AI and 5G
A Force Multiplier, or Lost Opportunity? (Part 1)
Mobile app builders and Telcos, this blog is for you.
If software is eating the world, AI is the chef cooking up a future that’s smarter, faster, and more innovative. It’s not just changing the game — it’s rewriting the rules. Existing boundaries are mere reference points of what was, and certainly not an indication of what’s ahead. That’s where we get excited. Today, KhasmX, the skunkworks division of the 5G Open Innovation Lab, alongside alumni Archetype AI, Radium Cloud, Graphiant, and partners like Accenture, Amdocs, and Dell Technologies are pioneering a new AI reference architecture. This architecture aims to bring AI compute closer to users and unlock untapped Telco opportunities within the software ecosystem. Telco networks touch over 5.6 billion mobile users globally. With PwC projecting AI to contribute $15.7 trillion to the global economy by 2030, the potential is immense. Let’s dig in.
The Transformative Potential of AI
Each technology led transformation cycle is built on the back of “new infrastructure” cycle. The Internet in the 90s connected homes and workers. The mobile and cloud infrastructure cycle since the late 2000s gave another healthy boost to the economy. We are now firmly and undeniably in the “AI Infrastructure” cycle. Each cycle has measurable impact on the country’s GDP as it starts to scale the adoption and availability of the platform. This opens up the opportunities for the players who build the infrastructure (just like it did in the Internet, Mobile, and Cloud cycles) and make the platform available to developers to dream up new applications and services.
In a speech at the World Government Summit in Dubai, Jensen Huang, CEO of Nvidia, stated “It is our job to create computing technology such that nobody has to program. We will democratize programming for everyone, allowing people to use their native language to code.” Assuming he’s right, AI will not only reshape industry norms but will also push computing closer to users. Why? By 2025, the world will generate 463 exabytes of data daily (that’s 463 million terabytes!). Average monthly mobile data is expected to more than double from 21 GB to 56 GB by 2029. Sure, we may be streaming a lot, but our AI agents could be streaming even more on our behalf via AI’s direct integration into mobile apps, laptops and consumer goods, leading to exponential data growth. How can AI change mobile app experiences?
Here are examples from Archetype AI:
1. Package handling demo (based on accelerometer data)
3. What’s the vibe at home? (based on smartphone camera video)
A new role for Telcos?
Clearly future mobile apps, and the AI models they are built on, will generate a significant amount of computational traffic. What sits between mobile apps and the Cloud? Telco networks. The question is, will Telcos continue moving packets or will they explore moving up stack by dipping their toes into the AI computing world? Imagine if they did, what would be their upside? For users (consumers and enterprises) it could be transformational. Here are seven ways we see the opportunity playing out:
1. Value-Added Services: This could include AI-driven applications such as natural language processing for customer support, predictive analytics for business intelligence, and personalized content generation. These services can be offered as premium features, generating additional revenue streams.
2. Enhanced Customer Experience: GenAI can significantly enhance the customer experience by enabling smarter, more responsive, and personalized services. For example, AI-driven chatbots can provide instant customer support, reducing the need for human intervention and improving customer satisfaction. Enhanced customer experience can lead to higher customer retention and increased average revenue per user (ARPU) AND cloud consumption services!
3. Improved Network Efficiency: AI can optimize network operations by managing traffic, and ensuring efficient resource allocation. Imagine processing GenAI queries on the network versus moving petabytes of data to the Cloud and back.
4. Monetization of Data: GenAI can help telcos monetize the vast amounts of data they collect by providing insights and analytics services to businesses. For example, telcos can offer AI-driven marketing insights, customer behavior analysis, and other data-driven services to enterprises looking to enhance their operations and marketing strategies. Going further, Telcos can utilize AI platforms like Archetype AI and Radium to fine-tune their own custom foundation models, optimized for the unique data insights that can only be unlocked by each Telco based on their private network data.
5. New Business Models: Integrating GenAI within their networks allows telcos to explore new business models. For instance, they can offer AI-as-a-Service (AIaaS) platforms, where businesses and developers can access and utilize AI tools and infrastructure on a subscription or pay-as-you-go basis. This opens up fundamentally a new revenue stream independent of traditional telecom services.
6. Partnership Opportunities: Telcos can form strategic partnerships with AI, Cloud, and technology companies to co-develop and market AI-driven solutions. These partnerships can expand their service offerings and customer base, leading to mutual revenue growth.
7. Support for Emerging Technologies: With the proliferation of IoT devices and the increasing demand for real-time data processing, having GenAI compute services can support the development and deployment of innovative applications in areas such as smart cities, autonomous vehicles, and telemedicine.
Telcos vs. Cloud?
Wrong question. Rather, Telco AND Cloud. While the industry has been talking about the convergence of communications and compute for a long time, the AI era presents a perfect opportunity to integrate. Telcos, with their infrastructure investments, have a unique opportunity to become key players in this cycle. Integrating AI compute services within Telco networks presents a compelling opportunity for both Telcos and developers without causing competitive tension with hyperscaler cloud providers. Developers should be able to leverage Telcos for local, lightweight AI computation, enhancing mobile app experiences with reduced latency and faster response times. Hyperscaler cloud providers remain essential for handling large-scale, advanced AI computations. Current day content delivery networks (CDNs) mirror the hub-and-spoke model we envision, where Telcos act as local spokes for immediate processing needs and hyperscalers serve as the central hub for intensive tasks.
5Gs true opportunity; it’s not all about connectivity.
The industry evolution into 5G is far bigger than packet moving and feature based network APIs. According to an article by Mike Dano of LightReading, US Telcos have spent $230B in “wireless capex and spectrum licenses from 2020 to 2023”. In the same article an analyst is referenced saying, “yet [the operators] have only grown wireless EBITDA [earnings before interest, taxes, depreciation and amortization] by ~$10 billion, the majority of which has come through cost-cutting synergies from the Sprint/T-Mobile merger.”
On a global level the GSMA is forecasting Telco spend on 5G will exceed $1.5 trillion in CAPEX between 2023–2030 yet in the same period operator revenue will is forecasted to grow by $140 billion from $1.11 trillion to $1.25 trillion. Let’s be honest, not a lot of uplift for the CAPEX outlay. In comparison, the cloud computing industry is projected to have a market size of $2.39 trillion representing a growth of $1.8 trillion from the $626 billion market in 2023. From an ROI standpoint, and assuming projections are accurate, Telcos could see a 9.3% ROI from their CAPEX investments through 2030 compared to 220.45% from the Cloud industry in the same period. Why? Packet moving, while important, isn’t where considerable market value is placed. There is, however, exponential value in distributed computing services and that’s the miss. Change is hard, uncomfortable, and risky. So is staying in the same spot, providing the same service (albeit faster), and expecting different results. The window for Telcos to think differently about future opportunities is now and it’s narrowing.
Tomorrow’s blog…
Tomorrow we’ll explore why now matters, what’s in the stack we are proposing, and insights from our participating alumni and partners. Go here to find it.
