Why did BoYong acquire VMware? BoYong explains it in 20 pages!

Source: The content is compiled from CRN by Semiconductor Industry Observer (ID: icbank) , thank you.

1. Michael Dell: “This is a milestone moment”

2. Transaction Details

3. Technical collaboration plan for building “next-generation infrastructure software”

IV. 'Go-Shop' Terms

After Dell spun off VMware, we knew it was only a matter of time before an acquisition was announced.

When Pat Gelsinger left VMware's top job to return to Intel a year ago , we thought Intel was the logical buyer for VMware , and we still think there's a good case for Intel to own the virtualization layer -- it runs on the physical chips used for compute, storage, and networking. But it became clear that Intel didn't have the money or credit to build a foundry, let alone spend enough cash to build three of them to control VMware. So forget it.

AMD can't swallow a giant like VMware, having gobbled up Xilinx for $49 billion last year and early 2022. But with the help of some wealthy friends on Wall Street, maybe AMD can make a deal with VMware.

We have long believed that Nvidia's acquisition of Arm Holdings would be a tough sell to antitrust regulators, as many chipmakers in and outside the data center want the Arm architecture to remain strictly neutral - the only way SoftBank would be allowed to acquire Arm Holdings. Nonetheless, we accepted the commitment of Nvidia co-founder and CEO Jensen Huang that he wanted to acquire Arm to grow and protect it and adopt its distribution model for all of Nvidia's chips .

When the Arm Holdings deal looked pointless last summer, and as Nvidia was building closer ties with VMware, we suggested that Nvidia should buy VMware and take control of the entire HPC and AI software stack, from the hypervisor to the server nodes across the data center. (We also recently suggested that Nvidia complete this chain of events by acquiring SUSE Linux so that it could have its own Linux and Kubernetes distributions.) Such a deal would give Nvidia access to a massive traditional installed base of more than 300,000 enterprise customers and more than 100 million virtual machines for printing money.

But alas, Nvidia got distracted.

Avago founder Hock Tan isn't distracted, having acquired chipmaker Broadcom in May 2015 for $37 billion and taking its name from it.

Tan is known for making expensive acquisitions, relentlessly cutting costs, and making them pay for themselves. Tan wanted a larger, more distributed software business, not Symantec's security and CA systems management businesses, which cost a fortune and still squeezed out a lot of profit. This chart illustrates how good Tan was at building and expanding the conglomerate:

As long as there's cheap money, that is. Cheap money is running out fast, so now is the perfect time to raise $32 billion in cash from the banking network, a lot of cash and stock, assume $8 billion in debt from VMware, and run off with $61 billion for the global virtualization giant.

Before rumors of a VMware deal hit Wall Street on Monday, VMware had a market cap of about $40 billion, plus $3.63 billion in cash and $12.67 billion in long-term debt, on revenue of $12.85 billion for fiscal 2022 and net income of $1.82 billion for the year ended January. The company's top line is still growing, but its revenue is not. That's a problem Tan knows how to fix. With that comes higher prices for software licenses and support, as well as internal cost cuts and operational synergies from getting rid of VMware's back-office functions. But there's more.

Tan said in a presentation announcing the deal to Wall Street that the plan was to address the workload expansion VMware has been seeing since it entered the data center with the launch of the GSX Server hypervisor in 2002. But Tan will also focus on VMware's R&D because it is "uniquely positioned to innovate and drive customer success," and centralize sales and marketing efforts across the Broadcom portfolio on the massive VMware enterprise base. That's why Tan sold Broadcom's nascent but technologically powerful "Vulcan" Arm server chip business to Cavium, which revived the Vulcan effort as ThunderX2, but that was also largely shut down after Marvell acquired Cavium.

In the fiscal year just ended, VMware earned $4.7 billion before income taxes, depreciation, and amortization. (That's at a non-GAAP level, which is a generous accounting of profits.) Tan thinks Broadcom can raise VMware's mid-line to $8.5 billion.

The obvious question is why VMware can’t do this on its own. It’s simple. Everyone at VMware knows everyone, and it’s hard to cut people with a long history in a company. Most of the increased profits will come from VMware eliminating back-office functions. That’s how Computer Associates built its mainframe and Unix groups, and it’s no coincidence that CA was eaten by Broadcom as its first software acquisition. Tan learned a lot from CA founder Charles Wang, certainly before Broadcom acquired CA for $18.9 billion in November 2018, and if not directly before Wang’s death, then indirectly from Wang’s long and dramatic career building Computer Associates. (Wang died a month before Broadcom announced the acquisition.)

But as the chart above shows, Tan knew how to buy something with Avago and then unlever Avago (now Broadcom) so it could go out and acquire again. After the Broadcom deal, free cash flow was $1.7 billion and dividends were $1.94 per share in fiscal 2015. In fiscal 2021, Broadcom Group had free cash flow of $13.3 billion and dividends of $14.40 per share, and Tan expects it to reach $16.40 per share this year.

This chart gives you an idea of ​​how Broadcom has changed over the years:

In January 2017, after Avago completed its acquisition of Broadcom, and Avago was renamed Broadcom again, trailing 12 month (LTM) revenues were $15.6 billion. By the end of fiscal 2021, which ended in October, Broadcom had nearly doubled in size, with $20.4 billion in sales of chips, boards, and now systems. The company's Symantec and CA software businesses had sales of $7.1 billion, and as you'd expect, are much more profitable than the hardware side. Operating margins are in the mid-70% range, because there aren't a lot of sales. Why?

The secret to Broadcom’s software business is that about half of its revenue comes from IBM mainframe shops, and a very small number of accounts – about 500, according to Tom Krause, former Broadcom CFO and current head of the software group – drive a large portion of that software revenue. If you add VMware revenue through October 2021 to Broadcom’s numbers, you’ll see the stacked bar on the right, with the now-expanded Broadcom half software and half hardware. Now, Krause explained on a call with Wall Street, Broadcom will have 1,500 large global accounts to manage and cross-sell, while the other 299,000 VMware accounts will be sold by resellers and OEMs and developed and supported at VMware.

Here's what Broadcom is betting on:

While containers are all the rage among hyperscalers and cloud builders and all the cool kids imitating them, in the enterprise the VMware ESXi virtual machine remains the unit of software consumption, and companies have spent a fortune — and will continue to spend a fortune — maintaining the “21st century software mainframe” comprised of ESXi, vSAN virtual SAN storage, and NSX virtual networking — a goal that former VMware CEO Paul Maritz explained was shaky in March 2009 when the world economy faltered and created the conditions for VMware’s success.

VMware is a little different now. Its software has expanded to cover all aspects of the enterprise data center, its customer base is large, and that customer base is as risk-averse as ever. So it's a safe bet that most VMware customers will stay put, and that the number of VMs and workloads they'll manage on top of the VMware stack will remain increasingly modest over the next few years, as it has for the past decade. And the ability to deploy VMware's software infrastructure on public clouds gives it a hybrid story comparable to the one IBM pushed after it bought Red Hat for $34 billion in October 2018 .

The VMware deal might raise some U.S. national security concerns if Broadcom had remained headquartered in Singapore after the acquisition, but Broadcom moved its headquarters to San Jose years ago. Regulators might turn up their noses, but the reality is that no one feels as proprietary about VMware as its licensees and the UK government do about Arm. The original Arm deal also wouldn’t have passed muster if SoftBank wasn’t a Japan-based conglomerate with a variety of different assets.

It will be interesting to see if Michael Dell, who owns 40.2% of VMware, wants to become a majority shareholder in Broadcom. The price Broadcom paid for VMware was about a 40% premium, so it's hard to imagine Dell the man being unhappy about "unlocking this shareholder value," as they often say on Wall Street. Dell approved the deal when it was announced, as did his longtime private equity piggy bank, Silver Lake, which owns 10.2%.

So if anyone wants to take advantage of the 40-day "go shop" option VMware has to get a better deal, they better bring some rich friends along. Since Broadcom owns 50.2%, the deal is already done with those two votes. By the way, the way the deal is structured is that half of VMware will be bought with cash and half with Broadcom stock, and the deal will be prorated to make everything balance out. So maybe Dell and Silver Lake are cashing out, and everyone else who holds VMware stock will get Broadcom stock, and that's it.

As for that "go shop" provision, there aren't many people who want to buy VMware. As we said, we can make a case for Intel or Nvidia. Broadcom is also a logical choice. Cisco Systems could do it, but it's too distracted trying to fight Broadcom in the merchant chip market for data center switching and routing. Dell owns this company and spun it off, and Hewlett Packard Enterprise can't afford VMware. US regulators will never allow Lenovo and Inspur to buy VMware. Amazon Web Services, Google, and Microsoft could all do deals like this, and you could make a case for them to take VMware to help move all those ESXi VMs to a foundation in the cloud instead of on-premises.

Broadcom paid a high premium for VMware to make this goshop configuration a formality rather than a real possibility, as it hopes to build a balanced hardware and software conglomerate, and it now has no plans to separate Broadcom into hardware and software businesses.

“We see a lot of benefits in bringing all these different franchises that we have — hardware and software — under one umbrella,” Tan explained on a conference call with Wall Street. “Think of it this way. Merchant silicon is driving a trend. The old model was that you sold black box hardware and software systems to customers in the IT department. That’s what you did in the past. If something went wrong, you called support and screamed for help because you didn’t know what was going on inside. We are creating a disaggregated model between hardware and software. We may still know a lot less about the systems, but we certainly know the technologies that enable the systems, whether they’re switches, routers, compute or storage. Broadcom has a model of separating hardware and software, but if we combine them, we’re much stronger than if we separated them.”

In other words, just because your hardware and software are disaggregated doesn’t mean you can’t sell both.

★ Click [Read original text] at the end of the article to view the original link of this article!