Telecom will not scale sustainably by building more infrastructure; it will only scale by managing networks as software.
Part 1 of this series argued that telecom outsourced control of its own future. The consequence was predictable: slow execution, rising costs, and an industry structurally unable to change itself.
Part 2 addresses the next question: if control is the root issue, what does a credible alternative look like in practice?
The answer is not another technology upgrade. It is a change in how networks are conceived, built, and operated. Intelligent Growth means scaling networks the way modern software businesses scale platforms, not the way industrial infrastructure has historically been expanded.
Traditional telecom growth follows a familiar pattern. Demand increases, so capacity is added. New spectrum is deployed. More sites are built. More hardware is installed. Headcount grows to manage the added complexity. Costs rise linearly with scale.
This model worked when mobile growth was primarily about general coverage, basic capacity and everybody didn’t have the service. It fails as networks become massive, agility becomes a requirement and market penetration becomes greater than 100%.
The industry now runs national networks with hundreds of thousands of active cell sectors, thousands of data centers, and real-time service expectations. Yet many operators still rely on planning cycles, organizational structures, and operational processes designed in the 1990s.
The result is visible everywhere. Software upgrades take nine to twelve months to reach production. Network changes require coordination across multiple vendors and teams. Automation remains localized rather than systemic. Even modest change becomes risky and slow.
Scaling infrastructure, operations, and business this way does not produce Intelligent Growth. It produces the managed decline in top-line and top-of-mind relevance – which we have seen.
Modern software businesses scale differently. They invest upfront in platforms, automation, and abstraction so that growth and cost are disassociated.
We call this hyperscale economics:
Once the software platforms exist, adding users, features, or capacity does not require proportional increases in people or cost. Software absorbs complexity so organizations do not have to.
This is not theoretical. It is observable across cloud platforms, digital marketplaces, and hyperscale services. The common word is “technology leaders”.
The same principles apply to telecom networks, but only if the network is treated as a software system end to end.
That requires three shifts:
Progress is measured in continuously improving economic units rather than generational uplifts. This is the difference between scaling infrastructure and scaling software.
Rakuten Mobile provides a concrete example of what software-centric scaling enables.
As of year-end 2024, the network covers more than 99 percent of the Japanese population, operates over 94,000 4G and 5G base stations[ref], and has continuously won quality awards for performance[ref].
The entire network is open, cloud-native and fully automated using software and AI. Operational scale is not achieved by expanding teams. The network is run by around 250 people. That number is intentionally capped, even as the network continues to grow. Scale is absorbed by automation, not headcount.
Sites are planned, built, commissioned, and provisioned through zero-touch processes. Nodes are named automatically. IP addresses are allocated automatically. Inventory is tracked in real time across physical and digital assets.
This was not possible at the start. It required years of hard engineering work to build the underlying software systems that make automation safe and repeatable. But once those foundations exist, growth accelerates rather than slows.
This is Intelligent Growth.
The most common misunderstanding in industry discussions is around treating technologies like Open RAN or cloud as the end goal.
They are not.
Open interfaces and cloud-native architectures are enablers. They matter because they allow the network to behave like software, not because they are fashionable standards.
Without a unified operational platform, disaggregation simply moves complexity around. With the right platform, disaggregation becomes a source of speed and control.
In a software-scaled network, every operational activity is digitalized. Planning, rollout, fault management, optimization, and customer experience are connected through shared data and automation.
Signals that matter most are not hardware alarms but customer impact. In practice, this means correlating network behavior with real-time experience data rather than reacting to isolated technical events.
It also means treating operations as a continuously improving system. Every automated process can be refined, reused, and extended. Improvements compound over time.
This is how software businesses improve margins while growing. Telecom can do the same, but only if operations are treated as a software product rather than a cost center.
One of the clearest indicators of infrastructure thinking is software delivery speed.
In many networks, introducing or upgrading software takes most of a year. Lab testing, integration, acceptance, and rollout are manual, sequential, and risk-averse.
This cadence is incompatible with modern digital services.
In a software-scaled network, software is distributed through controlled marketplaces. Applications are tested in sandboxes, published, and deployed through automated pipelines. Rollbacks are routine. Observability is built in.
This model already exists in consumer technology. Mobile app ecosystems moved from carrier-controlled deployment cycles to self-service platforms years ago. The same logic applies inside the network.
Reducing software iteration from months to weeks or days is not about convenience. Time is money. Faster iteration directly affects cost, reliability, and the ability to create new services.
Technology change is rarely the limiting factor. Organization and skills are.
Every successful transformation combines three conversations:
Most telecom organizations are structured to prevent change. Budgets are allocated vertically. Incentives reward stability over improvement. Skills are optimized for procurement and vendor management rather than system design and software ownership.
Scaling as software requires different mindsets, different incentives, and different governance. It starts with challenger leadership and belief in the possible even if different. It requires trusting platforms over process and accepting that automation will outperform manual coordination at scale.
This is uncomfortable, but unavoidable.
We have only mentioned AI once in passing. AI is important but must be seen as “yet another technology”, and not a business outcome – even though it is the most important technology to arrive in 20 years. The last technology revolution was web 2.0; this changed how we live and created completely new technology leaders.
However, web 2.0 did not help telecom change any of its business fundamentals – some say it made telecom reality worse. We must not miss on the opportunities of AI in the same way. But we must also remember that AI is not an outcome but an enabling technology.
Telecom transformation fails when it is treated as a technology replacement rather than an operating model change. Rakuten’s experience entering traditional industries is consistent: digitalize end-to-end processes, move interaction and support online, and then improve cost, speed, and performance through software.
Rakuten Mobile applies this approach. Traditional vendors are used everywhere, but only as component suppliers. Each vendor fits into Rakuten Mobile’s operating model, not the other way around. Control of system design remains with the operator, which is what enables differentiation, speed, and long-term relevance.
Rakuten Symphony exists to make this approach repeatable.
Platforms, blueprints, and proven operating models
Rakuten Symphony provides the same platforms, blueprints, and operating models used by Rakuten Mobile, refined through live national network operation. The goal is not to replicate Rakuten Mobile, but to allow each operator to retain ownership of its own system design while avoiding the cost and risk of discovering everything from scratch.
This gives operators control of their operating model, the ability to differentiate, and the benefit of learning from proven execution at scale.
Modernization without rip and replace
Transformation does not require everything to change at once. Successful modernization is phased and pragmatic.
Existing networks, vendors, and systems remain in place while new capabilities are introduced incrementally. Open interfaces allow new platforms to integrate with legacy environments, creating a common control layer first and modernizing underlying components over time.
The architecture shown illustrates how Rakuten Symphony platforms support TM Forum-defined interfaces across planning, orchestration, assurance, and automation. This interoperability is what enables modernization as a journey rather than a reset.
Retaining ownership of the future
The objective is not compliance with a standard or adoption of a specific technology. It is control.
Operators that retain ownership of their operating model retain the ability to evolve, automate, and monetize on their own terms. Rakuten Mobile has proven this at scale. Rakuten Symphony enables others to do the same, without losing control of their business in the process.
Part 3 of this series will focus on how telecom moves from connectivity to new revenue in the AI era. That conversation only makes sense if intelligent growth is already in place.
Without software-centric operations, networks cannot adapt fast enough to support ecosystem participation, customer-level personalization, or AI-driven services. With it, the network becomes a programmable asset rather than a static utility.
The opportunity is not to run networks more cheaply, although that matters. The opportunity is to make telecom relevant again as a digital platform. That starts by scaling networks as software, not infrastructure.
To revisit the first article in the series titled “The Control illusion: How Telecom Outsourced Its Own Future”, click here.
