Then came the war. In 2005, Microsoft launched Virtual Server 2005 (a rebadged Connectix product). In 2007, (open source) gained traction, and KVM entered the Linux kernel. But VMware had a three-year lead.
By 2020, VMware had over 500,000 customers and $11 billion in annual revenue, but growth slowed to single digits. The hypervisor was a commodity. The value lay in management and security. On May 26, 2022, Broadcom Inc. (the chip and infrastructure software giant known for aggressive acquisitions) announced it would acquire VMware for $61 billion in cash and stock. The deal closed in November 2023 after lengthy global regulatory reviews.
Today, under Broadcom, VMware is no longer a visionary leader but a cash engine. The name remains on products – vSphere 8, NSX, vSAN – but the soul is different. Yet every time a server runs 20 VMs instead of one, or a VM live-migrates without a hiccup, the ghost of that Palo Alto lab lives on. vmware inc. - display - 8.17.2.14
The reaction was immediate. Developers called it “sorcery.” For the first time, you could test a buggy kernel patch, crash the virtual machine, and simply restart the window. The host remained untouched.
The killer feature arrived in 2006: (VI3). It bundled ESX 3, VirtualCenter, VMotion, High Availability (HA), and Distributed Resource Scheduler (DRS). A single admin could now manage a thousand servers as one giant pool of resources. Wall Street took notice. Server consolidation projects paid for themselves in 6–9 months. Then came the war
But the execution was messy. Tanzu was complex, and customers complained of “confusing licensing.” Meanwhile, AWS launched (a joint engineering effort) – VMware’s olive branch to the public cloud, allowing customers to run their familiar vSphere environment on bare-metal EC2 hosts.
Then came the bombshell: In October 2015, announced it would acquire EMC (VMware’s majority owner) for $67 billion — the largest tech merger in history. VMware remained an independent, publicly-traded company, but Dell now controlled ~80% of the shares. But VMware had a three-year lead
August 17, 2016 – On the 14th anniversary of the first VMotion, Dell’s merger closes. Pat Gelsinger stands in front of employees: “Our north star hasn’t changed. We will run any app, on any cloud, on any device.” But behind the scenes, debt from the merger pressures VMware to deliver ever-higher margins. Part V: Multi-Cloud Pivot & The Kubernetes Gambit (2017–2020) The world had changed. Kubernetes had won the container orchestration war. AWS, Azure, and Google Cloud were giants. VMware’s on-premises dominance began to feel like a moat around a shrinking castle.
Each physical server—whether running Windows NT, Linux, or Novell NetWare—sat idling at 5% to 15% capacity. To run ten different applications, you needed ten different machines, each consuming power, cooling, and floor space. The industry’s solution was simply “buy more hardware.” Rosenblum and his colleagues, including Scott Devine, Edward Wang, and Edouard Bugnion, asked a different question: What if one physical machine could run many operating systems at once, safely and efficiently?
In a final irony, the date that once symbolized technical wizardry (first live migration) now marks a legacy of lock-in. Some engineers from that 2002 lab have left; others stay, maintaining the kernel of code that still runs inside data centers for 99% of the Fortune 500. Epilogue: The Virtual Legacy VMware did not invent virtualization – IBM mainframes had it in the 1960s. But VMware commoditized it, turning a mainframe luxury into a ubiquitous x86 utility. It enabled the modern cloud era, even if the cloud giants eventually ate its lunch.
Prologue: The Server Room Problem (1998) In the late 1990s, a small team of computer scientists at the University of California, Berkeley, led by Dr. Mendel Rosenblum (husband of Stanford professor Diane Greene), kept running into the same maddening problem. Their server rooms were graveyards of inefficiency.