Revolution Next: Mobile networks in the AI age

The internet revolution truly began around 1995, when web browsers began to take off. Everyone was talking about online business in Japan, but nobody wanted to take the risk.

I left my banking job and created the world’s first B2B2C online marketplace – an idea nobody believed in at the time – and became part of the first generation of internet entrepreneurs. Today, only a handful of those entrepreneurs still run their companies.

As I discussed recently with Rakuten TV CEO Cedric Dufour on stage during the 2024 edition of Viva Technology in Paris, throughout my career, I have seen numerous tech revolutions that rival that of the internet. Perhaps the biggest was the smartphone; today in online shopping, some 90% of people make purchases on a mobile device. They’re used to pay, learn, entertain and access medical services. Smartphones are an essential part of our social life.

Now, the same thing is happening again – this time with AI.

Empowering the smartphone revolution

A few years ago, one of the biggest problems we had in Japan was mobile phone plan prices. They accounted for a significant portion of the average household budget, and that number was rising. Why was this happening? Because the technologies we were using were too old.

Until Rakuten Mobile, the telecom industry still worked off old physical switches that have barely changed since the age of fixed line telephones.

“With mobile as the next major pillar, our ecosystem is even stronger.”

MICKEY MIKITANI, RAKUTEN GROUP, CHAIRMAN & CEO

The industry had been oligopolized by just a few players who had no incentive to change. We went from 2G to 3G, 4G and now 5G – the numbers changed, but the hardware architecture continued to be vertically integrated and locked in by a few dominant hardware vendors. There was simply no competition.

In launching Rakuten Mobile, we had the advantage of being able to build everything from scratch, unburdened by legacy technology. We virtualized our entire network, running it on software that traditionally required extensive hardware investment. We didn’t follow the path laid out for us by other telecom companies.

Rakuten Mobile’s path to profitability

There are three reasons I believe Rakuten Mobile will be a success. One is that it will likely become very profitable, because our operational costs are significantly less than our competitors’, thanks to our software-driven approach.

The second is that Rakuten Mobile is serving as a test drive for our technology, proving that it works. This has already borne fruit with the launch of 1&1’s network in Germany, built in partnership with Rakuten Symphony. Subscribers in Germany are already moving to Europe’s first cloud-native mobile network built on Open RAN standards.

The third is Rakuten Mobile’s contribution to our ecosystem. We already had e-commerce, credit card, banking and some of the richest data in the world. And this data is telling compelling stories.

Rakuten Mobile subscribers are buying 60% more from Rakuten Ichiba. They spend twice as much on our travel offerings and use our credit card and banking services more as well. With mobile as the next major pillar, our ecosystem is even stronger.

Ushering in the age of “open” Open RAN

In recent years, there’s been much talk about how we could open up the telecom industry – make it into an open-source operating system. Everybody was talking about it, but nobody was doing it. So we took the risk.

When we began developing our Open RAN software, people were more than skeptical – they laughed at the idea. It was too complicated.

But we did it. And now Rakuten Mobile has over seven million subscribers, connected by our nation-wide base station network with 99.9% population coverage when combined with our roaming partner. It’s one of the best performing networks in the world.

“I would love for the technology behind Rakuten Mobile to become the Linux of telecommunications.”

MICKEY MIKITANI, RAKUTEN GROUP, CHAIRMAN & CEO

Our work has essentially broken the grip that telecom hardware vendors had on the industry. This will lead to faster, more affordable, unlimited data access in markets around the world. Operators no longer need to depend on a few locked-in vendors to run the entirety of their networks.

Right now, Rakuten Symphony is one of the only companies with proven 4G and 5G Open RAN software and interoperability with most radio hardware. Other companies are still trying to catch up, but we want to accelerate this process.

That’s why we decided to launch our “Real Open RAN Licensing Program” during MWC Barcelona earlier this year. We will offer our source code at a very reasonable price – even to our competitors – so that operators won’t need to develop the software themselves.

I would love for the technology behind Rakuten Mobile to become the Linux of telecommunications. I think this is going to be a huge revolution that changes the nature of wireless networks.

The next revolution: AI

The shift from hardware to software is important for another reason: the rise of AI. AI cannot manage hardware so easily, but with software, it can manage everything from the radio to the core.

“I believe AI will redefine the future of not only online business, but the very nature of society itself.”

MICKEY MIKITANI, RAKUTEN GROUP, CHAIRMAN & CEO

This autonomous capability is one reason Rakuten Mobile’s engineering costs are a fraction of our competitors. AI plays a key part in our network, something that’s not possible for companies that rely on hardware and physical labor. What we’re doing is very different from our competitors.

More broadly speaking, Rakuten has two strategies with AI: One is our partnerships with OpenAI, Google and other AI companies. Simultaneously, we are also developing our own AI: We recently unveiled the Rakuten AI 7B large language model, which demonstrated exceptional performance compared to other open-source Japanese language LLMs.

While the Big AI players are very smart and innovative, large models require a lot of calculations, which consume a lot of money. We are considering smaller, more vertically integrated AI tools specifically for commerce, finance, travel and more. This makes sense for us to do with our highly valuable dataset.

The second strategy is on the service side – what we are calling our “Triple 20” project. We are looking to use AI to improve our operational efficiency by 20%, with specific targets for each business. We’re also going to improve marketing efficiency by 20% and empower our clients, merchants and partners – these smaller companies that do not have the capability to implement and use AI – by 20%.

I believe AI will redefine the future of not only online business, but the very nature of society itself.