Early this year Broadcom (AVGO) announced that it would acquire VMware (NYSE:VMW) for about $61 billion in cash and stock, the deal is expected to close in Q4 23. Last month the CEO of VMWare stated that things are on track and with no other bids on the table, things are looking pretty good.
Broadcom is confident. Shelling out $142.50 in cash which was a 49% premium to VMware’s last close before the news dropped and once that news dropped the technology market reacted and react it did.
To say there was a smidgen of concern would be an understatement. Before we get into that, let’s do a thirty-second primer on the acquirer and acquiree – if you feel like you know enough about these two you can safely skip to ‘The future of VMware’.
Thirty seconds on Broadcom
Broadcom is owned by its parent company Broadcom Inc. For the purpose of the following discussion, we’re talking about Broadcom Inc and the capabilities it has within its extended portfolio. Broadcom Inc does around $ 28B in annual revenue.
Its revenue can be split into the following general areas;
- Semiconductor solutions, which make up 76% of revenue
- Infrastructure software, which makes up 24% of revenue
By far the largest chunk of their revenue is generated in semiconductor solutions. What makes up that revenue? Its’ essentially infrastructure communication hardware. This consists of the following solutions:
- Top-of-rack ethernet switches. You might be familiar with its legendary tomahawk chipset.
- A large collection of physical network interface solutions connecting devices to both network and storage fabrics
- A variety of wireless controllers that you’ll find in everything from Dell notebooks to iPhones
Do you know what else is interesting about Broadcom? Much like Apple, it sells lots of hardware however it doesn’t manufacture the chips that sell to the market. It’s what’s known as a fabless company.
Broadcom’s chips are physically produced by contract manufacturing behemoths like GlobalFoundries, Silterra and TSMC.
Broadcom has a history of buying some well-known technology brands. They have proven successful in dramatically increasing the profitability of their acquisitions. They do this by enhancing profit margins, implementing a lean operation and ensuring the product roadmap supports only the profitable areas of their portfolio.
On paper, this sounds like good business. Still, much of the negativity created around the VMware acquisition has developed through the perceived lack of innovation and the sweating of existing product lines to their customers has been viewed as being analogous to a rent extraction business model. This is the reason why Broadcom’s previous M&A outings have been viewed by the technology sector as a negative outcome for both employees and existing customers.
The businesses that Broadcom acquires all have several traits in common, they already have significant market penetration and there is no clear long-term roadmap for growth. Symantec enterprise security is one of their most well-known acquisitions and is a good example of their ideal target acquisition. Since the start of my career, almost every business I’ve joined has at one point used Symantec enterprise security. Following the acquisition, Broadcom stated that they will eliminate over $1 Billion in spending across both R&D (40% cut) and Sales (82% cut). Symantec enterprise security with appropriate R&D should be riding on the crest of the ridiculous growth in cybersecurity products but when you half your innovation efforts and focus only on delivering solutions to your existing top 2000 customers any product manager worth their salt will tell you that’s not a winning formula for growth.
But let’s not paint Broadcom in a purely negative light. The CEO of Broadcom Hock Tan is a shrewd businessman and frankly, his move to buy VMware is genius.
Thirty seconds on VMware
Let’s start off by saying VMware is an inspirational company. It has rightly earned its place in the annals of history as the first business to successfully virtualise the x86 microprocessor architecture.
By every right, VMware should be one of the top three public cloud providers. However, for several reasons I won’t go into, it just couldn’t quite find its way.
As a business with a solid foothold in every fortune 500 company, it’s astonishing that it only generates the revenue that it does, around $ 12.85B annually. The fact VMware’s revenue isn’t 10X where it is now is a testament to the outright domination of open-source virtualisation software and its eventual acceptance by the enterprise in the form of AWS over a decade ago.
Its public cloud execution struggled multiple times and the customer outcomes delivered by its product portfolio can be delivered by many of its competitors. Microsoft can provide identical outcomes enhanced by its Azure capability and it’s also in Microsoft’s gift to offer their enterprise customers some compelling licencing arrangements to jump ship. If VMware found a way to embrace mass market adoption by navigating the problem of being forced to run in competitors’ public clouds while overcoming a challenging licence model from Microsoft then things might just have been different.
Most decision-making technology leaders earned their stripes with VMware and because of this, there is huge sentimentality and respect for the business in its core customer base. But time doesn’t stand still, and succession plans will unfold. Gen X leadership are starting to realise that they’d rather decide which poolside resort to select rather than stress over the content of their exec reports. Much of that well-earned sentimentality towards the brand will diminish as millennials take up the reigns.
The fundamental challenge with VMware there is no straightforward roadmap to continued growth, their competitive advantage is no longer unique, and the OS Virtualization game is over and has been for some time. The shift to Kubernetes simply doesn’t translate to a VMware infrastructure. They’ve been trying to pivot into telcos that have been on a path of embracing open-source software so the success of that will unfold in time. Their end-user compute portfolio like everything in their portfolio is exceptional, but in a world of M365 and Azure Virtual Desktop, it’s been constrained by the previously mentioned licensing challenges.
The future of VMware
VMware is still a growing business, defying the naysayers by more than doubling revenue since 2017. It’s a software business so its gross margin is around the 80% mark. By every indication, VMware is a healthy business. So why has its share price been declining for the last half-decade? The answer of course is the challenge from the public cloud. It would be unfair to say that was the only reason. VMware has been acquired by EMC, then Dell and now, we surmise, Broadcom. All of this since 2004. There are few things that will impact a business’s long-term strategic ambitions than the persistent disruption of M&A.
Now to safely have this discussion, it would be prudent for me to highlight that the major three public cloud providers have a wonderful relationship with VMware. In fact, the major public cloud providers even offer customers the ability to run VMware on their public cloud infrastructure – it doesn’t get cosier than that!
But let’s enter a hypothetical universe in which the top three cloud providers see VMware’s entire install base as revenue in waiting.
Many VMware customers, defence contractors, healthcare, shipping etc either don’t want to move or simply can’t move to the public cloud due to latency challenges, specific financing requirements or security restrictions. What can the public cloud providers do about that?
That’s right, the public cloud has gone private cloud.
Last week I wrote about the problem with the word ‘cloud’ and how the big three cloud providers have done such an amazing job of painting a crystal-clear idea of cloud that trying to modify its image is proving challenging. You remember the narrative from the last decade, don’t you?
Sales lead: “Public Cloud far more cost-effective than private cloud infrastructure. You only pay for what you need and thanks to our innovative special sauce, you can do some niche fancy things that you can’t do outside our cloud.”
CIO: “Ok, sounds compelling, IT isn’t our core business after all so this seems to make a lot of sense.”
Today the message is very different.
Sales lead: “It’s all about our fancy special sauce that supports your transformation efforts. The whole point of the public cloud is transformation. Wait, are you telling me you did a lift and shift without transforming? Sheesh. Anyway, let’s have a chat about our new private cloud infrastructure.”
CIO: “Wait. Wait. What did you just say? Can you repeat that?”
Here is the crux of the problem, even VMware’s enterprise private cloud business is under threat from public cloud providers. Welcome to Hybrid Cloud and without a credible public cloud narrative analysts can see the writing on the wall for VMware.
So why on earth would Broadcom wish to buy them?
I surmise there are two reasons, both are amazing.
Reason 1: Buying VMware is about protecting Broadcom’s semiconductor revenue.
Apple was the first large business to disassociate itself from the reliance on proprietary microprocessor hardware. It did this first with the shift from Motorola to the IBM Power platform. When they realised, they couldn’t navigate power efficiency challenges the market was blown away when Steve took to the stage and announced that “Hell froze over” and Apple was moving to Intel.
If Steve wanted to release a ground-breaking new phone with a battery life that exceeded three seconds, intel probably wasn’t the move. So once again they ported their software to ARM and following its success with the iPhone, they began developing their own chips initially for the iPhone and iPad and recently they have released the staggering M1 and M2 processors.
Geopolitical challenges aside, Apple is a business that’s very much in charge of its own destiny. In 2020, Broadcom estimated that 20% of its revenue comes from Apple and they have a $ 15B supply agreement for their wireless chips that runs until 2023.
Do you know who else is making their own chips? Microsoft.
To quote this article on Nasdaq.com “Microsoft MSFT recently hired key Apple AAPL engineer Mike Filippo to accelerate the process of creating homegrown chips and reduce the dependency on third-party providers.”. At this very second, you can download Windows software [https://www.microsoft.com/en-us/software-download/windowsinsiderpreviewARM64] that runs on ARM. Those aren’t Microsoft-designed chips of course, but the course is clear and as with all things Microsoft I expect they will offer spectacular energy efficiency and performance once out in the wild while continuing to support their growth. As with Apple, they are continuing to follow a path of both hardware and software verticalization. People might get upset about this but there is precedence with Apple pulling it off, so it could be a successful target model for Microsoft.
Apple, Amazon, Microsoft and Google are making their own chips, servers and network devices. Those chips will be produced by the same collection of contract manufacturers that make Broadcom chips. Not ideal. Seems like a reasonable time to diversify.
Broadcom customers are in no way dependent on Broadcom. There are between five to ten competitors offering the very same solutions. This combined with in-house chip design begins to put 76% of their revenue at long-term risk. Every one of their customers is also a key strategic partner of Microsoft. Upset Microsoft and you upset your customers and your risk your continued hardware revenue. Keep Microsoft onside by getting out of their way and even though we’ll begin to see the active VMware customer-based decrease, they’ll increase revenue with the previously mentioned business strategy alongside significantly increasing their semiconductor hardware sales as Microsoft directs their partners to Broadcom as their preferred hardware vendor. It’s genius really.
Reason 2: Buying VMware is about growing Broadcom’s semiconductor revenue.
Broadcom could (and should) be a next-generation challenger to Cisco.
You might not have heard of a DPU before and that’s not surprising, Data Processing Units are niche. But big things start from humble beginnings.
This fancy card contains a CPU, NIC and programmable data acceleration engines. Trying not to get too technical, it is effectively it’s an astonishingly powerful computer that interconnects a bare metal server to the network or storage fabric.
Broadcom makes amazing network equipment. What they haven’t had is an operating system, until now. DPUs have the option to radically transform computer networking, moving the advanced network complexity out from the core of the network to DPU devices themselves. The outcome for customers is the delivery of software-defined networking, identical to the type you would find in Microsoft Azure which leverages the open-source SONiC network operating system. This would allow customers of VMware/Broadcom to have the level of network orchestration found within the public cloud but locally within their non-public cloud infrastructure. This would immediately circumnavigate expensive SDN solutions such as Cisco ACI for the enterprise.
It’s no secret that the lasting value of VMware has always been the acquisition of Martin Casados software-defined networking business Nicira Networks which eventually became VMware NSX.
With NSX, VMware customers have an incredibly powerful set of network and security features to support their virtual machine and container infrastructure. From virtual firewalls, distributed security via micro-segmentation, virtual load balancers, routing and switching NSX has it all. Broadcom’s acquisition of VMware delivers a network operating system for their network hardware, along with the DPU revolution that’s about to take hold pushed NSX out of vSphere and into the bare metal server and network market – a direct challenger to every brand name network provider.
Summary
Yes, VMware is going to look very different to what it does now. That isn’t a bad thing.
The future for VMware has never been brighter, being bought by Broadcom opens a new world of growth for the company.
I believe that we’ll see VMware shift focus to creating a DPU/NSX agnostic network fabric, supported by Broadcom’s existing hardware portfolio which they’ll use to capitalise on an incredible network opportunity.
They will strategically allow Microsoft to gain traction in the Hybrid Cloud / Azure Stack arena. They will do this by increasing focus on their existing customer base and aligning their product roadmap to only focus on growth areas. As per their proven business model, they will increase pricing to the point where it’s noticeably uncomfortable but not uncomfortable enough for their existing customer base to move to a different solution.
Keeping Microsoft happy will ensure they protect their existing semiconductor revenue from partners that rely far more on Microsoft than Broadcom. It will also ensure that in the emerging new world of in-house chip design, Broadcom’s chips will continue to be purchased over their competitors.
The result of this strategy ensures both Broadcom and VMware revenue continues to grow in the short to medium term. And as specialist chip designers like Broadcom lose their uniqueness in the marketplace, Broadcom will ensure long-term revenue growth through their NSX-infused network and security hardware portfolio.
So is Broadcom buying VMware a mistake?
No, it is not. It’s far from a mistake, it’s genius.
