Telcos and Tokens: Back to the Future?

My good friends over at Mavenir announced a new “Integrated AI Platform” the other day and it provided a reminder of two things. The TM Forum’s DTW Ignite event is next week and, just in time, we’re currently entering a period of “Telco as Token Merchant” hype.  

Telco interest in monetizing AI Token consumption isn’t hard to explain. 

Tapping into the outsized fervour around AI in China, the major Chinese mobile operators have launched service plans including a package of AI services and the tokens to power them. While market dynamics in the People’s Republic do not always translate elsewhere, everyone is taking notice. After all, while AI has proven invaluable in driving operational efficiencies, Tokenomics promises a path to revenue generation. It also promises a veritable journey back in time to an era when mobile operators delivered metered services and revenues were more tightly connected to service consumption.  

Put simply, the hope is that tokens could be the new (or, rather, old) gigabytes.  

We dug into some of this in a recent report – the first in a series that will attempt to answer the question on everyone’s mind. Just how big is the potential market? 

But, before everyone gets too excited, a few reality checks are useful.  

• China is China. There’s a reason these plans are a Chinese phenomenon; Chinese telcos have invested in massive AI capacity while consumers line up to get OpenClaw installed on their laptops and phones (thanks to support from local governments and tech players). Few markets can match this level of supply meeting demand.  

• Why Metering Faded. Mobile plans with unlimited data or big buckets of gigabytes are an easy sell. They give consumers a sense of certainty that they won’t run out of data or face overage-induced bill shock. Plus, translating everyday app usage into an understanding of data usage isn’t a natural competence for more folks. But we expect token usage to be different?  

• The Fall of Token-maxxing. When Jensen Huang denigrates engineers who don’t gobble up enough tokens, it’s only natural to look for ways to monetize consumption across business and consumer markets. When Uber burns through its 2026 AI budget in four months and caps employee spending, we have to acknowledge that today’s consumption may not look like tomorrow’s. Budget constraints, model innovation and process inefficiencies will impact token usage, at least on an individual level.  

Will token consumption grow? As more people do more things with AI? Of course.  

Will token pricing also gyrate across countries and over time, making usage and revenue opportunities hard to predict in nascent consumer segments? Of course.  

The takeaway is that any plan to monetize AI usage needs to be broader than just metering token consumption. You see this in the Chinese plans where operators are bundling tokens with robust model packages. You see this in Mavenir’s launch which talks about “monthly AI token plans” alongside “per-department enterprise quotas, and SLA-backed tiers for operator-provided AI services” with a recognition that “no single deployment model fits every use case.”  

It all provides some hope that even if token metering has an uncertain future, building fuller solutions can still help telcos tap into AI demand.