Summary

  • Per Narvinger’s career spans research and development, product management, customer-facing sales, network systems, Cloud Software and Services, and the Networks business area. The breadth is relevant to Ericsson’s commercial challenge, but it does not prove that his decisions alone produced reported segment outcomes.
  • He became CEO on 1 October 2026 through a Board appointment. The strongest early test is whether Ericsson can show customer payback—through durable deployments, renewals, service quality, operating savings and monetization—alongside supplier margins and cash generation.

The appointment starts the measurement period

On 1 October, Per Narvinger became President and CEO of Ericsson, a company whose products sit inside mobile networks used by billions of people and whose results depend on decisions made by service providers, enterprises, regulators and suppliers. The appointment creates a clean formal boundary. The Board chose him; the executive team supports him; and the day-to-day operating responsibility now rests with the CEO and the people accountable under him. That is a public authority map, not a claim about who privately influenced each decision.

It also creates a clean analytical boundary. The career before 1 October can explain why the Board regarded Narvinger as prepared. It cannot be treated as evidence that the CEO period has already succeeded. He has not yet produced a full quarter as chief executive. At the time this profile was commissioned, Ericsson’s latest published group report was Q2 2026, and the Q3 report was scheduled for 15 October. Q3 ended on 30 September, before the appointment took effect. Its results would be a useful inherited baseline, not a CEO-period score.

That distinction matters because succession stories often collapse three questions into one. What gave a person authority? What did the businesses under that person report? And what value did customers receive? The first has a straightforward public answer: the Board appointed Narvinger. The second has reported metrics and declared management targets. The third requires evidence that a company’s own financial statements do not automatically supply. Revenue, adjusted margin and free cash flow describe the supplier’s economics.

They do not, without customer-side evidence, establish lower network costs, a successful migration, improved availability, revenue uplift or a renewal that would not otherwise have happened.

The most useful way to read Narvinger’s rise is therefore not as a victory lap through a long résumé. It is as a map of the operating surfaces he has touched and the conversion tests that follow from them. Research and standardization shape what a vendor can build. Product management turns capability into an offer. Sales and customer-unit roles reveal the path from offer to contract. Network and software leadership expose delivery and operating economics. The CEO role makes him accountable for whether those pieces meet in a result customers can recognize.

Ericsson’s management page says the Board appoints the CEO and that the CEO manages day-to-day operations with support from the Executive Team. The appointment release names the effective date and the Board’s succession decision. Those sources establish responsibility. They do not tell us what the first year’s outcomes will be.

A career that crosses the sale, the product and the operating system

Ericsson’s public profile says Narvinger joined the company in 1997 and held roles across research and standardization, development, product management and customer-solutions sales. It records long assignments in Australia and Spain and a period leading the Northern and Central Europe customer unit, where Ericsson says about 2,000 employees served customers across 16 countries. The profile then traces his path through Product Area Networks, Cloud Software and Services, Networks and, now, the CEO office. These are not interchangeable jobs. Each sits at a different point where technical performance can become or fail to become customer value.

Research can produce a radio design or a software function without making it easy to deploy. Product management can package a feature without showing that a customer will pay for it. Sales can win a contract that later proves costly to implement. A customer unit can meet an order target while leaving the customer with a difficult upgrade path. A business area can raise its margin by changing its portfolio, even while a customer values a product for a different reason. The path through all those roles is a reason to ask more precise questions; it is not a substitute for answers.

His early customer-facing experience is particularly relevant because Ericsson is not an app business in which a buyer can test a product for an afternoon and switch the next day. A network operator plans capacity over years, coordinates radio, transport, core and operations systems, tests a change against service continuity, and negotiates across engineering, procurement, finance and regulation. The buyer’s economics are shaped by installed equipment, software support, energy use, spectrum, staff skills, rollout cadence, end-user demand and the cost of failure.

A supplier that sells a technically attractive feature still has to fit that system.

The public record does not name every account Narvinger handled or disclose customer-level outcomes from his customer-unit period. The right claim is narrower: he has worked in roles where the path from product to customer contract was part of the remit. That experience could help a chief executive identify when an internal success measure ends too early. An announcement, a lab demonstration, a first deployment and a repeatable commercial service are four different states. A performance claim needs to say which state it has reached.

This makes the career breadth useful as a test of operating coherence. Can the leader who understands development also insist on the supportability of the delivered product? Can a commercial organization price the lifecycle instead of just the initial order? Can software and network units share the evidence needed to prove an outcome without forcing customers to reconcile incompatible metrics? These are organization questions. The available documents do not show that Narvinger alone settled them in the past, and this profile does not infer that he did.

The network product role put technical promises on a commercial path

Before heading Cloud Software and Services, Narvinger led Product Area Networks. Ericsson describes that portfolio as spanning radio hardware and software across cellular generations, along with cloud-native RAN, indoor systems and transport. The public executive profile also records work in research and standardization, product management and sales. That mix puts him close to a common telecom tension: operators want better capacity, coverage and reliability, but they must decide whether a proposed architecture lowers the total cost of ownership enough to justify migration, integration and operational risk.

In a 2020 Ericsson podcast, Narvinger discussed the move toward virtual networks and Cloud RAN. The substance of that public technical discussion should be read as a position about a direction, not evidence that all deployments have achieved the proposed savings or service outcomes. The distinction between architectural possibility and operational value is particularly important when a technology depends on new cloud platforms, accelerators, software releases, integration skills or a redesigned operating model.

A credible vendor case has to explain not just what can run in a virtualized environment but who carries the cost of moving there and how the customer validates the gain.

That test is not hostile to innovation. It is what makes innovation commercially durable. If a new RAN design can place software across several hardware options, the customer needs to know how that choice affects energy use, performance, supply resilience, maintenance and upgrade compatibility. If an AI-enabled network promises faster optimization, a customer needs before-and-after measures and a record of when automation acts without operator intervention.

If an API exposes network capabilities, the customer needs evidence that developers adopt it, that the underlying network can honor the service promise, and that the additional revenue is recurring rather than a one-time pilot fee.

The seller has a parallel responsibility. A product may work in a test environment yet become expensive to deploy across a diverse installed base. A software release can reduce cost for the vendor while creating more integration work for the operator. A new platform can show better energy efficiency per bit in a controlled comparison but arrive with migration costs that postpone the saving for years. Those cases are not proof that a technology is bad. They demonstrate why cost, performance, deployment and retirement evidence must travel together.

Narvinger’s former role provides a practical lens for evaluating the CEO’s claims. Does Ericsson publish reference measures that let operators compare the old and new operating models? Does a product promise include release support and exit paths? Are service-level commitments traceable to a specific deployment or aggregated across a portfolio? Can a customer see where reported savings come from—energy, fewer sites, staff time, lower maintenance, or better spectrum use? Until those answers are measurable, “technology leadership” remains a strategic input, not the final account of value.

The cloud-services assignment was a turnaround with more than one yardstick

In May 2022 Ericsson announced a new Business Area Cloud Software and Services, combining Digital Services and Managed Services, and named Narvinger to lead it from 1 June. The company’s stated logic was to bring cloud, software and services together, increase cloud-native expertise, and build combined offerings for automation and AI-based service delivery. Ericsson also said that 16 of the 20 largest operators had chosen its cloud-native 5G Core portfolio.

That last statement describes a portfolio selection count; it does not tell us the scale of deployment, the revenue recognized, the customer’s savings, the length of support or the profit on each contract.

The new unit inherited different businesses with different economics. A license or software subscription may have a different revenue-recognition pattern from a managed-service contract. Managed services can involve people, operations centers, customer-specific delivery and local support commitments. Core network products may require lengthy validation and integration. Bringing them into one reporting unit can make strategic sense, but a unit-level margin cannot reveal which customer cohort improved or whether the combined offer reduced the cost of an operator’s network.

At its 2022 Capital Markets Day, Ericsson gave Cloud Software and Services a target: reach at least EBITA break-even in 2023. The early-2023 action plan included exiting some subscale agreements and product offerings after a portfolio and customer-contract review. Ericsson later reported adjusted EBITA of SEK 1.7 billion for the full year and SEK 2.0 billion in Q4, saying it had met the break-even target. The arithmetic matters. A target was set; a reporting unit produced a positive annual adjusted measure; and the company described commercial discipline, automation and delivery efficiency as continuing priorities.

That is meaningful evidence of a financial milestone, and it should not be dismissed. A business that loses money indefinitely cannot fund product support or development forever. Portfolio focus can make delivery more reliable. Exiting an uneconomic contract can prevent a supplier from subsidizing work that does not scale. At the same time, the exit can transfer switching, integration or continuity costs to a customer. Without account-level data, we cannot know which customers faced those tradeoffs or how they resolved them. The source supports the company’s announced portfolio actions, not a claim that every customer benefited.

The next year complicates any simple “turnaround complete” narrative. Ericsson’s 2024 report put Cloud Software and Services sales at SEK 62.6 billion, down 2% from SEK 63.6 billion in 2023, while its adjusted EBITA margin increased from 2.7% to 3.2%. A positive margin and lower sales can coexist. The figures show that the unit remained large and that adjusted profitability improved modestly while the company continued a turnaround.

They do not answer whether the business was growing in customer value, whether renewals improved, whether support costs fell for buyers, or whether customer economics were strong enough to justify adoption across more of the installed base.

Accounting measures also have boundaries. Adjusted EBITA excludes certain items; reported EBITA can move differently because restructuring, impairment and other costs occur outside the adjusted measure. Adjusted margin is useful for comparing operating performance on the company’s defined basis, but it is not the cash flow of an individual contract. Nor is it customer ROI. A business can improve its adjusted margin through cost reduction, mix, pricing or exiting work. Some of those changes may reflect better execution; others may narrow the portfolio or shift obligations. A careful profile keeps the measure and its interpretation separate.

Ericsson’s 2022 structure announcement, 2022 strategy presentation, 2023 results and 2024 annual report document the unit’s setup, target and subsequent numbers. Together they establish that Narvinger was in charge during the break-even milestone. They do not isolate the result to him or supply the missing customer-level evidence.

The transfer to Networks changes the denominator

On 15 March 2025, Narvinger moved from Cloud Software and Services to lead Networks, while Jenny Lindqvist took over the software unit. The transfer matters to how the later results should be read. Ericsson’s 2025 report says Cloud Software and Services sales were SEK 62.7 billion and its adjusted EBITA margin was 11.4%, up from 3.2% in 2024. That increase is a striking segment result, but most of the 2025 period came after Narvinger had left the role. It is evidence about the unit’s trajectory after the leadership handoff, not a fair personal credit claim for his tenure.

The role transfer is also a natural continuity test. A strategy should leave behind systems that the next leader can operate, measure and improve. If the 2023 break-even achievement depended on a repeatable process, later results should reveal which parts carried forward and which changed. If it depended primarily on exiting work or one-off cost actions, the customer base, product roadmap and renewal data should eventually show whether the new economics remained healthy. Public segment reporting alone does not provide that decomposition, and this article does not claim to know it.

Networks gives the CEO a different operating surface. Ericsson reported 2025 Networks sales of SEK 151.0 billion, down 5% on a reported basis from SEK 158.2 billion, and adjusted EBITA of SEK 31.2 billion with a 20.7% margin, compared with SEK 27.8 billion and 17.5% in 2024. The company attributed its performance to a mix of cost-reduction actions, operating efficiency, currency and market conditions. Organic sales grew 1% for the full year even as reported sales fell, a useful reminder that exchange rates can change the picture a shareholder sees in kronor.

Narvinger led Networks for most of 2025, but the reporting year includes a short period under his predecessor and many forces outside any one executive’s control. A quarter-by-quarter reading is more informative than assigning the annual result to one leader. Networks adjusted EBITA rose from SEK 5.252 billion in Q2 2024 to SEK 6.506 billion in Q2 2025 and SEK 10.058 billion in Q4 2025, while the same report records changes in sales, currency, product mix, gross margin and expenses. Those numbers describe a business that improved profitability in a market that was still uneven.

They do not isolate how much came from decisions made by Narvinger, how much from earlier investments, or how much from cyclical and accounting factors.

The adjustment from a falling reported revenue line to a rising margin line is strategically important. If the company earns more from fewer reported sales, that can signal discipline and a stronger mix. It can also mean that customers are investing less, that the portfolio is concentrated in higher-margin contracts, or that suppliers are preserving price while volumes change. The answer requires evidence about what the buyer received and what it paid. For a network operator, a margin can improve even while deployment is deferred.

For Ericsson, a customer might extend a contract because the technology is valuable, because migration is costly, or because no alternative can meet its requirements. Those explanations have different implications for both parties.

The next test for the CEO is to connect his Networks experience to the group’s wider customer proposition without recasting every prior result as a personal achievement. A useful report would show deployment lead times, support response, energy use, upgrade effort and customer renewal by product cohort, alongside the supplier’s margin. It would explain whether the improvement arose from lower operating costs, better utilization, new features, pricing or a changed project mix. It would also show what happened when a project was delayed or a contract ended. These details convert an internal score into a customer-return account.

The 2025 financial report is therefore a useful baseline with a precise attribution boundary: Narvinger led Networks for most of the year; the segment reported a higher adjusted margin and lower reported sales; currency and operational factors shaped the result; and the report does not identify a personal causal contribution.

The CEO profile now includes the Enterprise segment

Ericsson’s current executive profile lists Narvinger as both President and CEO and Head of Segment Enterprise. His leadership remit therefore touches more than the Networks and Cloud Software and Services businesses that frame his earlier path. The company describes Enterprise as including enterprise wireless solutions, private 5G, indoor coverage, wireless WAN, and network APIs, alongside the Global Communications Platform. The enterprise story joins physical connectivity to software and communications platforms, but each offer has a different customer, implementation path and revenue model.

That breadth can help the company bring products together. It can also make aggregate reporting harder to interpret. “Enterprise” can include subscriptions, hardware, integration, software platforms and portfolio changes. A strong headline margin can be influenced by a transaction, a divestment or a one-off gain. In 2025 Ericsson reported a SEK 7.6 billion capital gain from the divestment of iconectiv; the annual report’s Enterprise segment result must be read with that footnote. Treating a reported segment margin as an ordinary recurring operating return would mislead readers.

For customers, the question is not whether a portfolio crosses business-area boundaries on an organization chart. It is whether the handoffs work. A private-5G buyer may need coverage planning, radio equipment, a managed wireless platform, security integration and support. A developer considering a network API needs predictable access, documentation, service levels, a usable billing relationship and evidence that the network capability is available where users need it. A communications-platform customer may care about developer experience and delivery reliability rather than Ericsson’s radio portfolio.

The group needs a clear owner for each promise.

Network APIs illustrate the gap between a plausible market story and measured returns. Ericsson has publicly described an ambition to expose network capabilities for developers. That may create new revenue and allow applications to request quality or location-related capabilities under defined conditions. To show that the strategy works, the company would need to separate live production usage from pilots, count paying developers and active end-user transactions, disclose renewal and concentration, and explain how value is shared among the operator, Ericsson and the application provider.

A single launch, partnership announcement or test does not establish a recurring business.

The same evidence discipline applies to AI-enabled networks. A demonstration can show that existing towers or radio signals support a new sensing or optimization capability. It cannot by itself show that operators have adopted it, that the result is reliable in ordinary conditions, or that customers receive more revenue or lower costs than the system requires. Where the commercial claim depends on future deployment, the article should call it an opportunity rather than a realized return.

The Enterprise role makes a cross-portfolio scorecard especially useful. It should distinguish hardware sales from recurring software and service revenue; booked orders from deployed capacity; platform activity from paying usage; and a divestment gain from operating earnings. It should explain how a private-network customer can move between support levels, what happens when a product is discontinued, and who remains responsible when a service crosses corporate units. The point is not to demand one accounting metric for every business. The point is to keep unlike metrics from being presented as if they proved the same outcome.

A customer-return scorecard starts after the contract is signed

Supplier reporting usually begins with recognized revenue, cost of sales, operating expenses and cash. Customer value begins earlier and continues later. The customer chooses a network investment because it expects a specific change: more capacity, wider coverage, fewer faults, lower energy bills, faster service activation, new wholesale revenue or a stronger product for its own subscribers. The provider may not disclose every customer’s commercial data, but it can still report aggregated, independently defined measures that make its claims testable.

The first dimension is deployment. What share of announced orders reaches production within the planned window? How long does the change take from signed contract to a stable operational service? How many integration steps are required and which are customer-specific? If an order includes multiple countries, sites or software components, do the reported progress figures count each separately? A rising backlog is not a substitute for a deployment series. A completed lab trial is not the same as service carrying real traffic.

The second dimension is operational outcome. Reliability should be tied to a defined scope, measurement interval and comparator. A promise of fewer faults needs the rate per relevant unit or customer population. A performance claim should distinguish peak lab capacity from the experience operators obtain under traffic, interference, load and maintenance. Energy efficiency requires a boundary: per site, per bit, per covered household, or per period. A reduction at the radio layer may shift energy use elsewhere in the system. A customer should be able to understand what the reported measure includes.

The third dimension is economics over the lifecycle. An initial order is only one part of the total cost of ownership. Hardware, licenses, integration, training, energy, support, security updates, upgrades and eventual migration all affect the return. The relevant horizon depends on the equipment and contract. A product that lowers annual operating cost but takes years to deploy may produce a weaker near-term return than the headline model suggests. Conversely, a higher initial cost can be economic if it reduces outages or improves capacity for long enough.

Published comparisons should state the assumptions and avoid portraying a forecast as realized savings.

The fourth dimension is durability. Renewal rate alone can mislead if customers renew because migration is difficult or if small accounts exit while a few large accounts remain. A credible story would pair renewals with customer counts, contract value bands, product use, support burden and stated reasons for retention where those can be reported without disclosing confidential accounts. Churn, end-of-support events and migration assistance should be visible alongside new orders.

Software lifecycle management becomes commercially fair when the customer knows how long a release is supported, how an upgrade is priced and whether data or configurations can be exported at the end.

The fifth dimension is monetization. For a vendor, recurring revenue and cash conversion show whether demand persists. For the customer, the equivalent question is whether a capability increases revenue or reduces cost. Network APIs, service orchestration and enterprise connectivity need a chain from capability to actual usage and from usage to money. Count active production customers; separate pilot and live services; report average usage and renewal, while explaining the role of the operator or third party. A forecast about the size of a market does not prove a product is capturing it.

Finally, an accountability report should show both sides of a transition. When a provider exits a subscale product or contract, what notice, support and migration assistance are offered? What proportion of customers moves to another Ericsson offer, selects another vendor, or retires the service? This is not an argument against pruning. Portfolio decisions can protect long-term delivery quality. It is a way to make the cost of the decision visible to the people who depend on the service.

No one metric can settle all of these questions. The usefulness comes from a small, stable bundle of measures that can be followed over time and compared with a baseline. If Ericsson publishes only supplier margin, the reader can judge supplier efficiency but not customer payback. If it publishes only engineering benchmarks, the reader can judge technical capability but not commercial adoption. If it publishes bookings without deployment, the reader can judge sales momentum but not execution. The new CEO has experience across each of these surfaces.

A coherent customer-return account would let investors and customers see whether those surfaces are connected.

Authority, continuity and the limits of a biography

Heng Lu’s Note 73 argues that attendance or visible participation does not itself create mandate; authority must be established through a documented decision. Applied narrowly here, that principle points to a simple fact: Narvinger’s formal authority comes from the Ericsson Board’s appointment, not from his length of service, his technical credentials or a succession narrative. His résumé can explain why the Board selected him. The public appointment record establishes that it did.

The analogy should stop there. A corporation has legal bodies, shareholders, executives and commercial counterparties; its authority structure is not the same as a multi-stakeholder internet institution. Note 73 is a way to keep the evidence categories clear, not a source about Ericsson. This profile therefore does not turn Heng Lu’s governance claims into claims about Swedish corporate law, nor use the note to validate the Board’s decision.

Note 70’s continuity lens is also useful if kept factual and modest. Telecom service must continue while leaders change. Operators do not replace a nationwide system when one executive leaves; they need service levels, release schedules, operating knowledge, support capacity and escalation paths that outlast the person currently in the CEO office. Continuity is an organizational capability. It is not proof that a leader has authority over every customer system, nor a reason to attribute every historical outcome to one executive.

Ericsson’s appointment release says Börje Ekholm would act as executive adviser to the new CEO until 15 June 2027. The statement establishes an announced advisory period; it does not disclose the private scope of that role. For investors and customers, the practical question is how responsibilities are recorded and how decisions are escalated while a departing CEO remains available. A clear arrangement can preserve institutional memory without obscuring who now has decision rights. Public evidence should be allowed to answer only what the company has said.

A biography has a natural temptation to make every role part of a single inevitable ascent. The evidence supports a more useful account. Narvinger has held positions across technology, product and commercial operations. He led a software and services unit that met a break-even target. He later led Networks during a year when the reported margin rose and sales fell. A different leader then oversaw most of the following year’s software margin expansion. He now has a formal CEO mandate and a listed Enterprise remit. The facts are interesting because they expose boundaries and tests, not because they remove uncertainty.

The inherited baseline and the first full-year questions

Ericsson’s Q2 2026 report gives the group’s latest published pre-CEO baseline at this writing: reported sales of SEK 52.7 billion, down 6% year on year; organic sales down 1%; adjusted gross margin of 48.4%; adjusted EBITA of SEK 6.9 billion, or 13.1%; and free cash flow before M&A of SEK 0.4 billion. The company said the sales comparison was affected by a partial IPR settlement in the prior year. The reported decline and nearly flat organic result are not contradictory: they use different comparison bases. Nor does the adjusted margin show whether an operator’s network costs fell.

The first CEO quarter is an awkward measurement unit because the company’s Q3 reporting period ended the day before his start. Even when those results publish on 15 October, the accounting period belongs to the organization Narvinger led as Head of Networks and to the outgoing CEO’s executive period. It can inform an assessment of the inherited group, but a fair report would label the date boundary and refrain from calling the quarter an early CEO success or failure. The next quarter will still be short. One year and several customer deployment cycles will be more informative than a single headline.

The full-year test is not simply whether Ericsson grows. Demand for radio equipment is cyclical. Operators can stretch replacement intervals when capital budgets tighten; government decisions can alter deployment timing; currencies can move reported results; and licensing agreements can create lumpy comparisons. A capable CEO cannot eliminate those forces. He can make the company’s exposure and response legible: identify which offers have durable demand, which contracts are profitable after support obligations, which roadmap investments customers will co-fund, and which capabilities should be maintained even when a market pauses.

Three questions will be practical for readers to follow. First, do Networks and Cloud Software and Services convert technical advances into adopted customer capabilities? Second, does Ericsson maintain or improve margins through efficiency and mix while leaving enough investment for security, support and the next product cycle? Third, can Enterprise demonstrate repeat usage and returns without relying on transaction gains or market-size forecasts? The answers should be measured over cohorts and time horizons rather than inferred from one presentation.

There is also a governance question. Does the Board ask management to report realized customer outcomes alongside adjusted financial results? Are milestones set in terms that customers can observe, or only in terms that the supplier controls? How are service exits, extended support and migration obligations considered when compensation and portfolio decisions are made? These are reasonable questions for shareholders to ask. The public materials cited here do not answer them, so this profile frames them as future oversight tests rather than as allegations.

A demanding but fair test

Narvinger’s path gives him unusual context for the work: he has been close to research, products, sales, service delivery and network systems. It does not grant him a free pass on customer outcomes. The 2023 break-even result deserves recognition as an achieved unit target; the later improvement in Cloud Software and Services belongs primarily to the period after his transfer; and the 2025 Networks result deserves attention with the market, currency and cost context intact. Each result is more informative when its ownership and limits are stated clearly.

The CEO job now begins with authority that the Board can document and performance that customers will experience through products, releases, service and support. The most credible record would link those levels: a product claim, a deployment result, a customer-side measure, a renewal or usage pattern, and a supplier return that persists after support and migration costs are counted. Where confidentiality limits detail, Ericsson can still publish consistent, aggregated cohort measures with an explanation of methodology.

If customer outcomes strengthen while the supplier’s economics remain durable, Narvinger will be able to point to more than a long career and a well-timed appointment. If the company’s margin improves but customer deployment, usage or renewal does not, the financial result may still be real, but it will answer a narrower question. That is the standard this profile sets: a career is evidence of exposure to the problem; only measured delivery can show that Ericsson has solved it.

Sources