• Rakuten Mobile argues operators must reorder their priorities for 6G
• Control, customers, business and operations should trump technology
• Do operators want to cede sovereignty for yet another 10 years?
• Or do they want to own their future?

The mobile industry has a habit of marketing each generation based on the same metric: Peak downlink speed. It served its purpose in justifying infrastructure upgrades driven by user demand for devices with faster download capabilities, at least up until 5G. 

However, speed is no longer the marketing hook it once was. Few operators are likely to pin their 6G marketing around ‘faster downloads’ and even fewer customers are likely to pay a premium for the privilege. The International Telecommunication Union (ITU) has been strongly suggesting that 6G should feature a five- to 10-times increase in theoretical peak data rates from those of 5G. And yet standards organisation the 3GPP is currently working with an agreed air interface KPI that only supports a 1.8-times peak rate increase from 5G. 

So where does this leave network operators?

In a new strategy paper, Operator-Controlled 6G: From Connectivity Infrastructure to Guaranteed Digital Services, David Soldani, senior VP of next-generation advanced research at Rakuten Mobile, argues this framing is precisely the sort of trap operators must avoid. He states that 6G is the last realistic chance to escape a structural problem that has been building ever since the analogue age and has based his evidence on Rakuten's own national-scale Open RAN network.

The problem, as Soldani frames it, is not technical capability but ownership. 

From 2G to 5G, the OSS, BSS, RAN and core have been delivered as vertically integrated stacks in which the vendor controls the data model, the upgrade cycle, the integration interface and now the AI inference layer. Operators have become highly proficient at procuring, integrating and operating networks they do not own, on platforms they cannot modify, with AI systems they cannot audit. Changing a network behaviour means raising a change request and waiting for a software release which, with some vendors, can take a year. His verdict is that this is not a procurement inconvenience but a progressive loss of strategic autonomy. 

The question for 6G, he argues, is not whether it will be technically superior to 5G (that's a given) but whether operators will design and control the platform or merely consume it.

Reorderings and new frameworks

The paper's organising idea is a reordering of five priorities:

• Control First demands that operators own the software-defined control plane – the layer that expresses policy, enforces guarantees, manages spectrum and directs automation. 
• Customer First reorders what networks are optimised for: Customers do not buy peak data rates, they buy outcomes – a video call that holds, a surgical robot that does not lag, a campus network that guarantees throughput to every machine tool. 
• Business First confronts the reality that connectivity pricing has hit its ceiling, with average revenue per user (ARPU) flat or declining across mature markets for a decade despite rising investment. 
• Operations First accepts that a fully instrumented 6G network will generate data and decisions at volumes no human team can process, making agentic AI a necessity rather than a luxury.
• Technology Last. Yes, this does indeed appear to be somewhat counterintuitive. However, Soldani explains that this is not a downgrade but a sequencing principle. When operators let vendor research agendas and standards-committee momentum define the network, he argues, they reproduce the dependency cycle of every prior generation. Technology selection must follow from the four priorities above it, rather than drive them.

Soldani has also created two paired constructs. The first is the ‘6G Control Compact’, which he explains as a three-layer ownership taxonomy that sorts every architectural component into what operators must own, what they can federate, and what they should consume as commodity. This approach allocates sovereignty according to strategic value rather than treating the whole stack as something to be bought.

The second construct is the ‘Guarantee Economy’, described as a five-tier, outcome-priced commercial model that turns control into revenue. Its logic is that a guarantee without enforcement is merely a marketing claim. To make one contractually meaningful, an operator needs the ability to define service-level objectives in granular, measurable terms; the ability to enforce them in real time through closed-loop automation; and the ability to verify compliance through transparent, auditable telemetry. Pricing combines an upfront capability premium, SLA (service level agreement) penalties that give the guarantee credibility, and usage-based metering for applications needing assured connectivity only intermittently.

Soldani is candid about where this runs into limits. He acknowledges the 5G enterprise service-level agreement (SLA) proposition of private networks, guaranteed slicing and deterministic quality of service (QoS) has largely failed to materialise at the projected scale, with many enterprises buying spectrum and equipment directly and bypassing operators altogether. He argues the 6G version is distinct because it is a pure service contract requiring no enterprise capital expenditure. This is because the AI-inference edge and sensing tiers create demand with no 5G precedent, and because closed-loop enforcement enables outcome contracts that the manual operations of 5G could not support. 

The evidence base

What distinguishes this paper from the usual 6G manifestos is that its claims are anchored in an operating network rather than a roadmap. Soldani notes that Rakuten Mobile is the only operator to have built and run a fully cloud-native, Open RAN network at national scale, passing 10 million subscribers in 2025 and reaching its first full-year earnings before interest, taxes, depreciation and amortisation (EBITDA) profitability in 2025. (And its progress continues, as CEO Sharad Sriwastawa noted recently.)

Through the mobile operator’s sister company, Rakuten Symphony, the architecture has been commercialised for partner operators across multiple markets. The deployment, he argues, has generated operational insight available to no other operator or vendor: The failure modes of cloud-native network functions under production load, the tooling for zero-touch RAN configuration, and the organisational processes that a software-driven network demands.

The operations target he sets out is explicit: NGMN Level 4 autonomous, closed-loop control within a Level 5 governance framework, where humans set policy and exception thresholds rather than approving individual agent decisions. This is a combination that Soldani ties directly to EU AI Act compliance.

He lists seven calls to action, along with a caveat that the Guarantee Economy model applies most readily in high-ARPU markets. The paper concludes with a three-phase roadmap:

• 5G-Advanced as the on-ramp until 2027
• Early commercialisation to 2029
• 6G at scale from 2029 to 2032.

Soldani’s central claim is that the architectural decisions taken in the 2026-28 window will determine whether the industry’s default 6G model is operator controlled or vendor dependent.

For three decades, the industry has sold each generation on a peak data downlink number. But no enterprise has ever signed a contract for a peak data rate, and no customer has ever experienced one. Customers, whether consumers or enterprises, buy outcomes. Operators should treat 6G as an architectural reset: Owning the control plane, measuring outcomes, pricing guarantees, and running the network as software with agentic AI at the helm. They should stop selling the downlink and start owning the network.

– Guy Daniels, Director of Content, TelecomTV