- Analysts, assemble! Ericsson is holding its annual Enterprise Industry Analyst Day tomorrow at the Boston Public Library
- Network APIs have been early, overhyped and commercially disappointing—but they will become the DNA of the global digital economy
- Ericsson is missing today’s optical-networking boom, but may be better positioned for the physical AI economy that follows it
Tomorrow, Ericsson (Nasdaq: ERIC) hosts its annual Enterprise Industry Analyst Day at the Boston Public Library. (Shhhhh!)
This being an Ericsson event, none of the speakers will be late.
More importantly, the strategy that it will share is early. Years early. This is both a boon and a bane for Ericsson.
It has correctly identified where value is heading in the AI economy. But its vision is so far ahead that Ericsson risks looking like an outlier, isolated from its peers.
Those optics are magnified by optical networking and the extraordinary financial bounty being enjoyed by companies selling optical technology into the AI-factory and data centre madhouse, including Ciena (NYSE: CIEN), Cisco (Nasdaq: CSCO), Nokia (NYSE: NOK) and Huawei.
Ericsson sells optical-fronthaul products for mobile transport, but it does not offer a broad hyperscale optical-networking and data-center-interconnect portfolio comparable to those companies. It therefore does not participate meaningfully in the greatest optical bull market since 2000.
However, that disadvantage may become its secret weapon. When the bubble surrounding centralized LLM data centers deflates—and North America realizes it has mistaken the AI factory for the AI economy—Ericsson may be better insulated from the fallout than today’s winners.
The API economy is coming
There is a quite reasonable argument that network APIs are another telecom solution in search of a problem. Progress has been slow, revenues modest, and most of today's applications still treat connectivity as an invisible utility.
But dismissing network APIs because they have not taken off yet is like dismissing the commercial internet in 1993 because nobody had worked out how to make money from a website.
The API layer will be essential. Its time has not yet arrived. Yet. But it is coming.
For it is written. And so shall it be, Inshallah.
Anyone who does not understand this is guilty of viewing networking as a 20th-century phenomenon defined by bandwidth, coverage and latency. Those things will remain important, but they will no longer be sufficient.
The networks now emerging will be utterly different: ubiquitous in reach, extending beyond the traditional carrier edge and deep into enterprises and industry; embedded in almost everything; essential to the functioning of the physical economy; and, above all, mind-blowingly complex.
Consider Tianjin Port, which I visited in China. One network may simultaneously connect autonomous cranes, cargo sensors, customs systems, security cameras, ships, drones, trucks, workers and the local power grid. Each will have a different identity, owner, application, priority and level of authorization. A crane may require guaranteed performance; a customs agent access to specific cargo data; a drone permission to cross an operational zone; an ambulance immediate priority over everything else. Thousands of such decisions must be made securely, across public and private infrastructure, in real time.
The network’s job will no longer be merely to carry the traffic. It must determine who—or what—is asking, what it is permitted to do, under which conditions and for how long.
Now extrapolate that across the world and everything that moves or breathes upon it.
So, yes, complicated at a level never before seen in the history of human advancement.
The API layer will be the gateway through which autonomous applications request access to critical infrastructure. It will not be the control plane itself. It will be the strategic control point—the front door to multiple control planes.
And the $64 trillion question is: who controls it?
Ericsson is living in the future
This is the importance of Aduna, the network API company created by Ericsson and major international operators.
Aduna gives developers one route to capabilities across multiple carrier networks by aggregating CAMARA-aligned APIs behind a common interface. The organisation says it simplifies access for hyperscalers, aggregators, enterprises and developers.
Initial applications are dull: number verification, fraud detection and connectivity on demand.
The significance becomes clearer when machines and AI agents begin configuring infrastructure. An autonomous vehicle will not call an account manager for low-latency connectivity. A factory robot will not negotiate a contract before changing networks.
They will use APIs.
Ericsson believes the great prize is not another incremental improvement in radio performance; it is control of the gateway through which the autonomous economy reaches the network.
Private 5G meets the physical economy
Ericsson’s position as a 5G incumbent—and shortly a 6G incumbent—is a fundamental advantage.
Demand for network APIs will not be driven primarily by people watching better video on phones. It will be driven by private 5G and eventually private 6G connecting AI to machines.
A production line cannot hesitate: robotic arms, machine-vision systems and safety controls must respond instantly or production stops—and people may be injured. A power grid must authenticate thousands of devices while detecting and isolating a compromised system before it triggers a blackout. A hospital must know, with absolute certainty, whether an application or AI agent is authorzed to control a medical device upon which a human life depends.
These are questions of identity, permission, policy, resilience and control. They are also questions of operational sovereignty: whether an enterprise or nation retains meaningful control over the autonomous systems on which it depends.
The API becomes the mechanism through which those decisions are requested. Whoever controls that interface sits between the AI agent and the physical infrastructure it wants to command.
Today’s winners are riding another wave
Ericsson is performing considerably worse at the top line than several networking peers. Its business remains constrained by subdued carrier spending.
Ciena, Cisco and Nokia sell the infrastructure connecting GPUs, AI factories and hyperscale data centers. Ericsson is weighted toward mobile networks and carrier investment.
That makes Ericsson’s life harder today. It may also leave the company less exposed when this particular bubble bursts.
AI demand will not disappear, but today’s spending on giant training clusters cannot continue indefinitely. Infrastructure cycles overshoot. Sometimes they also miss the point. The optical boom at the turn of the century showed that transformative technologies can attract too much capital, too quickly, into one part of the architecture.
AI factories are one element in the means of production. They are not the finished economy.
The larger prize emerges when intelligence enters vehicles, factories, grids, healthcare and robotics—requiring secure, programmable connectivity among billions of physical systems.
Everyone wants the gateway
Ericsson does not have the field to itself. Powerful groups are approaching the same control point from different layers.
The table reveals the real contest. Ericsson is trying to aggregate carrier capabilities and expose them through one gateway. But carriers could build that layer themselves. Lumen Technologies (NYSE: LUMN), for example, is placing a programmable interface above its own infrastructure. “Lumen is putting innovation back where it belongs—inside the network itself,” CEO Kate Johnson told me recently.
The hyperscalers pose a different threat. Amazon (Nasdaq: AMZN), Microsoft (Nasdaq: MSFT) and Alphabet (Nasdaq: GOOGL) could absorb network APIs into the cloud platforms where developers already work. Security companies could make identity and authorization—not connectivity—the primary gateway. Cisco (Nasdaq: CSCO), meanwhile, spans networking, silicon, security and observability, giving it several possible routes to the same destination.
Ericsson nevertheless has leverage. The hyperscalers control compute and software, but they cannot reach factories, vehicles and machines without carrier networks. Ericsson sits between those two worlds.
The danger is that carriers bypass it, hyperscalers subordinate it or another platform seizes the gateway first. Ericsson has identified the control point. It has not yet secured it.
Aduna needs industries
For now, Aduna remains primarily a partnership among Ericsson, carriers and distribution platforms. It lacks meaningful participation from vertical industries—and that must change.
The most important APIs will emerge from robotic production lines, mines, grids, hospitals and transportation systems, not telecom product departments. Manufacturers, energy companies, transport operators, healthcare providers and others must therefore help shape Aduna’s capabilities and guarantees, rather than being treated as consumers at the far end of the platform.
Otherwise, Aduna risks exposing what carriers know how to sell rather than what enterprises need to buy—creating a vacuum that competitors will happily fill. Success could lift operators above basic connectivity. Failure would leave them providing the roads while somebody else owns the traffic lights and tollbooths.
Persuading the market to arrive
Living in the future does not pay particularly well when everyone else is spending money on the present.
But Ericsson’s financial weakness and strategic position arise from the same fact: it has less exposure to the centralized AI-infrastructure boom and greater dependence on a mobile, industrial future that has not yet arrived.
The API economy is painfully early. That does not make it optional. Bandwidth and latency describe how data moves. APIs determine what the network may do, for whom, under what conditions and at what price.
One executive who will not be in Boston is Ericsson’s incoming CEO, Per Narvinger. Unlike his Americanophile predecessor, Börje Ekholm, Narvinger is an Ericsson engineer formed almost entirely within Sweden’s telecom-industrial establishment: understated, technically literate and low on CEO theatre.
Having worked across standards, radio networks, cloud software and sales, he understands Ericsson’s machinery in microscopic detail—an advantage as the company attempts to unite private 5G, 6G, AI and APIs into one proposition.
“As AI continues to industrialize, this will increasingly require advanced connectivity solutions,” he said when Ericsson announced his appointment in June. The question is whether he can turn that engineering inheritance into a movement. Ericsson needs more than someone who understands the future. It needs someone who can persuade the market to arrive there.
I expect to interview Narvinger before the end of the year.
Ericsson is already living in that future. Narvinger’s task is to drag the rest of the world into it.
Stephen M. Saunders MBE is a communications analyst and USPTO-registered inventor examining how digital infrastructure — 5G, cloud and AI — is reshaping industry, power and society, as well as underpinning the emerging, ubiquitous global digital economy. As anchor of FNTV and a longtime industry insider, he focuses less on growth narratives and more on execution, risk and how hyperscale technology is distorting markets, governance and society at scale.
Opinion pieces from industry experts, analysts or our editorial staff do not represent the opinions of Fierce Network.