One of the online issues raised in the last week was whether telcos were inherently bad at doing platforms, and whether they should instead focus on APIs. I want to offer my view on this, starting with a simple answer and then justifying it. The simple answer is that yes, they’re bad at platforms, and no, they shouldn’t focus on APIs. The justification? Well, we’ll get to that now.
What is a “platform?” It’s a framework for offering a service, in the context of our discussion. What is an API? It’s a way to expose features for exploitation by others. The position the LinkedIn post takes is that telco platforms, to reverse a popular saying were too big to succeed. If telcos had simply framed APIs they could have taken a smaller role suited to their monopoly-utility mindset, and let someone else build on them. I don’t disagree, but let’s keep going here.
What telcos are looking for is relevance. They’ve bitched for decades that they’d been “disintermediated”, meaning disconnected from the business models and revenues that came out of the Internet, which they saw themselves as the foundation players for. So the obvious question here is whether, by pushing a strategy that centered on APIs, the telcos would increase their relevance, decrease their disintermediation. To that question, I’d have to offer a “No!”
The whole purpose of an API is to expose something, a feature you have and can offer. Whether that’s good or bad depends on what happens when you do the exposing. If it involves you in something new and profitable, it’s a good thing. If it only contributes to further disintermediation, it’s a bad thing. So, which is it? We don’t really know for sure.
The option the referenced LinkedIn post offers is “APIs that guarantee identity, trust, and quality for enterprises, governments, and AI agents.” I can’t say for sure, but I think it’s likely that these features are expected to be available from telco infrastructure. What that would mean is a combination of positives and negatives. On the positive side, it wouldn’t cost the telcos much to do this sort of exposure since they already have the features. It could attract partners who’d exploit the APIs to create a new set of retail offerings, and in theory the telcos could charge for the APIs and so earn revenue. On the minus side, telcos would have to set a very attractive price for those APIs, somehow protect themselves from having competitors exploit them, and overcome the fact that using telco APIs would eliminate a potential source of differentiation for their partners. If everyone bases their super-new-service-set on the same API primitives, what differentiates the service? Only price, which is absolutely not what startups and innovators and VCs look for in an opportunity.
This doesn’t mean that the basic premise of the post is wrong, though. Telcos are terrible at platforms, because they’re terrible at identifying a viable business opportunity. They could be good at exposing what they already have because, well, they already have it. No insight required. The problem is that what they have is what they think already got exposed, resulting in that disintermediation. If they want to “re-mediate” they have to create APIs that link to something valuable that they can own, and that can’t easily be used against them by competitors.
You’d be right to think that this is taking an excursion toward platform-hood, which I just agreed telcos were awful at. How do we resolve that tension? Half-way measures get a bad rap, but maybe in this case they’d be justified. Not a “platform”, but a half-platform? Syntactic-speak could say that’s a “plat” or a “form”.
“Plat” here is “platitude” which is what an “API strategy” is, because an API is just a technical coupling. If you imagined some API bazaar peddling its wares, the first question you’d ask as a prospective buyer wouldn’t be price, it would be “what does it expose?” Jumping on APIs as a strategy is a mistake for that reason; it’s just a way of disguising the fact that what you’re doing is at best exposing what’s already been done and already known to be needed, or discarded. At worst, you’re offering disintermediation as a feature.
So we need to look at the “form” option. The future, for both telcos and OTT rivals, is a set of new services. What telcos dare not do is imagine they can be the retail providers of these new services; they have too many barriers bred into their DNA to accomplish that. At their first move, which would surely be to start on an endless round of meetings, rivals would be out selling what the meetings hadn’t even decided to consider. “Form” as a strategy is based on the notion that while retail offerings are beyond them, telcos can surely see the foundation elements that will form retail futures. What those elements are is also pretty obvious; the future has to lie in real-time, real-world changes in telecommunications.
Ah, IoT. Yes, but not just IoT. Back when IoT came out (the term dates back to 1999), when an MIT executive director said “If we had computers that knew everything there was to know about things, using data they gathered without any help from us, we would be able to track and count everything and greatly reduce waste, loss, and cost. We would know when things needed replacing, repairing, or recalling and whether they were fresh, or past their best.” This envisioned a world where the Internet was connecting all these “things”, which meant public sensors or at least a community of sensors, and the telco response was to immediately focus on selling “things” cellular subscriptions and providing the identity assurance that thing-billing mandated as a service feature to those exploiting the thing-data stream. Obviously, that didn’t work.
Things like high-quality networking for AI or AI agents won’t work either. Most agent connectivity requirements today exist inside the company data center, with some overflow into the VPN. Might that change? If it does, the question is what forces the change, and how do telcos get involved with that. You don’t create a “form” by depending on some other force to justify it. Which takes us back to IoT.
Real-world IoT, to succeed, can’t boil the real-world oceans all at once. An investment in things and thing communication demands a return, so you’re not going to start with a whole planetary thing deployment. The early need won’t be any of the things telcos saw as obvious, it will be for the tools to build the applications that justify those “obvious” things.
The LinkedIn piece is right in that sense. Telcos could and should avoid trying to do a full retail platform. They should focus on the underpinnings of valuable future retail experiences, identifying and deploying the API services to reduce the cost of further retail exploitation. That’s the model that creates telco partnerships; do the heavy, unattractive, lifting so that glamour can be built on top. But you have to move early enough to do this, or the glamour won’t need foundation from you at all because they’ll do it themselves.
