The Six Five Team discusses VMWare Explore 2023.
If you are interested in watching the full episode you can check it out here.
Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.
Daniel Newman: All right, so I’m going to be broad on this. I know you like to geek out a little bit more than I do. So, I’m going to basically say for me, there were three themes of VMware. One, the company is actually one and two go together because it’s about the combination of Broadcom, its leadership and its vision. The company actually had a very strong keynote, which I was concerned about when we’d had some of the internal briefings because it was very hyper product focused. But I thought that the keynote was actually well-delivered and hit on some very important high notes.
Having said that, Hock Tan’s presence, the visibility of Broadcom VMware, it felt to me for the first time that VMware was truly acquiescing to the fact that this deal was getting done.
Patrick Moorhead: Yeah, it’s like this is a done deal.
Daniel Newman: I just think whether it was the way they presented products, the safe harbor statements that were hyper emphasized, which they have to do, but the amount of emphasis on it was unique.
Patrick Moorhead: I’ve been going to events for 30 years and that is the longest statement I’ve ever heard in my entire career.
Daniel Newman: Yeah. It usually it gets flashed and they go on with it. But I mean they were very clear that like, “Look, this is kind of how we’re developing. If VMware remains an independent company, this is kind of the roadmap we’re on, but basically we have no idea if this is how we’re going to continue.” And that’s the bottom line is because under Hock Tan, you can absolutely be sure just by looking at his track record of acquisitions, that he’s going to be very bold in looking at the company, looking at his products, looking at its services, figuring out how to optimize, maximize, create efficiencies.
This is his playbook, so of course I’ll never speak to anything that we would know that’s specific. But what I would say is just generally speaking, his playbook and VMware’s current playbook, VMware’s probably too broad, has too many products, too many distributed focuses. They have too many things, they’re moving slowly, need more focus on a few of its best products, maximizing the margins and doing it with probably less headcount. And that’s just like I said, that’s the playbook, but this deal’s getting done.
I just absolutely left the event more sure than ever and obviously having the US regulators basically walk away decided to not pursue anything was probably the biggest news of the week. Because in the UK we knew they decided and then it went official. But the US one, everyone was waiting on. And that was because the US has been busy trying to litigate everything that’s bigger than $10 billion. LeadIcon apparently decided this one’s going to go. The only other thing I wanted to talk about was AI. I wasn’t quite sure how VMware was going to talk about AI and I thought it did a really good job.
Besides bringing Jensen on stage, which every company does now. They’re like a rule, you’re not really an AI if Jensen doesn’t come out. And then you’re not really an AI if Jensen doesn’t come out and present you with a GPU, even if it is a gaming GPU.
Patrick Moorhead: Patrick Moorhead. I know, man. It’s basically “kiss the ring,” that’s what this is about.
Daniel Newman: But in all seriousness, what was he supposed to do with that RTX?
Patrick Moorhead: I don’t know, you can’t run any AI anything on it. So, by the way, it almost amplified the we can’t get any H100s.
Daniel Newman: I can’t even give you a fake one backstage.
Patrick Moorhead: In my head, it totally backfired. I’m like proof positive, you can’t get H100s-
Daniel Newman: But he didn’t even say, even on stage, you can’t get H100s then he made a joke about it?
Patrick Moorhead: He did.
Daniel Newman: But I do think that the VMware private AI foundation, which is its services to basically enable private data in generative AI in a single stack, sounded compelling. Ecosystem, infrastructure partners, software stack, it seemed to have all the pieces, it seemed to be well thought out. And the weird bringing our lawyer on stage thing, that was probably the worst part of the keynote for me. I didn’t really understand what they were trying to do there. But I think what they were basically trying to say is that large enterprises are going, and especially in regulated industries that run hybrid architectures are going to have a load of complexity about trying to do generative applications and run generative tools in their business.
To have an offering that’s hybrid that focuses on data privacy and security and allows for prem is going to be important to those companies. And I thought the company, like I said, did a good job at obviously having all the parts, pieces, the GPUs and DPUs from NVIDIA and partnering with the likes of Dell, HPE and Lenovo on the hardware side to enable prem and cloud and private AI to be delivering generative tools. So, those were my two big takeaways. Basically they did a good job on AI, which I thought they might fumble and they didn’t and this Broadcom deal’s getting done.
Patrick Moorhead: Wow. For not knowing quite what to say, you certainly had a lot to say.
Daniel Newman: Hey, that’s why it’s called the six-ten.
Patrick Moorhead: No, exactly. Just wait, sit back, get a cup of coffee folks. No, there is so much to talk about at this show. We could probably take the entire show-
Daniel Newman: It costs extra, by the way.
Patrick Moorhead: And talk about it. First off, it’s an odd show for me because I’m used to showing up and meeting one-on-one with CEOs, presidents and the business unit leaders. That doesn’t happen at this show which limits my ability to come to conclusions. Because part of my research methodology is to meet with, now I did meet with the acquiring company CEO, Hock Tan but not with Ragu. By the way, I did appreciate him doing unmoderated NDA Q and A, but there’s nothing like meeting face-to-face to really get to know what’s going on. And hey, that’s part of my research methodology, it’s what I think makes more insights great.
So, a lot of stuff was announced in areas that wouldn’t be a shock. Just like you said, increases in multi-cloud and a new on-prem and orchestrator for AI. Not a whole lot about the client stuff with the exception of automated workspaces. So, let me dive into the multi-cloud part. First, I had to contend with, hey, Tanzu is changing, Arita is changing and I’m going to have to do a lot of work to untangle that. I do think that VMware has the best hybrid multi-cloud story that encompasses apps. It encompasses networking and certain elements of security. It’s not taking off financially, I don’t understand why. But then again, on the other side of my mouth, that growth is around 30%, which is larger than the hyperscalers.
So, it is growing just not adding enough as many dollars as I would expect. And I’ll be honest, I don’t have the answer at this point. My instinct tells me it’s a go-to-market issue and it’s a sales issue. That partners can’t punch it through, their salespeople can’t punch it through. And I think that enterprises are contending so much potentially with this new AI burden that it could be an issue. But out of the other side of my mouth, I’ll say to really pull off AI, you really have to be an arbiter of multi-cloud.
So, I need to get underneath that. I especially liked the end of the multi-cloud section where they talked about the capabilities and it’s like, “Oh, it’s all rainbows and unicorns,” I love that. Their president is a great guy in terms of how he can deliver this message. But I think it’s a go-to-market and sales execution issue, that’s the only thing I can place. Now, things like NSX don’t have all the capabilities that these enterprises needed for multi-cloud. That’s where NSX plus comes into the stage. I also thought it was interesting just how deep the company is getting into ransomware, which is interesting.
Because if you look at Veeam, Cohesity, Haiku, Dell Technologies, they’re now competing with many of the folks that they partner with. But let me move to AI here. Dan, like you, I thought the story was one of the best ways of showing how complicated it’s going to be to get your on-premise or private data infused into generative AI. And I really liked the discussion about the difference between what they call predictive AI, which is essentially high-end analytics, ML and DL versus generative AI. But one thing, again, the story is great.
I thought what enterprises do need to fully understand is that it’s not just putting a customer service chatbot out there and getting access to public data. You got to combine marketing, supply chain sales, HR, legal, R&D, customer experience, manufacturing. It’s the ultimate connecting the front end to the backend, if you really want to do what enterprises want to do. And if you think about the companies, let’s just say 75% of the data’s on-prem, does this drive people to put all their data into the cloud, this type of data? Or is this going to be a boon for on-prem?
And I don’t have a definitive answer yet on that. First of all, I need to know how many GPUs NVIDIA is actually going to send to enterprise. So, if you don’t send GPUs to enterprise, not only can you not do inference, you can’t do training. And yes, the big companies with the proprietary data are going to do training on-prem. The chief legal counsel going on stage, I thought the timing was off, which made it a little bit off. But I did like the message which was, “Hey, you’ve got private intellectual property, you’ve got private data and you need private access. How are you going to work with all this in Generative AI Land?”
Finally, and I’m going to breeze over these because we’re running out of time and this can’t be the 650. Two main offerings, first one was an open architecture that they specified. Now, you can’t buy an architecture but it’s a framework for everybody else other than NVIDIA to come in and add value. But the second thing they did is, like you said, they rolled out Jensen and there is a VMware private AI foundation with NVIDIA, which essentially is more of a prepackaged, pre-tested, it works. It’s leveraging not only NVIDIA hardware, but CUDA, but also, and I think even more importantly, the libraries, the frameworks that come in there that come along when you have NVIDIA.
So, with NVIDIA means to me that they’re opening the door with AMD, with Intel, with Groq, somebody like that. Net-net got a lot of work to do to figure out everything. There was not a date for delivery for private AI, not even a year. And I think Dan, this might lead to or connect with what you talked about related to the disclosure at the beginning. Are these announcements real? Could they be reversed in six months? We’ll see, and again, I’m not trying to spread FUD here. This deal is going to go through, but with a gigantic statement like that at the beginning of the show and not having dates for private AI, you got to wonder.