The idea that telecom networks can be used for weather sensing is not new. GNSS-based atmospheric estimation, rain attenuation on microwave links, and radio refractivity effects have been studied for more than 20 years. The physics is well understood and already embedded in network planning and synchronization systems.

Starlink’s decision also comes after the company disclosed in December that one of its satellites suffered an in-orbit anomaly, producing a “small” amount of debris and cutting off communications with the spacecraft at an altitude of about 260 miles. The company said the satellite, one of nearly 10,000 currently in orbit for its broadband internet network, rapidly lost 2.49 miles in altitude, suggesting an internal explosion, in what was described as a rare kinetic accident.

La faillite de Kodak est souvent présentée comme l’illustration parfaite de la théorie de la disruption popularisée par Clayton Christensen. L’étude d'Albéric Tellier montre d’abord que Kodak n’a ni ignoré ni sous-estimé le numérique. Dès le début des années 2000, sous la direction de Daniel Carp, l’entreprise consacre près des deux tiers de sa R&D aux technologies numériques et élabore en 2003 un plan stratégique ambitieux prévoyant trois milliards de dollars d’investissements. C’est précisément ce plan qui se heurte à une opposition frontale des actionnaires.La chute de Kodak ne s’explique pas uniquement par une incapacité managériale face à la rupture technologique, mais par l’impossibilité de déployer une stratégie de rupture dans un cadre de gouvernance dominé par des investisseurs institutionnels orientés vers le court terme.

The killer app is one of the most persistent myths in technology. It promises salvation through brilliance. One product so compelling that customers flood in, competitors fall behind, and monetization somehow takes care of itself. History shows the opposite: killer apps are excellent at creating demand but poor at protecting value.The iPhone is the cleanest way to expose the myth, precisely because it is so often misused as proof. People talk about it as if the device itself were the business, and it never was. A smartphone is a standardized object the moment it succeeds: glass, battery, radios, silicon, and software. Once adoption explodes, expectations harden, and competitors converge. That is what killer products do: they define the reference design that others copy.What prevented the iPhone from becoming a commodity phone business was not the phone itself. It was everything built around it. App distribution, payments, identity, accessories, defaults, services. Remove those layers, and Apple would have faced the same fate as every other hardware success story. Indispensable product, shrinking margins.

The more realistic role of 6G is to enable an AI-driven economy without collapsing network economics. AI traffic is not primarily about terabytes but about transactions, latency variance, trust, and coordination. Supporting that efficiently requires lower control density, tighter timing, integrated compute placement, and autonomous operation. Those capabilities protect margins first and create optionality second.Standards strategy reflects this shift. The emphasis is moving from feature accumulation to architectural coherence, AI-native operation, and system-level efficiency. Success will be measured less by headline speeds and more by reduced cost per delivered service, lower energy per transaction, and fewer human interventions per network element.6G therefore carries a narrower but more complex value proposition. It does not promise revenue miracles, but it aims to stop economic deterioration while preparing networks for continuous machine interaction and AI-driven workloads. The ROI case is defensive and makes it less exciting, but far more grounded than the narratives that accompanied 5G.

L’idée, si on résume, c’est un peu de traiter le soleil comme un interrupteur. Il fait nuit, mais seulement sur une partie de la Terre. Le soleil ne s’arrête jamais de briller donc ses rayons sont toujours là. C’est une source d’énergie quasiment inépuisable pour encore cinq milliards d’années en tout cas, ce qui laisse le temps de voir venir. Les deux jeunes fondateurs de Reflect Orbital pensent qu’il ne faut pas forcément améliorer le rendement des panneaux solaires ou concevoir des batteries pour stocker l’énergie. Il suffit "juste", selon eux, d’allonger les journées.

From pipes to machine governance: how 5G and 6G turn the telco stack into deterministic coordination, trusted identity, and real time control for billions of robots

Revolut has not entered mobile communications to outperform network operators on coverage or spectrum efficiency. The company has entered the market because mobile connectivity has become another digital capability that fits naturally within a platform already handling identity, payments, subscriptions, rewards, travel, and daily financial behavior. When connectivity is viewed as software rather than infrastructure by the customer, the economics shift upstream.

On Dec 19, 2025, the White House issued a national security presidential memorandum ordering federal users to begin relocating from 7.125 to 7.4 GHz to free the band for full-power, licensed commercial 6G. Agencies have 12 months to submit relocation plans with costs and timelines, while protecting national security missions and critical infrastructure operations. The memorandum also triggers immediate studies on 2.69 to 2.9 GHz and 4.4 to 4.94 GHz for potential additional full-power 6G capacity. It directs the State Department to advance US 6G positions through international engagement ahead of the next set of standards and spectrum cycles.

According to new data from Eurostat in a press release from the European Environmental Bureau, the amount of electronic equipment sold in Europe in 2023 skyrocketed to almost 16 million tons, an 89% increase over the amount sold in 2012. On top of that, 5.7 million tons of e-waste were collected in 2023, and collection rates of e-waste left much to be desired. 

At the Fangchenggang project, Walker S2 units will help border staff guide passenger queues, direct vehicles, and answer simple questions from travelers. Some robots will patrol corridors and waiting areas, watching for blocked exits or crowd patterns that might require human officers to intervene. Others will move between cargo lanes to support logistics teams, checking container IDs, confirming seals, and relaying status updates to dispatch centers. Away from the border itself, the fleet is expected to inspect steel, copper, and aluminum facilities, walking structured routes through hot industrial yards.

The robot in question ran off an internal AI agent. By exploiting a flaw in the software, the researchers were able to take over the robot while it was connected to a network, at which point the researchers had the robot use local wireless communication to spread the hack to another nearby robot that actually wasn't even connected to the network at the time. Spreading the hack from one robot to another only took a matter of minutes. Even worse, the researchers were able to issue a command for the robot to physically strike a mannequin on stage.

Telecom is rarely associated with startup innovation. Software, fintech, and consumer AI dominate venture narratives, while telecommunications is viewed as slow, regulated, and capital-heavy. Despite structural barriers, a non-trivial startup ecosystem has formed around telecom. Analysts identify roughly 7,000 active telecom startups globally, spanning RAN software, AI-driven operations, security, satellite systems, private networks, and connectivity platforms. This is small relative to SaaS or fintech, but large enough to bring disruption.The backdrop for this new wave of telecom startups is a mix of necessity and disruption. Operator revenues are flat, while complexity continues to rise across multi-generation networks.

As of today, GPUs remain appropriate for co-located edge inference where workloads are heterogeneous, and utilization can be amortized. In a commoditized, energy-constrained RAN, inline radio inference follows a different economic and physical logic. Architectures optimized for deterministic execution, integration, and cost per decision outperform throughput-oriented designs. As learned functions move deeper into the radio processing chain, hardware selection will increasingly be driven by these constraints rather than by data-center inference defaults.

In most industries, heavy usage correlates with high value. Airlines reward frequent flyers. Banks price transaction volume. Cloud providers charge per compute cycle. Telecom does the opposite.Mobile data usage follows a power law. Operator and regulator datasets consistently indicate that approximately the top 10% of users account for 40–60% of total traffic, particularly during peak hours. Under flat or unlimited pricing, users pay approximately the average price rather than a premium.This outcome is not really accidental. Usage-based pricing increases bill shock, which is a major churn trigger. To stabilize retention, operators removed usage signals entirely, accepting margin dilution in exchange for lower churn risk.In telecommunications, the customer who relies most on the network is often the one the network least wants to serve.

96 $, c’est ce que coûte aujourd’hui la minute de calcul quantique chez un grand cloud provider. Le quantique n’est plus un fantasme de laboratoire : les machines sont disponibles, les tarifs clairs, les capacités démontrées. En d’autres termes, il suffit à un acteur malveillant, étatique ou criminel, d’estimer ce qu’il lui en coûterait de déchiffrer des données en sa possession et le bénéfice qu’il pourrait en tirer.

In 1982, Bruce Springsteen recorded a set of stark, one-take demos in his home studio. His label executives expected him to polish them in a professional studio with the E Street Band. Every instinct in the music industry said the same thing: take the raw material and improve it. Springsteen refused. He released the unrefined version as Nebraska, a haunting, minimalist album that critics later hailed as one of the most influential of the decade. Its impact came not from production quality but from preserving what made it authentic and different. Inside most organizations, that kind of protection rarely happens. Employees with their own “Nebraska moments” often see their boldest ideas diluted through layers of approval until the originality is gone.

Telcos did not misunderstand the internet; instead, for many years, they treated it as an add-on rather than as a new organizing principle. They optimized existing processes rather than redesigning decision-making, pricing, and product creation to substantially reduce coordination costs.AI sits in the same category today. It reduces the cost of prediction, pattern recognition, and decision making by orders of magnitude. According to recent studies by McKinsey and Analysys Mason, more than 70% of telco AI spending between 2022 and 2024 has gone into customer care automation and network operations, with primary success metrics tied to opex reductions of 5-15%. Fewer than 20% of initiatives target pricing, revenue optimization, or new product creation, and even fewer have the authority to execute changes autonomously. AI is treated as an efficiency layer rather than an operating model.

L’essor de DeepSeek redonne de l’élan à l’IA chinoise. Mais entre retard logiciel, manque d’ouverture et faiblesse du soft power, Pékin doit encore relever plusieurs défis pour s’imposer à l’échelle mondiale.

Airspace activity has crossed the threshold where ignoring it becomes a strategic mistake. In 2024, commercial aviation operated approximately 40 million flights worldwide, carrying nearly 9 billion passengers. At the same time, tens of millions of drones are already active worldwide, with commercial fleets scaling at double-digit growth. Urban air mobility programs and military aviation introduce an additional layer of permanent airborne traffic. These systems require deterministic, high-availability connectivity. This could be a new trillion-dollar control-and-connectivity opportunity for Telcos.