretches across Europe, North America, and Asia. The company says it delivers service in 129 countries. The number represents commercial reach, not a count of owned fibre systems or direct PoPs. A buyer needs three maps: where Arelion’s own backbone and facilities sit, where service can reach through partners, and where a given service is technically available.

Europe remains a dense part of the network’s history. North America is a major transit and enterprise market, and recent additions such as the Oklahoma City PoP sit inside cloud and AI data centre demand. Arelion’s Danish investment in 2026 attached to paths and a landing station environment important for Nordic and submarine connectivity. In Mexico, path development and a channel programme extended reach into a growing market. The file provides less public detail on paths in Asia compared with Europe and North America.

Submarine connectivity deserves separate treatment. Arelion can buy spectrum or capacity, install equipment in landing stations, and operate an end-to-end service without owning an entire submarine system. Cable faults, landing station incidents, and shared terrestrial backhaul can affect several carriers. Logical path diversity should therefore be tested against cable and duct diversity.

Geopolitics can also reshape path economics. Cross-border rules, sanctions, permits, cable security, and national resilience policies affect where carriers can build and how customers price risk. The file does not establish a specific live dispute involving Arelion, so these factors belong to structural constraints rather than claims.

Competition takes place on several layers simultaneously

At the Tier-1 and global transit layer, Arelion competes with networks such as NTT, Lumen’s Level 3 lineage, GTT, Tata Communications, Cogent, and Sparkle. Direct comparison requires dated evidence on route relationships, geography, latency, capacity, security, support, and price. A single interconnection ranking does not decide the choice.

At the physical and optical layer, firms such as Zayo compete for fibre and wavelength demand. At the cloud connectivity layer, Megaport and Equinix Fabric offer software-defined interconnection across partner and facility ecosystems. At the enterprise layer, regional carriers, SD-WAN providers, and managed service companies can build alternatives over internet connections. These are overlapping categories, not identical ones.

Hyperscaler private backbones present a different challenge. A cloud provider can keep traffic between its own regions inside its own network and bundle transport with compute. Arelion remains relevant when a customer needs connectivity across clouds, enterprises, data centres, and networks beyond a single provider’s control. The more traffic stays inside hyperscale scopes, the narrower some wholesale opportunities become; the more workloads distribute across providers, the more valuable neutral wide-area access becomes.

Internet exchange points can both complement and partially replace transit. A network with enough traffic can peer directly with major destinations and buy less transit. But it still needs access to the rest of the internet, transport to exchange points, and operational support. Arelion’s IX Connect and transit services place it on both sides of this decision.

Arelion’s lasting advantage is not that a single element is impossible to copy. Fibre can be leased, routers bought, cloud on-ramps joined, and DDoS systems deployed. The assembled system—paths, peering relationships, customer prefixes, PoPs, optical routes, support, and reputation—needs years to build and continuous investment to stay relevant.

Interconnection creates both resilience and contagion

A highly connected backbone provides more path options and direct reach. It can help traffic avoid intermediate networks and absorb failures by shifting flows. But the same position increases the number of downstream organisations exposed to its errors. A route leak, filter mistake, or capacity incident can spread beyond direct customers because other networks depend on paths learned through AS1299.

Physical concentration can create the same double effect. Dense PoPs improve interconnection, but a facility outage can affect many services. Logically diverse topologies may share a single fibre duct. A DDoS platform can protect customers, but a diversion-policy error could move a large volume of legitimate traffic. Scale amplifies both capability and damage scope.

Resilience therefore depends on controls less visible than network maps: maintenance discipline, staged change, route filters, RPKI validation, community governance, optical protection, capacity reserves, scrubbing locations, out-of-band access, incident communications, and tested back-out. Arelion’s public tools and product pages show elements of this system exist, but the file does not contain a complete incident record or independent audit.

Customers have responsibilities too. Multi-carrier arrangements, prefix design, route policy, testing plans, and local access diversity determine how much of the provider’s resilience turns into actual protection at the customer edge. Buying from a Tier-1 does not remove the need for architecture that can survive losing that carrier.

Sustainability claims need disciplined measurement

Backbone growth consumes equipment, energy, colocation, construction, and maintenance. New coherent optics can increase capacity per watt and per rack-unit. That efficiency can coexist with rising total energy consumption if traffic and network spread grow faster. Arelion publishes a sustainability framework, but the available evidence does not provide a complete, independently audited footprint of the network and supply chain.

Responsible editorial treatment separates efficiency from absolute impact. A path upgrade may cut energy per transported bit; a new path may add total equipment and power. Leasing infrastructure also shifts some emissions and energy use to suppliers rather than eliminating them. Without consistent scope, baseline, and traffic data, broad claims of net environmental improvement cannot be supported.

Sustainability matters strategically because customers and pension-backed owners may demand credible energy, procurement, and resilience evidence. It matters operationally too: optics density, cooling, and power availability affect where capacity can be deployed. The worth-watching question is whether reporting becomes precise enough to link network expansion with measured resource use.

What public evidence cannot show

The network-scale numbers in this file are mostly current claims from Arelion itself. The research file does not contain an independent, route-by-route audit of more than 80,000 km, a standardised definition of each PoP, or a complete list of what is owned versus leased. Nor does it contain a comparable traffic series across competitors.

The interconnection claim is metric-dependent. The “95% within one hop” phrase describes Arelion’s claimed reach in the United States and Europe, not a guarantee for every destination, protocol, or moment. Monthly performance reports are useful but provider-selected. A purchasing decision still needs path-specific testing and contractual terms.

Product availability is also conditional. Ethernet lists speeds up to 400 Gb/s; that does not mean every site, link, or path offers 400G. AI Direct is a connectivity bundle, not evidence of AI customer revenue. A 6.1 Tb/s observed attack does not reveal contracted scrubbing capacity. MEF/Mplify certification supports service-definition confidence but does not certify every access link or incident outcome.

The financial gap is wider. The provided materials do not disclose current revenue, margin, debt, capex, valuation, or customer concentration. That prevents building a full economic model. The gap should be kept visible rather than filled with assumptions drawn from path ranking, customer count, or owner identity.

Renewal is continuous even when the routing identity looks stable

AS1299 can look strikingly stable from the outside. The AS number persists, known peer relationships appear intact, the Twelve99 name continues to signal operational continuity, and customers continue to see a global routing table. This continuity is useful, but it can hide how much change is required beneath the surface. A backbone can maintain a single public routing identity while swapping out routers, optical platforms, line cards, software releases, power arrangements, fibre paths, and facility connections. The real product is the continuity that results from managed change.

Hardware generations create a first renewal clock. Router capacity must track larger interface speeds, growing table sizes, telemetry demands, and security functions. A chassis or line card that was sufficient when 100G was dominant can become a constraint when 400G ports and higher aggregate traffic become normal. The timing is not just about whether a device still forwards packets. Sparse, vendor support, power density, software compatibility, and the cost of operating two generations together can make a technically working platform economically obsolete.

Optical systems create a second clock. Coherent technology can push more usable capacity onto the same fibre pair, but the upgrade is not a universal software switch. Distance, fibre characteristics, spectrum, amplifiers, ROADMs, open-line-system design, and operational qualification determine what can be carried on a path. So Arelion’s ability to offer higher-capacity services depends on path-specific engineering, not the nominal capability of a single optical component.

Software adds a third clock. Routing platforms receive security patches, protocol changes, new features, and operational fixes. A stable network may avoid the newest release until it proves itself, but staying too long on an old release increases support and security risk. An operator must stage upgrades so that fixing a software defect does not create a bigger availability problem than the defect itself. For a Tier-1 backbone, change-management quality is as important as code quality; a single policy or control-plane error can affect traffic far beyond one customer.

Internet routing standards and operational practices create a fourth clock. RPKI adoption, filtering expectations, peering policies, BGP community conventions, and routing-security practices evolve without changing an AS number. A network considered well-operated ten years ago cannot assume its old controls remain sufficient. Customers and peers gradually demand better validation, more transparent path policies, and stronger responses to leaks and hijacks. Maintaining trust therefore requires procedural renewal as much as capacity renewal.

Facilities have a slower but influential cycle. A PoP can stay in the same building while power density, cooling, cross-connect economics, and the peer mix change around it. A carrier may find that a historically important site is no longer optimal for new high-capacity equipment, while moving existing customer and peer connections is expensive. New data centre clusters can pull traffic away from traditional carrier hotels. The network must add access to emerging hubs without weakening the interconnection density that made the older sites valuable.

Customer geography creates another source of movement. Enterprise applications shift to cloud regions; content platforms add edge locations; AI clusters appear around power-rich data centre markets; wholesale networks open new national points. Arelion can hold the same PoP count and become less relevant if those points are no longer close to important demand. Conversely, targeted path additions can increase the value of the existing backbone without a radical change in the overall footprint.

Security capacity follows a more uneven demand curve. Normal traffic can be planned against familiar peaks, but DDoS systems must absorb inherently rare events. The largest observed attacks in one year can become ordinary events years later as botnets, access speeds, and attack techniques grow. A carrier that sizes scrubbing capacity to average usage is unprepared; a carrier that builds unused unlimited capacity spoils economics. The renewal problem is to maintain a distributed margin and diversion flexibility sufficient for reasonably probable attacks while using the same infrastructure efficiently under normal conditions.

Access partners also have a lifecycle. A local carrier may improve, merge, change ownership, withdraw a product, or become less competitive. The path that was the best enterprise access option three years ago may not be the best today. Global service quality therefore depends on continuous supplier qualification and the ability to steer new orders, and where possible existing customers, toward better alternatives. The partner count measures breadth; the renewal burden is deciding which relationships still deserve traffic.

Contracts can age even when the infrastructure does not. Long-term capacity rights can look economically attractive when signed and then become expensive compared with later market prices. Short-term contracts preserve flexibility but expose the company to repricing. Facility, fibre, and supplier contracts have renewal dates that may not align with customer contract terms. The network operator must manage these contractual clocks alongside the physical ones so that a path does not become financially unattractive simply because several commitments renewed at a bad time.

People create a less visible renewal requirement. Global backbones depend on operational knowledge of path policies, optical characteristics, historical incidents, supplier behaviours, and customer exceptions. Some of this knowledge can be encoded in inventory systems and automation; some remains institutional. Retirements, reorganisations, and outsourcing can remove context whose importance only becomes clear during an unusual fault. A stable network therefore needs functional succession, documentation, and training just as it needs spare optics.

Automation does not stop these clocks. It can improve inventory accuracy, speed up provisioning, and detect dependencies, but it also increases the need for trustworthy data. A wrong facility identifier, a stale path record, or an incorrect supplier hand-off point can propagate faster through an automated workflow than a manual one. The more Arelion standardises provisioning and operations, the more important the underlying data model verification becomes.

The commercial consequence is that renewal spending should not be assessed only by what appears as “expansion”. Replacing a line card before it fails, adding protection capacity to an existing route, moving a customer away from a shared-risk path, or updating a security platform may not add a country or a new PoP. These investments protect existing revenue. A private owner measuring growth by headlines alone could underweight them; an operator treating every preventive upgrade as mandatory could overcapitalise.

The useful discipline is to connect each renewal decision to a failure domain, a capacity constraint, a customer commitment, or a measurable operational improvement.

The simple question “Is AS1299 still Tier-1?” is therefore too narrow. The routing relationship can remain intact while other aspects of the service improve or degrade. A broader test asks whether the network keeps pace with traffic growth, whether physical diversity remains real, whether peers and customers retain confidence, whether security capacity tracks the threat environment, and whether capital arrives before technical debt appears as outages.

Arelion’s long history is both evidence and burden. More than three decades of operation mean the system has survived several generations of hardware, traffic, and ownership. It also means the network contains decisions made in very different eras of technology and demand. Renewal is the process of preserving what still creates value and replacing what no longer fits. An AS number can stay constant precisely because so much underneath it does not.

Buyers should verify the service boundary, not settle for the Tier-1 label

For a customer, useful due diligence begins after confirming that AS1299 is a Tier-1 network. The next question is where Arelion’s direct control starts and ends within the service being purchased. A transit port inside an Arelion PoP, a managed enterprise circuit delivered via a local partner, a wavelength on an Arelion fibre path, and a Cloud Connect service expose different failure domains. The same brand and the same backbone can sit behind fundamentally different operational contracts.

The physical hand-off point should be explicit. A customer needs to know whether the delivery point is inside an Arelion PoP, a third-party data centre, the customer site, or a partner facility. If the service includes a local access link, it must know who owns that link, where it enters the building, and whether a second circuit genuinely uses a distinct physical path. Two circuit IDs do not prove diversity. The evidence needed is separation in duct, building entry, facility, power, and upstream path.

The routing boundary matters independently. A transit customer controls its ASN and can use communities, preferences, and multi-homing. A DIA customer delegates more edge policy. An Ethernet customer may receive a Layer-2 hand-off where internal routing is invisible. A Cloud Connect customer reaches an on-ramp where a second administrative scope begins. Procurement should record these differences because they determine who can act during a fault.

Capacity should be tested at the exact purchased boundary. A backbone may support 400G while a local port, exchange point, cloud on-ramp, or access carrier does not. A customer planning high-capacity AI or data centre traffic must distinguish between core network capability and the service that is actually orderable at both ends. Path qualification, port availability, and protection options are more useful than a generic statement that the provider “supports” a given speed.

Latency claims also need path-specific definition. Arelion’s interconnection can reduce intermediaries, but the path with the fewest AS hops is not automatically the shortest physical distance or the lowest latency. BGP policy, fibre distance, local access, and destination location all matter. Customers with latency-sensitive applications should measure the path from relevant sites and understand whether the contracted service uses traffic engineering, protected paths, or ordinary internet routing.

Availability commitments need the same discipline. The difference between backbone, PoP-to-PoP, and end-to-end SLAs shows why a single number is limited public evidence. A customer should ask which components are included, how planned maintenance is treated, what exclusions apply, how it is measured, and what remedy follows a breach. A high percentage can still permit damaging downtime if the measurement window and compensation do not align with the application’s cost of failure.

DDoS purchasing must separate detection, diversion, scrubbing, and emergency blackholing. A carrier’s ability to observe multi-Tb/s attacks is useful evidence of visibility, but the customer needs to know its own protection status, activation policy, protected prefixes, clean-traffic return path, and escalation procedures. It also needs to know what lies outside the service—in particular application-layer attacks that require different controls.

Cloud connectivity requires a three-party operating model. Arelion can deliver the carrier path; the cloud provider owns the interface and internal fabric; the customer controls the account and routing. Troubleshooting works best when identifiers and communication channels are recorded across these scopes before an incident. “Private connectivity” is valuable because it narrows the path and increases control, but it does not merge the three parties into one.

Partner-delivered access deserves special attention because it is where a global carrier’s consistency is hardest to maintain. A buyer should ask whether Arelion monitors the local circuit, receives proactive alarms, controls supplier escalation, and can obtain path information. It should also know what happens when a local provider repeatedly misses repair targets. The value of a global contract is higher when the primary carrier has enough data and commercial leverage to change the outcome, not just to pass tickets.

Operational transparency should be assessed before purchase, not only during an outage. Looking glasses, performance reports, and the MyArelion portal provide useful views, but a customer should define what metrics are available for its service, how often they are refreshed, and whether raw evidence can be exported. When performance becomes disputed, shared timestamps, path notes, and incident logs are more valuable than a blanket service-status message.

Change management is another boundary. Customers need to know which changes they can execute through the portal or BGP community, which need provider approval, and which Arelion can initiate during maintenance or security response. Delegated control is useful only when the action, blast radius, and rollback path are understood. This is especially important when several logical services share a single physical port.

A multi-homed customer should test the failure case rather than assuming redundancy simply because it bought from two providers. Path preferences may keep traffic on a failed or degraded path longer than expected; a customer’s prefixes may be filtered incorrectly; inbound and outbound traffic can behave differently. Regular failover tests reveal whether the architecture actually uses AS1299 and the alternative carrier as intended.

The same principle applies to optical diversity. A protected wavelength may use two fibre paths, but buyers should confirm how they are defined and where they rejoin. A dual-link service with diverse ends can give a stronger boundary than two channels that share a site. Application architecture may need to span the same data centres if the facility is a critical failure domain.

Contract renewal should re-examine these assumptions. A service bought three years ago may today be delivered via a different access provider, facility, or network design. New clouds and data centres may offer better endpoints. The provider may have added direct paths that change the optimal design. Treating renewal as an administrative price negotiation misses the opportunity to update the resilience model.

For BTW readers, this due diligence explains the practical meaning of Tier-1. Arelion’s AS1299 gives it a strong routing position, but customers buy services at specific hand-off points. The backbone’s value is only realised when routing policy, physical infrastructure, partner access, security, and operations work together at those boundaries. The label reduces one class of dependency; it does not remove the need to understand the rest.

Central judgement

Arelion illustrates what a Tier-1 carrier has become. Settlement-free access on AS1299 remains the defining network position, but the commercial product is a multi-layer services platform. Fibre and optics provide path and capacity. PoPs provide market access. BGP, communities, and path security provide policy. Ethernet, cloud, exchange, and managed services create different hand-off points. DDoS systems use backbone visibility as a security function. Operations and customer support turn all these layers into something an enterprise or carrier can buy.

The moat is historical accumulation. Arelion inherited decades of routes, facilities, peering relationships, and operational knowledge from the Telia Carrier lineage. Polhem Infra ownership gave the system a focused infrastructure investor and a new brand without replacing the technical identity. Twelve99 remains visible because network memory is longer than corporate marketing.

History itself imposes an obligation. Tier-1 status does not freeze a network at the moment of achievement. Traffic prices change, attack peaks grow, clouds internalise transport, AI creates demand for new corridors, optics shift to higher speeds, and customers expect more automation. Arelion must renew the platform while protecting the routing trust that gives AS1299 its value.

The evidence supports confidence that Arelion is active, globally relevant, and technically broad. It does not yet support confidence in current profitability, indebtedness, or the speed at which AI announcements turn into durable revenue. The most useful conclusion is therefore practical: Tier-1 today is a routing relationship surrounded by capital-intensive fibre, optical, packet, security, and operational systems, whose quality is repeatedly proved at service boundaries rather than granted once by the name.

Indicators that will reveal whether the backbone is strengthening

Arelion’s next chapter should be judged by what turns into results, not just by slogans. The company already has a mature routing identity, global-scale claims, and a broad catalogue. The open questions are whether new paths become activated demand, whether higher-value services offset transit price pressure, and whether resilience keeps pace with growing interconnectedness. The indicators below connect technical activity to commercial and operational outcomes.

Peer and path stability

Tier-1 status depends on continuing settlement-free relationships. Material changes in AS1299’s visible peer relationships, AS-path lengths, or propagation would be an early signal that the economics of interconnection are shifting. Looking glass results and external BGP monitoring can show change but cannot explain the commercial terms behind it.

Activated capacity rather than announced capacity

New 400G products, optical upgrades, and path announcements matter when ports light up, traffic grows, and customers commit. Watch for qualified 400G availability, completed Nordic and North American corridors, new PoPs, and named customer deployments. A high volume of announcements without activation evidence weakens the AI and capacity narrative.

Product mix beyond transit

Traffic volume can grow while the price per bit falls. SecureConnect, Cloud Connect, IX Connect, managed optical networks, and Ethernet attach rates to existing accounts would signal that Arelion is earning income from integration rather than relying on commodity transit alone. The company does not currently disclose revenue by product, so customer and service announcements are the available proxies.

Access partner performance

Growth through about 450 access partners extends reach but can introduce variability in delivery and repair. Watch the gap between direct PoP outcomes and end-to-end service, local circuit lead times, repeated regional incidents, and whether Arelion discloses stronger partner qualification. The published SLA gap already shows that the last mile changes the risk profile.

DDoS scale and mitigation evidence

The 2026 report establishes the attack environment Arelion observed. Subsequent reports should be read for peak growth, botnet composition, mitigation time, customer impact, and scrubbing distribution, not for a single headline number. A larger observed attack is not itself proof of stronger protection.

Capital and governance signals

Private ownership limits direct financial analysis. Watch leadership changes, completed path projects, hiring, supplier commitments, any disclosed funding, and Polhem Infra’s shifting investment posture. A slowdown in deployments or increased reliance on partners could reflect capital discipline, demand uncertainty, or both.

Five evidence-supported scenarios

Wide-area AI demand becomes a sustained growth driver

Neoclouds, enterprises, and researchers distribute data and workloads across facilities. Arelion activates 400G EVPL and wavelength services on named corridors, and AI Direct becomes a measurable source of traffic and customer expansion.

Traffic grows while economics tighten

AS1299 carries more bits, but transit prices fall faster than usage rises. Arelion remains technically important while returns increasingly depend on cloud, security, Ethernet, and managed optical services.

Security convergence improves customer retention

Customers buy transit or DIA with automated mitigation, and SecureConnect gains a larger role in renewal and incident response. Success shows in attach rates, customer references, and evidence of sustained availability during large attacks.

Hyperscalers internalise more transport

Cloud providers carry more traffic inside their own backbones and reduce some wholesale demand. Arelion remains relevant where enterprises need neutral, multi-cloud and data centre paths, but corridor economics become more selective.

An interconnected shock in a path, facility, or capital reveals concentration

A peering change, fibre incident, facility outage, or upgrade delay affects several services simultaneously. The strategic outcome depends on physical diversity, capacity reserves, incident transparency, and the owner’s willingness to fund remediation.

Professional implications by stakeholder

Network operators should test route policy, community behaviour, and multi-carrier design rather than treating Tier-1 as synonymous with backup. Enterprise buyers should map the direct and partner-delivered portions of each service. Cloud teams should identify the boundary between Arelion and the provider on-ramp. Security teams should separate observed attack scale from contracted mitigation. Investors and owners should connect route expansion with usage and return. Regulators and resilience planners should examine concentration in facilities, cables, and autonomous systems without assuming that ownership is the only form of control.

Control, incentives, and the decisions that shape AS1299’s next chapter

Arelion’s leadership problem is not choosing “the network” or “the business”. The network is the business, and every technical choice locks in capital and multi-year contractual commitments to suppliers and customers. The control map is distributed: Polhem Infra controls ownership and capital; management allocates capital and sets product priorities; engineering and operations control path and capacity changes; peers define settlement-free relationships; facilities and access partners control parts of the physical path; customers control prefixes, demand, and multi-carrier arrangements; cloud providers control the far side of the on-ramps.

A strategy only works when these authorities stay aligned.

Decision one: Fund physical diversity before marketing reach

Arelion should measure path diversity at the cable, duct, landing station, facility, and power levels, not just by logical topology. A new PoP or path can add visible reach while sharing a hidden failure domain with existing capacity. The hard-to-reverse risk is a long-term commitment to infrastructure that cannot deliver the resilience sold on top of it.

The leadership choice is to make physical diversity evidence part of capital approval and customer service design. That may increase cost or slow some launches, but it reduces the chance that several revenue-producing products fail in a single incident and protects the credibility of the global map. A carrier that uses a mix of physical control forms can be trusted when it demonstrates its control over failure boundaries.

Decision two: Preserve interconnection neutrality while monetising direct access

AS1299’s advantage depends on peers and customers continuing to see the network as a trusted exchange partner. Product growth must not distort routing policy in ways that weaken that trust. Commercial pressure can encourage aggressive preferences, concentration on large accounts, or under-investment in paths that appear less profitable but preserve the topology.

Leadership should protect a clear separation among settlement-free peering policy, customer transit policy, security interventions, and product traffic engineering. The second-order benefit is easier incident diagnosis and stronger counterparty confidence. The third-order risk is the loss of direct relationships that cannot be restored simply by buying more fibre.

Decision three: Demand activation evidence for the AI portfolio

AI Direct gives Arelion a compelling language for describing high-capacity data centre interconnect demand. It can also become a wrapper for ordinary circuits relabelled without new economics. Management should calibrate the portfolio through per-path availability, activated ports, named use cases, utilisation, protection requirements, and renewal behaviour.

This discipline prevents capital from chasing market enthusiasm without customer evidence. It also improves product design because training data, inference traffic, replication, and cloud access have different latency and security requirements. The hard-to-reverse risk is over-building corridors or optical capacity where demand rests on forecasts that do not materialise.

Decision four: Turn security evidence into an operational feedback loop

Arelion’s DDoS report gives it a valuable attack-traffic perspective. Leadership should connect the publication to capacity planning, customer architecture, scrubbing distribution, false-positive review, and post-incident learning. Publishing the peak size alone rewards spectacle; publishing mechanisms and outcomes improves the service.

A stronger loop distinguishes observed attacks, mitigated attacks, customer impact, and the role of blackholing. That helps customers price risk and helps the owner assess whether security capital is protecting revenue. The long-term consequence of weak evidence is that SecureConnect becomes a commodity label rather than a trusted operational capability.

Decision five: Make partner access governable

Access partners are essential for global enterprise delivery, but accountability can fragment across contracts. Arelion should treat partner selection, service list, incident data, diversity, and exit rights as part of the product architecture. The customer should know which parts are direct, which are managed, and which are merely coordinated.

The second-order effect of strong governance is a wider addressable market without equivalent owned build. The third-order effect of weak governance is that a single local supplier lowers the global backbone’s reputation. Some failures will always lie beyond Arelion’s physical control; they need not lie beyond its evidence and escalation control.

Decision six: Disclose enough economics to sustain institutional trust

Private ownership gives Arelion flexibility, but it leaves customers, partners, and infrastructure stakeholders unable to assess indebtedness, capex, and customer concentration. Full public-company disclosure is not required to improve trust. Management and the owner can publish consistent metrics on operational investment, capacity activation, resilience, and sustainability without revealing sensitive contract detail.

Better disclosure would make path announcements more meaningful and reduce the temptation to treat interconnection as a proxy for financial health. It could also internally discipline capital allocation by tying strategy to comparable outcomes. The risk of continued opacity is that counterparties apply their own—often more extreme—assumptions during market stress.

Second-order effects of execution success

If Arelion turns long-term capital into physically diverse capacity, preserves peer trust, and attaches higher-value services, AS1299 becomes more than a commodity transit path. It could become a broad neutral platform for clouds, enterprises, carriers, and distributed AI infrastructure. That could improve customer retention, support better capacity forecasting, and make security investment more economically defensible.

Success also affects the surrounding ecosystem. Access partners gain traffic and pressure toward standards; data centres gain interconnection value; cloud providers gain enterprise access; exchange points gain remote entities; customers gain an alternative to concentrating all transport inside one hyperscaler. The benefit is distributed, which is why no single entity controls the full outcome.

Third-order effects of failure

A severe routing-policy error or a hidden physical concentration could harm more than one product. Customers may reroute traffic, peers may reassess trust, security services lose credibility, and the owner faces a larger capital need at the moment revenue is at risk. A highly connected infrastructure can transmit reputational damage as fast as the packets themselves.

Prolonged capital constraint would have slower but lasting effects. Delayed optical and PoP upgrades could lengthen paths, increase partner dependency, and make the network less attractive to high-capacity customers. Once direct relationships and anchor accounts shift, rebuilding them can take years.

Irreversible risks

The hardest risks to reverse are the loss of settlement-free peering relationships, commitment to physically shared-risk paths, under-investment during a hardware generation, dependence on a single large customer segment, and a security incident that reveals weak operational control. Brand damage can be repaired; a lost path position or a stranded long-term infrastructure contract can persist.

Leadership should therefore separate reversible product experiments from hard-to-reverse network commitments. A portal feature can be changed; a fibre right, a landing station dependency, or an optical build can constrain the company for years. A commercial offer can be withdrawn; a peering relationship broken by policy misuse may not return on demand.

Leadership judgement

Arelion’s strategic asset is not the Arelion brand or the Twelve99 hostname alone. It is the coordinated system behind AS1299: the interconnection relationships, the physical paths, the optical capacity, the packet policy, the security, the operations, and the trust of customers and peers. Polhem Infra and management inherit this system rather than starting from a blank sheet. Their task is to renew it without breaking the trust accumulated under its previous identity.

The critical test is disciplined conversion. New routes must become usable, diverse paths. New ports must become customer traffic. AI positioning must become activated services. DDoS visibility must become preserved availability. Partner access must become governable delivery. Private capital must become timely upgrades without hiding risk. When these conversions succeed, Tier-1 status remains economically meaningful. When they fail, the label outlasts the advantage it once described.