Microsoft Volume II

The episode opens by challenging the familiar arc in which Microsoft rose with Windows, declined for years, and was rescued by Satya Nadella. The hosts' research points to a less tidy account: between 1995 and 2014, annual revenue grew from $6 billion to $80 billion, and the company became highly successful at selling enterprise software, even as many consumer products and initiatives disappointed. Their argument is not that the failures were imaginary, but that the standard 'lost years' label leaves out much of what the company built and obscures how uneven the period was.

Windows 95 is the hinge between the two volumes. Its consumer launch was a celebration of software, but the operating system initially lacked an Internet browser. In retrospect, its lasting importance was that it became the platform on which many people first used the Internet. That outcome was not obvious at launch: in 1992–1995, consumer access to online services usually meant entering proprietary, dial-up environments such as CompuServe, Prodigy, or AOL. These were controlled destinations, not an open network where anyone could connect a server and publish a website. Content providers needed a negotiated relationship with the service operator, so the emerging online market still looked like a set of gatekeepers and distribution deals.

Microsoft responded to that uncertainty by placing adjacent bets rather than waiting for one settled definition of the future. It explored acquiring AOL, then began Project Marvel as a proprietary online service that would eventually be repurposed under the MSN name. At the same time, Microsoft pursued an information-superhighway vision built around interactive television and partnerships with cable companies. The division of labor made sense given the assets of the moment: cable operators already had the pipes, household relationships, and television endpoints; Microsoft could supply software for set-top boxes and back-end servers; Silicon Graphics could provide powerful hardware. The hosts present this as a strategy that had worked before: enter a market with partners and investments that keep the company in the mix while the eventual shape of the market remains uncertain.

The interactive-TV thesis was not irrational simply because the open web later won. A networked entertainment system needed screens, controls, content, and connections into homes, and cable companies already possessed much of that infrastructure. Building a new device ecosystem and new delivery network from scratch looked less practical than extending what existed. The limitation was that the installed base also shaped Microsoft's expectations: the company was investing in a future mediated by existing television distributors, while the open Internet and browser were developing along a different path. This is the episode's recurring distinction between seeing a broad technological shift and correctly timing, framing, and executing the product that will define it.

Bill Gates's book The Road Ahead makes the change in understanding unusually visible. The hardcover, published after Windows 95 had shipped and the Netscape IPO had occurred, centered on the information superhighway. The later paperback revised the framing: references to that phrase were replaced with references to the Internet and web browsers, and the Internet index expanded from a few discussed portions to an entire page. The contrast captures how quickly the market's center of gravity became legible in hindsight. Microsoft's early activity was not proof that it had already understood the web; it showed a company exploring plausible routes into online computing, then revising its story once the web's importance became clear.

A January 1994 memo from Microsoft engineer Jay Allard argued that Mosaic, a browser spreading rapidly from the University of Illinois, could make the Internet an exponential phenomenon. His proposed response was to 'embrace, extend, innovate': accept the behavior users already wanted, then build it into Windows and extend it in ways that fit Microsoft's existing business model. That mattered because Microsoft sold Windows through OEMs to both businesses and consumers. The browser memo therefore linked a new user behavior to an established distribution engine; it was not simply a prediction that the Internet would grow.

Steven Sinofsky's experience at Cornell gave Microsoft another kind of evidence. He remembered the Internet as an academic network for exchanging research, but a snowstorm left him on campus while students used it to flirt, register for classes, message one another, and send personal email. His memo, 'Cornell is Wired!', captured a shift from a specialist tool toward everyday social infrastructure. Bill Gates later said that hearing Sinofsky describe Cornell made him take the Internet seriously. At an April 1994 leadership offsite, Gates concluded that exponential phenomena could not be ignored. The account connects a broad technical idea to observed behavior rather than to a mature commercial market: at that point, the Internet was still largely associated with academic communication, not established business use.

The same uncertainty affected even people closest to the technology. Jim Clark, frustrated that SGI was not moving more aggressively on the information-superhighway opportunity, contacted Marc Andreessen about Mosaic. Their first plan for Electric Media was not to build a web-browser company: they imagined making set-top boxes for interactive television, including a proposed partnership around Nintendo's N64. The hosts use this reversal to show how hard it was to recognize compounding potential in a network whose uses and infrastructure were still developing. Even Andreessen, whose browser was already helping make the web practical, initially considered the Internet a route to the more widely endorsed television vision. Early confidence in the general direction did not make the winning product obvious.

The shift came as usage made the browser opportunity harder to dismiss. Clark and Andreessen abandoned the N64 information-superhighway plan, raised money, and released Mosaic Navigator as a free download in October 1994. Their proposed business model separated the consumer browser from paid server software for organizations hosting websites. The hosts stress an important naming distinction: this commercial Mosaic used new code, despite borrowing the name and experience of the university project. Meanwhile, the University of Illinois licensed the original Mosaic to Spyglass, creating overlapping commercial claims that disrupted deals and led to litigation and a settlement. The episode's browser race was therefore not a simple handoff from one product to another; it involved different codebases, licensing rights, distribution channels, and business models.

Online-service companies also saw the web coming and tried to add browsers to their proprietary platforms. Netscape preferred to distribute its browser freely and sell server software, while Spyglass licensed the original Mosaic and pursued deals with service providers. Microsoft, having decided that the Internet should become central to Windows but lacking a browser, first negotiated with BookLink; AOL bought the startup before the deal closed. Microsoft then licensed Mosaic source code from Spyglass, which became the starting point for Internet Explorer. This sequence supports the episode's larger point: Microsoft recognized the strategic importance of the Internet early enough to act, but it entered through a series of bets, negotiations, and borrowed code rather than through a fully formed web strategy. The hosts also signal that the familiar public account of IE's origins is broadly right but incomplete, leaving room for further nuance.

Microsoft's browser strategy was broader than adding one application to Windows. A team working ahead of Windows 95 considered making the operating system deeply web-enabled: build HTTP into Windows, provide reusable HTML-rendering components, and let applications use them. Internet Explorer would be one consumer of those components, and Microsoft even offered Netscape the underlying code. Netscape declined, and IE became the main application using the Windows components. The hosts stress that Microsoft's license from Spyglass did not amount to simply taking Marc Andreessen's original browser and shipping it: Spyglass had substantially altered the code, Microsoft could not license the original NCSA version, and engineers spent considerable effort undoing parts of what they received. The origin story is therefore less a clean shortcut than a mix of integrated operating-system engineering and substantial follow-on work.

Gates's 'Internet Tidal Wave' memo marked a decisive change in priority. He called the Internet the most important development since the IBM PC and said it was crucial to every part of Microsoft's business. Netscape's rapid growth made the threat concrete: the company went from one million to 15 million users in a year and went public at a $3 billion market capitalization in August 1995. Marc Andreessen's prediction that Netscape would reduce Windows to a poorly debugged set of device drivers captured a platform-level risk. If developers could build rich applications for the web and users could reach them through a browser, Windows might become little more than the layer that starts the computer and connects its peripherals. Gates's memo explicitly recognized that Netscape wanted to move key APIs into the client and commoditize the underlying operating system.

The competitive irony was that Netscape could use Windows itself as its distribution vehicle. Windows was already installed on the vast majority of personal computers, giving a browser distributed as a Windows application a route to reach users at scale. At the launch of Windows 95, Internet Explorer was not included in the retail box; consumers could buy it for $50 in the Plus Pack. That did not answer the threat from a free Netscape Navigator. On December 7, 1995, Gates announced that IE would be free and bundled with every future copy of Windows 95. Netscape's stock fell by about a third that day and did not recover. The hosts frame the move as a distribution contest: Microsoft did not need browser revenue to defend Windows, while a platform owner that makes competing distribution a priority can quickly undermine a business that depends on that channel.

Microsoft extended that advantage through deals with AOL, CompuServe, and Prodigy, which replaced their browsers with Internet Explorer. The hosts describe IE's share rising from more than 20% in 1996 to more than 40% in 1997 and more than 60% in 1998, before it effectively approached the entire global browser market by 2000. Their account also keeps the later reversal in view: IE's share eventually fell back toward zero by 2010. The dominance was powerful but not permanent, and the episode treats the loss of that browser position as the next chapter rather than as an inevitable consequence of the first victory.