🌐 AI & Telecom Trends – Week of November 10, 2025
From margins to space: how artificial intelligence is rewriting the DNA of telecommunications.
This week something that’s happening quietly but constantly catches my attention: AI is stopping being a tool we use to become the nervous system of everything we do. It’s not just that there are more models or more power — it’s that artificial intelligence is embedding itself in the structures that sustain our daily life.
Each week I share key developments that are shaping the intersection of artificial intelligence and telecommunications. The goal: understand what’s changing, how it will affect real projects, and where new opportunities are emerging.
This week’s signals show five movements that, together, paint a clear picture: from operational margins to satellite orbit and digital security, all signals converge in the same direction: intelligence as infrastructure.
1️⃣ AI to the rescue of telco margins
What happened:
The global telecommunications market (including pay TV) will reach US $1.53 trillion in 2025, with annual growth of 1.7%. However, margins continue to fall due to competition, declining ARPU, and rising operational costs.
AI is consolidating as the key tool to sustain profitability and reduce OPEX. According to IDC, telcos that integrate AI into operations achieve efficiency improvements of 8 to 20% and cost reductions of up to 15%.
Why it matters:
Growing focus on AI for predictive maintenance, support automation, and network optimization. Telcos are redesigning entire flows to measure ROI in savings, not just innovation. Operational efficiency becomes a real competitive advantage.
Concrete example:
A European operator uses AI to detect network micro-failures before they affect customers, reducing repair time by 30% and support costs by 15% in the first year. The system analyzes more than 2 million daily network events and predicts failures with 87% accuracy.
Broader context:
Margin pressure is structural. Price competition and rising operational costs force telcos to seek radical efficiencies. AI is not optional: it’s the only tool that can maintain profitability while investing in new capabilities.
Sources:
- “IDC Report: Telecom Operators Turn to AI to Boost EBITDA Margins” — techblog.comsoc.org, 2025
2️⃣ Artificial intelligence in satellite networks: new playing field
What happened:
The expansion of LEO and MEO constellations is taking network complexity beyond the atmosphere. AI and machine learning enable managing thousands of satellites that constantly change position, optimizing resources and anticipating congestion in real time.
Currently there are more than 8,000 active satellites in orbit, and it’s projected that by 2030 there will be more than 50,000. Managing this complexity without AI is impossible: each satellite generates terabytes of telemetry data, and routing decisions must be made in milliseconds.
Why it matters:
The hybrid “ground + space” architecture demands intelligent orchestration. AI enables dynamically prioritizing traffic according to demand, weather conditions, or orbital congestion. Opens “connectivity everywhere” business opportunities — key for remote or maritime regions.
Concrete example:
A global operator employs AI to automatically reassign bandwidth on cruise and flight routes, reducing latency by up to 40% and maintaining stable service quality. The system processes data from more than 500 satellites in real time and automatically adjusts resources.
Broader context:
By 2030, more than 30% of internet traffic is expected to pass through satellites. AI is essential: without intelligent algorithms, it’s impossible to manage thousands of satellites moving at 27,000 km/h.
Sources:
- “AI and ML Powering the Next Leap in Satellite Networks” — Telecoms.com, 2025
3️⃣ “Cloud-native” architectures and integrated AI: the new foundation of networks
What happened:
The future of networks lies in software. Migrating to cloud-native architectures — containers, microservices, and open APIs — allows AI to not be a patch, but a structural part of the system.
BCG estimates that migrating a telco core to cloud-native can reduce TCO (total cost of ownership) by up to 60% over five years, by simplifying maintenance, scaling automations, and reducing downtime. More than 70% of global telcos have active cloud-native migration plans, and it’s expected that by 2027 more than 50% of core networks will be fully virtualized.
Why it matters:
5G Advanced and 6G networks are designed from scratch “for AI”. Observability and automation will be as or more important than transport capacity. Telcos move from managing equipment to managing models.
Concrete example:
A European telco virtualizes its core, uses AI for failure self-supervision, and achieves halving new service launch times (12 → 6 weeks). When a service fails, AI identifies the root cause in less than 30 seconds and executes automatic remediation in 85% of cases.
Broader context:
Cloud-native is a cultural and organizational change. It requires DevOps, CI/CD, and AI models integrated into every layer. Traditional telcos that don’t migrate will become obsolete in less than a decade.
Sources:
4️⃣ Fraud detection in telecoms with AI
What happened:
Telecom fraud causes global losses exceeding US $38 billion annually. AI enables analyzing traffic, call patterns, and authentications in real time to detect spoofing and anomalies before they impact.
According to a recent study, more than 60% of telecom fraud attacks use identity spoofing techniques, and 40% involve credential theft. Without AI, detecting these attacks can take hours or days. With advanced AI systems, detection occurs in milliseconds, reducing losses by more than 70%.
Why it matters:
AI models process billions of records per day to identify fraudulent patterns. Blocking and alerts are automated, reducing damage in seconds. Operators are already commercializing “fraud as a service” for businesses and MVNOs.
Concrete example:
Major UK mobile networks have achieved a 60% reduction in fraudulent calls after implementing AI for “number spoofing” blocking. The system analyzes more than 500 million daily calls and automatically blocks fraudulent numbers before they complete a call.
Broader context:
Telecom fraud is a growing threat. Attackers use AI to create sophisticated attacks, so defenses must also use AI. Fraud detection is no longer reactive: it’s predictive and preventive.
Sources:
- “Telecom fraud in 2025: what operators need to know” — Telecoms.com, 2025
5️⃣ Cloud-telco collaborations to power generative AI
What happened:
Verizon and AWS announced a partnership to deploy ultra-low latency fiber that supports generative AI models at the edge. These collaborations show a global trend: connectivity merges with AI infrastructure.
More than 15 similar partnerships have been announced in 2025 between operators and hyperscalers (AWS, Microsoft Azure, Google Cloud). The edge computing market for generative AI is projected at more than $25 billion by 2027, with 45% annual growth.
Why it matters:
Generative AI needs processing capacity close to the user to reduce response times. Telcos move from selling bandwidth to selling “connected intelligence”. New revenue sources open in edge computing and industrial AI services.
Concrete example:
A European provider launches an “AI zone” at the edge of its network with a hyperscaler. Inference latency drops from 50 ms to 10 ms in industrial robotics applications, improving productivity by ~8% and reducing defects by ~5%.
Broader context:
Cloud-telco collaborations represent a new business model. Telcos provide low-latency connectivity, hyperscalers provide AI infrastructure. Together, they create ecosystems that didn’t exist before. The revenue model changes from “gigabits per second” to “inferences per second”.
Sources:
- “Bridging the services revenue gap with AI and cloud-driven solutions” — Telecoms.com, 2025
🔹 Bonus Track — Novice hackers and AI: the new digital battlefield
What happened:
AI is also democratizing cybercrime. Models like WormGPT and FraudGPT, available on illegal forums, allow users without technical knowledge to create phishing, malware, and deepfakes with previously unimaginable accuracy.
According to ENISA, more than 80% of social engineering attacks in 2025 used AI to generate personalized messages. KELA detected a 200% increase in mentions of AI-based malicious tools on cybercrime forums in the past year. Automated phishing, synthetic voice, and auto-generated malware are no longer the territory of experts, but of anyone with access to a model.
Why it matters:
The threat landscape changes radically. Attackers no longer need advanced technical skills: AI does the heavy lifting. This means more attacks, more sophisticated, and harder to detect. Defenses must also evolve: defensive AI is not optional, it’s essential.
Concrete example:
AI-assisted phishing campaigns achieve success rates above 35%, compared to 11% of traditional manual attacks. Messages adapt to the language, tone, and local references of each victim in seconds.
Broader context:
The democratization of cybercrime with AI is an existential threat to telecommunications. Networks are the main attack vector, and telcos must defend not only their infrastructure, but also their customers. Offensive AI requires defensive AI.
Sources:
- “ENISA Threat Landscape 2025” — ENISA, 2025
🧭 Conclusion
AI is no longer a layer added to the network: it is the network itself.
From operational margins to satellite orbit and digital security, all signals converge in the same direction: intelligence as infrastructure.
Those who understand this transition — telcos, providers, regulators, or industrial partners — will have a substantial advantage in the next decade of connectivity.
If you’re an engineer, architect, or project manager in this space, you should look from two angles — technical and strategic — and act proactively. It’s not enough to know “how,” you also need to understand “why” and “for what.”
The future of telecommunications won’t just be faster, but smarter, more autonomous, and more secure. And those who lead that convergence will have the advantage.
📚 Recommended Readings
AI and ML Powering the Next Leap in Satellite Networks
Perspective on how artificial intelligence is revolutionizing satellite networks.
👉 https://www.telecoms.com/ai/ai-and-ml-powering-the-next-leap-in-satellite-networks
A Smart Way for Telcos to Accelerate Their Cloud Strategy
BCG’s guide on how telecommunications companies can accelerate their cloud-native journey.
👉 https://www.bcg.com/publications/2025/how-telcos-can-accelerate-cloud-journey
Telecom fraud in 2025: what operators need to know
Analysis of evolving fraud threats and how AI is helping operators stay ahead.
👉 https://www.telecoms.com/security/telecom-fraud-in-2025-what-operators-need-to-know-to-stay-ahead-of-evolving-threats
I invite you to comment: Which of these trends do you think will have the greatest impact on your sector? How is your organization preparing for this transformation where AI becomes living infrastructure?
See you next week with more news.
✍️ Claudio from ViaMind
Dare to imagine, create, and transform.
