Artificial intelligence for cities, infrastructure and investor strategies

Large technological transitions rarely happen in a single sector. Here the thread that connects urban transport, real estate development, compute infrastructure, agentic platforms and even the diamond market is artificial intelligence. Artificial intelligence acts as a lever to transform resources, trust and services. The article explores how concrete projects and products show viable paths for Italian founders and investors.

When infrastructure becomes agent capability thanks to artificial intelligence and cities learn to speak

Peerium-style compute innovation proposes to make “one million processors work like a single computer” using unified buses and shared memory, with products like the Atlas 950 SuperPoD already deployed in clusters of 256,000 cards. Huawei describes this architecture as the way to make available the power needed for large-scale training and inference. One million processors can cooperate as a single computer.

This level of power intertwines with urban needs shown by Vinhomes. The Vinhomes Green Paradise - Can Gio project obtained the WCCD ISO 37122 smart city certification and relies on digital platforms. The platforms serve over 35,000 users, that is already a significant data base for automated services. Smart city certification creates a measurable data base. The pairing of large compute clusters and urban data opens practical possibilities for models that manage traffic, energy and safety in real time.

Compute capacity and data quality turn artificial intelligence into an active urban service.

Business agents and platforms: how to industrialize autonomy

On the platforms side, Huawei Cloud proposes the idea of “Agentic Infra” aimed at running reliable agents in production, with solutions like Context Memory Storage that offers petabyte-scale memory space and terabyte read performance with 50% more performance than peers. The AgentArts ecosystem provides over 5,000 context protocol assets and over 1,000 vertical assets, while the open source edition openJiuwen has exceeded 50,000 stars and 3.29 million downloads, signs of adoption and community. Contextual memory at petabyte scale makes long-term agents possible.

The same concerns emerge at large sector expos like MetroTrans 2026. The event convenes more than 500 exhibitors and over 40,000 visitors to discuss signaling, communications and “Rail + artificial intelligence” solutions. Here the practical question is: how to translate digital assets and models into reliable operations for critical infrastructure? The answers integrate unified scheduling, fast error recovery and end-to-end monitoring. Huawei cites fault recovery within 10 minutes and support for over 40 days of stable training as examples of operational reliability. Fault recovery in production must occur within a few minutes.

Markets, trust and value: from the diamond chain to urban services

The adoption of artificial intelligence also changes the perception of value. De Beers Group and Chow Tai Fook launched the Desert Rose collection, 20 pieces available in 25 cities, and focus on transparency and traceability technologies to strengthen consumer trust. Consumer trust is sustained by digital traceability. Traceability projects in luxury materials show that AI and data are not only for efficiency, but also to support the value of durable goods.

This dynamic intersects with real estate development. Since 2021 Vinhomes has built a TechnoPark building certified LEED Platinum and a mini-app of services for residents. These initiatives have generated operational data used to improve services and attract investment. According to junto, McKinsey estimated that AI can unlock up to 550 billion dollars in the real estate sector. Industry reports find that 75% of professionals use AI, that is three out of four. Artificial intelligence increases perceived value in real estate.

Better data and digital traceability support both commercial value and reputation.

Practical choices for founders and investors: risk, interoperability and scale

For those building products or scaling startups, the aggregated case teaches three priorities: infrastructure interoperability, data governance and operational resilience. Peerium-type architectures show that scale requires unified protocols to connect CPU, NPU, memory and network. Agentic platforms highlight that reusable assets are needed: 5,000 context assets and 1,000 verticals are useful to reduce development times. Interoperability and reuse accelerate time-to-market.

From the governance point of view, trust in assets, from diamonds to homes, goes through digital proofs and certifications. De Beers and Chow Tai Fook combine design and verification technologies, while Vinhomes uses international certifications to attract capital and residents. Finally, managing operational risk remains crucial. MetroTrans and cloud infrastructure examples indicate that prolonged training and fault recovery determine the sustainability of solutions in the real world. Managing operational risk is a condition for scale.

The scenario remains evolving. Some technologies are already mature for local-scale experiments. Others require open standards and partnerships between infrastructure providers, agent developers and urban operators. The decisions made today on protocols, governance and service design will decide who will manage to turn artificial intelligence into measurable value tomorrow.

Sources:

• Huawei Cloud Rolls Out Enterprise AI Products Across the Board… | europeanbusinessmagazine.com
• Connecting Global Resources, Empowering Industry Growth | europeanbusinessmagazine.com
• World Cities Turn to AI: Vinhomes Green Paradise - Can Gio Builds… | europeanbusinessmagazine.com
• Huawei Pioneers a New Computing Architecture for the AI Era: Making… | europeanbusinessmagazine.com