To The Moon

$NVIDIA(NVDA)$  | AI Infrastructure | Telecom Edge Computing

Nvidia’s next growth opportunity may extend beyond the giant data centres powering today’s AI boom.

At GTC Berlin, Jensen Huang outlined a broader vision of the AI economy, spanning energy, computing, AI factories, models and applications. One particularly interesting development is the potential transformation of telecommunications networks into distributed AI computing platforms.

This could open a new market for Nvidia, bringing AI processing closer to users rather than concentrating almost everything in centralised data centres.

But there is an important question: Can telecom companies turn their existing networks into profitable AI infrastructure quickly enough to create a meaningful new revenue stream for Nvidia?

📡 1. Telecom Networks Could Become More Than Connectivity Providers

Traditionally, telecom operators have invested heavily in mobile networks, fibre connections and infrastructure that moves data between people and devices.

The next phase could involve using that infrastructure to run AI workloads.

Telecom operators already have distributed sites, connectivity and access to local markets. By adding accelerated computing, they could potentially offer AI services closer to where users and businesses need them.

This matters because not every AI application needs to send data to a distant data centre.

Applications involving voice assistants, industrial automation, robotics and real-time video analysis may benefit from processing data closer to its source. Lower latency can improve responsiveness, while local processing may help organisations manage data and connectivity requirements.

Nvidia and its telecom partners are working towards this model through AI-RAN, which combines AI computing with radio access network infrastructure.

If adoption expands, Nvidia could gain another route to deliver computing capacity beyond traditional hyperscale data centres.

🏭 2. Why Nvidia Is Expanding Its Supply Chain at the Same Time

The telecom opportunity is developing alongside Nvidia’s efforts to address another challenge: the capacity required to manufacture and deliver its AI infrastructure.

On 21 October, Nvidia announced new agreements with manufacturing partners covering advanced packaging, semiconductor testing, substrates and optical components.

The announced plans include:

🔹 A targeted 2.5-fold increase in substrate capacity over three years.

🔹 An increase of more than 300% in advanced packaging capacity with Amkor by the end of 2028.

🔹 Plans to double total semiconductor testing capacity between 2026 and 2027.

🔹 A targeted 15-fold expansion in optical-engine capacity for co-packaged optics by mid-2027.

These are planned capacity expansions, not completed increases in production. Nevertheless, they show how much manufacturing infrastructure is needed to support Nvidia’s ambitions.

The important point is that Nvidia’s growth depends on more than chip demand. It also depends on whether its suppliers can deliver the components required to build complete computing systems.

Expanding this network could help reduce supply constraints and support higher production volumes, although the benefits will take time to materialise.

🔍 3. The Opportunity Is Bigger Than Selling More GPUs

The longer-term opportunity may lie in Nvidia becoming part of a much broader AI infrastructure ecosystem.

Centralised data centres will remain essential for training large AI models and handling demanding workloads. However, AI inference—the process of using trained models to generate responses or make predictions—can also take place closer to users.

Imagine a factory using AI to analyse production lines in real time, a telecom operator offering AI-powered business services, or a robotics company processing information across local and central computing systems.

These applications could create demand for computing at multiple locations.

For Nvidia, this potentially means more opportunities to supply accelerated computing platforms, networking technology and supporting software.

It could also broaden the range of customers investing in AI infrastructure.

However, there is a difference between a promising technical demonstration and a commercially successful business. Telecom operators must establish which services customers will pay for, while Nvidia and its partners need to demonstrate that these deployments can operate economically at scale.

The size of the potential market is exciting, but actual revenue will depend on adoption, utilisation and customers’ willingness to spend.

⚖️ 4. What Could Slow the Opportunity Down?

There are several challenges worth watching.

First, deployment takes time. Telecom operators need to integrate new computing capabilities into existing networks, and expanding semiconductor packaging and testing facilities also requires investment and execution.

Second, the economics must work. Telecom companies may be interested in AI services, but they still need sufficient paying customers and utilisation to justify the additional infrastructure costs.

Third, competition will remain intense. Nvidia is not the only company seeking opportunities in AI infrastructure. Customers will compare performance, operating costs, power efficiency and available alternatives before committing to large deployments.

Finally, capacity expansion does not automatically create revenue. More manufacturing capacity helps Nvidia fulfil demand, but the financial payoff depends on actual orders, product deliveries and customer spending.

These factors make execution just as important as the underlying AI opportunity.

💡 My Take: Nvidia’s Next Chapter May Be About Where AI Runs

The first phase of the AI investment boom focused heavily on building powerful computing systems in large data centres.

The next phase could involve distributing AI capabilities across a wider network of locations, from centralised facilities to telecom sites and enterprise environments.

If that transition gains traction, Nvidia could benefit from a broader range of AI infrastructure deployments rather than relying solely on continued expansion at large data centres.

Its supply-chain investments also suggest that manufacturing capacity and reliable component availability will remain important to supporting future growth.

However, I would not treat telecom-edge AI as an immediate earnings catalyst simply because the technology is advancing. Commercial adoption, deployment schedules and customer economics will determine how quickly this opportunity translates into meaningful financial results.

For investors, the key is to distinguish between a potentially large future market and revenue that the company can actually deliver.

The opportunity is worth watching, but execution will determine its value.

Do you think telecom networks will become an important new market for Nvidia’s AI computing technology, or will most of the economic value continue to concentrate in large data centres?

Disclaimer: This post is for informational and discussion purposes only and does not constitute financial advice or a recommendation to buy or sell any securities.