🌐 AI and Telecommunications Trends – Week of October 27, 2025
This week I want to start with something that’s changing much more than just one sector: how AI is redefining work, industry, and the networks that connect us. Then we’ll see how those dynamics clash, complement each other, and evolve within the telecommunications world.
1️⃣ Automation, AI, and Work in the Age of Amazon.com, Inc.
What happened:
Amazon introduced a robot called Blue Jay, working in one of its warehouses in South Carolina, reducing delivery times by 25%, injury rates by 40%, and required space by 30%. The company said this robot is already making concrete differences: it processes more than 1,000 packages per hour, compared to 600 with traditional methods.
At the same time, Amazon announced it plans to cut around 14,000 corporate positions in the coming months, as part of a strategy where AI plays a key role. This represents approximately 1.5% of its global corporate workforce, but the company projects that automation with AI will generate savings of more than $2 billion annually.
Why it matters:
Amazon combines robotic automation, operational management AI, and logistics optimization to achieve greater efficiency. But the story isn’t just technical: it raises fundamental questions about work, talent, roles that will be replaced or transformed.
What’s interesting is that these industrial changes offer clues—and lessons—for the telecommunications world: an industry also entering a stage of profound transformation.
Concrete example:
Imagine an Amazon logistics facility where robots move shelves, an AI system assigns tasks, and humans focus on optimizing exceptions. Now take that same model to the telecom network: automatic nodes, AI that detects failures, humans who act where AI cannot. The analogy is direct.
Broader context:
Human work is being redefined: at Amazon, robots don’t just do “physical tasks,” AI also decides routes, manages inventory, and optimizes processes. Human work shifts toward supervision, maintenance, strategy.
For telecom professionals: if your role is tied to repetitive operations, prepare to think “how can AI do this better?” and “what will my function be in that scenario?” At an industrial level, automation combined with AI requires a new employment contract: training, retraining, human-machine hybrid roles.
Sources:
- “Amazon sees faster delivery speeds with hi-tech driver eyeglasses, AI” – Reuters, October 22, 2025
- “Amazon to cut about 14,000 corporate jobs in AI push” – Reuters, October 28, 2025
2️⃣ AI Risks and Governance in Telecom: A “New Normal”
What happened:
A recent report from consultancy Ernst & Young Global Limited (EY) warns that network operators face an increasingly complex risk environment:
- Privacy, security, and trust remain the number one risk.
- Ineffective technological transformation ranks second.
- A changing geopolitical environment also enters the list, meaning data sovereignty, regulatory blocs, and global alliances matter much more than many thought.
For example, only 59% of operators say they have a robust methodology to identify and mitigate AI risks, compared to 66% in other sectors. Additionally, less than 45% of telcos have continuous algorithmic bias monitoring systems, and only 38% perform regular audits of their AI models. The study, which surveyed more than 200 global operators, reveals that 72% consider AI governance their main challenge for 2026.
Why it matters:
Telcos operate networks that are critical infrastructure. Adding AI means greater complexity: models, data, automation, continuous learning. But without a clear framework, those advances can become vulnerabilities.
Concrete example:
A telco launches an AI module to predict network failures. But it doesn’t evaluate whether the data used has biases or if the model favors better-served urban areas. Result: a certain customer segment receives worse service, complaints arise, reputation is impacted, regulatory sanctions. The AI didn’t fail due to technology, but due to lack of governance.
Broader context:
It’s not enough to “do AI.” You have to design governance, ethics, traceability, bias verification. Integrating security and privacy from design (“security by design”, “privacy by design”) is no longer optional.
As a project manager or engineer, your challenge is getting technical, business, and compliance teams to work together: AI is no longer just “AI work,” it’s “organizational work.”
Sources:
- “Telco sector faces rising risks in AI, ineffective transformation and evolving geopolitical environment” – EY, Oct 2025
3️⃣ From Pilots to Scale: AI + Telecom Entering Commercial Mode
What happened:
According to a study by NVIDIA Corporation of more than 450 telecom sector professionals:
- 84% say AI has helped increase annual revenue, with an average increase of 12-18% in AI-based services.
- 77% say it has reduced costs, with average savings of 15-25% in network operations.
- 68% report customer experience improvements of more than 30%.
- 52% are already monetizing AI services directly, generating between $5-50 million in additional annual revenue.
In parallel, a recent GSMA Intelligence analysis notes that telcos’ focus is changing: now more than efficiency, revenue generation and AI-sovereign models are being sought. The AI services market for telecom is estimated at more than $15 billion for 2025, with projected growth of 35% annually.
Why it matters:
A few years ago, AI in telecom was mostly pilot: chatbots, predictive maintenance, internal optimization. Today the conversation is more mature: “how do we monetize AI?”, “how do we reach scale?”, “how do we move from cost savings to value creation?”
Concrete example:
A telco decides that the AI used to optimize the network won’t just save internal costs, but will be packaged as a service for businesses (Industry 4.0, healthcare, smart cities). Thus the network stops being “passive” and becomes an “income-generating asset.”
Broader context:
Experimentation is no longer enough: adoption must be operational, scalable, measurable. Competition accelerates: whoever does it first will gain advantage.
If you lead a roadmap, it’s not enough to “activate AI”; you have to define how it will generate revenue, what new services will emerge, how success will be measured.
Sources:
- NVIDIA: “State of AI in Telecommunications” Survey Report
- “Telcos shift AI goals towards revenue generation” – TelecomTV, Oct 28 2025
4️⃣ Edge AI + Cloud-Native Architecture: When Intelligence Moves to the Edge
What happened:
A clear trend is that intelligence is moving toward the “edge” of the network and adopting cloud-native architectures. For example: an STL Partners article at MWC 2025 states that the “trigger” for the resurgence of edge computing is AI. The edge computing market will grow from $40 billion in 2024 to more than $155 billion by 2030, with AI as the main driver. More than 60% of new edge investments are specifically allocated to AI infrastructure.
Another IoT-Now analysis shows how edge-AI drives innovation and competitive advantage in telecom: companies implementing edge-AI report latency reductions of 70-90%, bandwidth efficiency improvements of 40%, and data transmission cost savings of 50%.
Why it matters:
With 5G, IoT, connected vehicles, smart cities… latency, bandwidth, and local processing requirements increase. Traditional architectures (sending all data to the center) are no longer enough. That’s why edge + AI + cloud-native is the formula many telcos are adopting.
Concrete example:
Imagine a telco network in a smart city: traffic sensors, cameras, real-time analysis at edge nodes, AI that adjusts traffic lights, reduces congestion. All thanks to the “brain” being at the edge, not at a distant center. The network stops being passive and becomes reactive.
Broader context:
Project teams must define where AI will be processed: in the data center, at the edge node, on the device. Hardware, software, orchestration requirements change: multiple nodes, microservices, containers, frequent updates.
For end customers: improved experience (lower latency, more reliability). For the company: new architecture, operation, and governance challenges.
Sources:
- STL Partners: “Edge computing at MWC 2025: AI is the trigger”
- “How edge AI is driving telco innovation & competitive advantage” – IoT-Now
5️⃣ AI-Native Networks, Data Sovereignty, and the Path to 6G
What happened:
The networks we deploy today are preparing for what’s coming: 6G, automation, integrated AI, and environmental data control. Recently, an academic paper titled “Sovereign AI for 6G” describes how AI models will be embedded from network design, and highlights data sovereignty as a key element. It’s estimated that 6G networks will require processing more than 1 terabyte of data per second per square kilometer, and AI will be essential to handle this scale. More than 15 countries are already investing in 6G research, with combined budgets exceeding $10 billion, and more than 70% of these projects include sovereign AI components.
Why it matters:
In a few decades, the difference won’t just be how many gigabits per second you transfer, but how many decisions the network makes on its own, how autonomous it is, and who controls the data. “AI-nativeness” implies the network is designed for AI, not that AI is “added” to the network. And data sovereignty means where they’re processed, who controls them, how they’re updated, matters a lot.
Concrete example:
A European operator defines its 6G roadmap thinking that, when the network goes into operation, all its AI models will reside in nodes within the country, with federated updates and automatic audits. This way it ensures compliance, control, and advantage over competitors who just “add AI.”
Broader context:
Telcos must start answering: Where will my AI models reside? In the country? In distributed nodes? How do I guarantee regulatory compliance?
Network architecture changes: multi-node, federated, autonomous. It’s not enough to have “more power,” “more intelligence” is needed.
For you as a project leader: coordinate technology, legal, operations, security today for a future that’s being built now.
Sources:
Final Reflection
The combination of the five topics shows a clear movement: AI is no longer separate from telecom or industry, but the nervous system of the next generation of networks, services, and jobs.
From Amazon we get the signal that intelligent automation redefines operations and human work; in telecom we see that risks, monetization, architecture, and future vision are already underway.
If you’re an engineer, architect, or project manager in this space, it’s worth looking from two angles—technical and strategic—and acting with anticipation. 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 ethical. And those who lead that convergence will have the advantage.
📚 Recommended readings
State of AI in Telecommunications (NVIDIA)
NVIDIA’s comprehensive survey on AI adoption and impact in telecom sector.
👉 https://www.nvidia.com/en-us/lp/industries/telecommunications/state-of-ai-in-telecom-survey-report/
Telco sector faces rising risks in AI (EY)
EY’s report on risk management and governance challenges in telecom AI implementation.
👉 https://www.ey.com/en_gl/newsroom/2025/10/telco-sector-faces-rising-risks-in-ai-ineffective-transformation-and-evolving-geopolitical-environment
Edge computing at MWC 2025: AI is the trigger (STL Partners)
Analysis on how AI is driving edge computing adoption in telecommunications.
👉 https://stlpartners.com/articles/edge-computing/edge-computing-at-mwc-25/
6G Infrastructures for Edge AI (ArXiv)
Academic perspective on AI-native networks and data sovereignty in 6G architectures.
👉 https://arxiv.org/abs/2506.10570
✍️ Claudio from ViaMind
Dare to imagine, create and transform.
