What's Up with Tech?

Broadcom’s Telco Cloud Playbook

Evan Kirstel

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 10:32

Interested in being a guest? Email us at admin@evankirstel.com

The hallways at MWC were buzzing, but our conversation cut straight to what matters for operators right now: shipping new services faster without getting boxed in. We sat down with Broadcom to unpack VMware Telco Cloud Platform 9 and how it blends cloud-native speed with the reliability telco networks demand. This release pulls previously separate components into a single, horizontal stack, pairing one automation engine for both virtual machines and containers with an embedded Kubernetes runtime. The payoff is simpler operations, shorter launch cycles, and real freedom to choose best-of-breed network functions.

We also dive into AI where talk meets traction. Broadcom is rolling out integrated management for NVIDIA GPUs directly in the platform, making it practical to offer GPU-as-a-service and manage accelerator resources with the same tooling operators already trust. That creates a clean path for enterprise AI workloads and sets the stage for the next leap: virtualization of GPUs. By bringing their virtualization heritage to accelerators, Broadcom is targeting higher utilization, better isolation, and elastic scaling for inference and training across the network and at the edge.

Vendor lock-in has long slowed telco transformation, so we explore how a true horizontal platform changes the equation. CSPs can run virtualized and containerized network functions from multiple suppliers on the same infrastructure, swap vendors as needs evolve, and rely on a rigorous ecosystem program for pre-qualification and certification. That means less integration risk, fewer surprises in deployment, and faster routes to revenue. We round out the conversation with Europe’s rising demand for sovereign cloud—country-level compliance, data privacy, and security—and how recent national certifications prove the model is working in practice.

If you care about open ecosystems, sovereign-ready architectures, and an AI-enabled telco cloud that actually reduces time to market, this conversation is for you. Follow the show, share it with a colleague who’s building next-gen networks, and leave a quick review to tell us what you want covered next.

Support the show

More at https://linktr.ee/EvanKirstel

SPEAKER_00

Hey everybody, I'm here at Mobile World Congress, day one with Broadcom. Gianluca, how are you?

SPEAKER_01

I'm very well, thank you.

SPEAKER_00

Well, great to see you guys. Great to see you. Of course, everyone knows Broadcom. Give us a little bit of an overview of your business, your mission within the team. Yes. And your impressions of day one.

SPEAKER_01

Yeah, no, I look after inside Broadcom the telecom, the telco market. We are organized in a vertical organization, and we look after the top brands in EMIA from a telco perspective. Exciting days, very intense day one, as usual it is at Mobiwa Congress here in Barcelona, indeed.

SPEAKER_00

So let's talk a little bit about roadmap. You have VMware telco cloud platform nine. That's a mouthful, but talk about the innovation that's coming from Broadcom these days. You've really tripled down on RD to have lots going on. Give us a snapshot.

SPEAKER_01

Indeed, indeed. No, the Techno Cloud Platform 9 is in the uh is indeed the very first version that we are releasing as after the acquisition is the results of the last couple of years of our engineering development. And it's bringing together basically the whiffer modules that so far we have marketing in different in different forms into one single stack with a couple of key elements from an innovation point of view. The single automation engine that is capable to orchestrate and automate both virtual machines and containerized functions to create additional efficiency and shorter time to market for the adoption of these functionalities. And the other key element is the uh Kubernetes runtime uh capabilities embedded into the stack, again, meant to accelerate the adoption of the containerized functions into the virtual private networks.

AI Strategy And GPU Management

SPEAKER_00

Yeah, it's a fantastic platform. Um it's gaining so much traction uh for a variety of reasons, not the least of which being private AI, AI in general. Talk about the demand for AI services. And um you have quite a unique strategy that's pretty differentiated in the market. Maybe talk about the strategy and the go-to-market for AI.

Avoiding Lock-In With A Horizontal Stack

SPEAKER_01

Yeah, no, indeed. Indeed, at Broadcom, as everybody knows, right? We are super active when it is about the journey to the to the AI technology adoption from multiple aspects. Much known when it is about the semiconductor business, but indeed we are now actively developing our strategy in the software domain as well. And indeed, the version 9 of our softwares that we have just announced are coming out with the first very first integration of the management software layer for the NVIDIA GPUs. So we are capable inside the stack to manage the GPUs from NVIDIA to then provide the first proposition to the market from the ability to even market as a service GPU management. And um and it is a very first uh, of course, step to what is gonna be the next the next evolution when talking about uh particularly the um virtualization of GPUs. Virtualization is the core of our business, so we are looking at that as the next frontier, the next step for uh uh for the GPUs as well.

SPEAKER_00

Yeah, the uh VMware traditionally has had a pretty incredible strategy about avoiding lock-in and interoperability, and that's very much at the core of what you're doing with CSPs. How do how do you help CSPs avoid lock-in and into proprietary uh kind of solutions? Indeed, indeed.

Speeding Integration And Time To Market

SPEAKER_01

Our view when the Techco Cloud platform in particular is about is the uh implementation of an horizontal platform from a cloud infrastructure point of view, which is the vision through which you can and any customer can uh implement on the same platform whichever function they want to implement. Could it be virtual, could it be in containerized mode, coming from whichever supplier network function providers they can be of their choice. The ability to have multiple different vendors on the same platform, the ability to change vendors time by time when and where this is eventually needed, is the vision behind the horizontal cavalry. Of course, this requires a very uh well-developed ecosystem of partnership, right, with all of the network equipment providers, which is what we are quite well in a structured form addressing inside our organization. Um we are collaborating with the top uh you know vendors in this in this domain, and there is a very firm strict program through which we do run pre-qualification of the of the functions anytime they need to come implemented in the platform. Any customer can receive a package where the functions that they have chosen are pre-qualified in our lab. So 80% of the job, if you like, is validated. And then, of course, there is the customization element that customer my customer might require that is done on uh on the lab together with the customers. And uh and then the certification program, right? So we have a very well defined program for which any kind of the network providers equipment that the that requires to certify their roads over our platform need to apply to a certification program and comply with a certain number of standards that again can facilitate then the implementation of the functions in the over the platform or in any active projects with our customers.

SPEAKER_00

Fantastic. I mean, a lot of that is reducing time to market and a lot of trial and error, I guess, in in integration, but also getting new services to market quickly. That seems to be a theme I'm hearing from Broadcom.

Private Cloud As A Service And AI

SPEAKER_01

Correct. And um, and of course, you know, when private cloud is about development in this in this form in a complete in a complete stack the ability to complement functionalities that might help, especially from a sovereignty capabilities point of view the possibility to promote it as a service. So cloud private cloud functionalities that in particular in the telegram sector we see all of our customers are promoting as a service to any kind of vertikal, could it be automotive, healthcare public sector in the market? Is a revenue generation element. When is it coming and when will it come with the DAI embedded into it? Again, with the version 9 of our stack, it's a possibility to again add additional capabilities, even from an AI as a service point of view. Could it be GPU as a service or could it be any other use case that is coming through and that actually many of uh of the people are experimenting these days in uh in a starting mode?

Europe’s Push For Sovereign Cloud

SPEAKER_00

Yeah, it's a great time to be in the industry, a great time to be an operator and create new services. And also sovereign cloud, you know, it used to be a nice idea, maybe it was a compliance thing, but it's really caught on, right? Every country uh seems to really demand it now. I don't uh give us a snapshot of Europe in particular and what the uh sovereign cloud requirements are. It is.

SPEAKER_01

It is uh for different reasons, uh good or less good. Uh I would say not good. It's a very felt uh you know, priority inside the market these days. It's also a set of requirements that, of course, are defined at the European level that from a security data management and data privacy, if you like, handling point of view are generally defined. And each and every country has their own specific customized, if you like, requirements that through our plattform we are going to implement. And very recently in France actually we had one case with one of our kustum that has been achieving the national sovereignty certification using our technology, which is always a very good point for us. We are very proud of it. This is basically happening in every single uh country. We are very active, primarily with the telco and incumbent players, which are, of course, on the front lines of promoting sovereignty uh capabilities in each uh in each of their own countries.

SPEAKER_00

And you continue to really invest in the platform. Uh RD is a big part of what Broadcom does. Um talk us about what customers are looking for maybe in the future. I think you work very closely to integrate their requirements and uh deliver them.

Closing Thanks

SPEAKER_01

Yes, of course, then it pretty much goes into the direction we were just discussing, right? So then the ability to, on one end, uh create efficiency, reduce costs, making sure that from a cloud uh private uh private cloud standpoint, the same capabilities that sometimes are offered to the development uh let's say community in a public in a public form are coming available on the on the private. So the same, if you like, flexibilities for the developer community, we offer it pretty much in private, not just on the IT kind of applications point of view, but also again on the networking point of view, where on top of the flexibility and efficiencies, there are requirements in terms of uh you know uh, of course, uh reliability, uh mission-critical kind of functions that those uh those networks need to need to handle, and uh and more and more AI-enabled capabilities. The AI again is gonna be the the next frontier of the uh of the innovation, both from a form through which they can contain cost while operating uh the networks, and also in the possibility to generate new new revenue streams in in different forms when this is coming. I mentioned to you GPU as a service is one case where TECCO can generate new revenues. This is the direction that basically this market is taking these days.

SPEAKER_00

Fantastic. Well, congratulations on the continued success, onwards and upwards.

SPEAKER_01

Thank you very much. Thanks a lot.

SPEAKER_00

Thank you. Thanks everyone for listening and watching.