Steve Ballmer Speech Transcript – PC EXPO ’97

Remarks by Steve Ballmer, Executive Vice President, Sales and Support, Worldwide Business Strategy Group

MR. MAUND: My name is Bob Maund. I’m the associate publisher of Fortune, and before I get into my introduction of Steve Ballmer, I’d just like to take a second to mention that following Steve’s remarks this morning, David Kirkpatrick, Fortune’s technology editor, will be conducting a live Q & A.

Okay. Fortune is back for a third year to sponsor the keynote speaker series at PC Expo. It’s a tremendous opportunity for us to present the top executives in the industry to our readers. People like Michael Dell, Ellen Hancock of Apple, and Steve Ballmer from Microsoft. These are the movers and shakers that are watched and covered by our editors issue after issue. Microsoft is, of course, the powerhouse that has affected the way we all do business. From the operating system up, Microsoft has driven our need for more and more computing power on the desktop. We scrutinize Microsoft’s developments and alliances to understand the potential of technology as a strategic business asset.

Steve Ballmer met Bill Gates when the two were students at Harvard. Steve finished his degree there, then launched his career at P & G as a product manager. He was in Stanford’s graduate business program when he decided to join Gates at Microsoft in 1980. As executive vice president of sales and support, Steve drives all activities related to Microsoft’s sales and marketing worldwide. Today, nearly 8,000 of Microsoft’s 20,000 employees are in sales, support and marketing functions, all dedicated to Steve Ballmer’s vision of building long-term relationships with Microsoft’s customers. Ballmer, the number two man at Microsoft, also shapes the company’s strategy and business planning as part of the company’s executive committee.

Ladies and gentlemen, please welcome Steve Ballmer.

MR. BALLMER: I have to admit, I got a little nervous during Bob’s introduction. Now that I know he’s scrutinizing me and us so carefully, I’m sitting here shaking a little bit. I was a little worried earlier this morning about whether we would have the crowd we wanted. I figured at least half of New York is mourning the Met’s loss last night, and I guess you Yankee fans are all here. We appreciate that.


MR. BALLMER:It is my pleasure to have a chance to open things up this morning and have a chance to chat with you. What I want to try to do is take a little bit of a step back on some of — a little historical perspective, shall we say, on some of the important things we have tried to pursue, particularly with the Windows platform. Give you a little bit of an update on where we think that is as an industry phenomenon. And then have a chance to take some questions from David Kirkpatrick of Fortune Magazine.

If you look over the course, certainly, of the time we’ve been in business, there have been perhaps five — I’m not sure if the word “visions” are correct, but five articulations, five things which Bill Gates, Paul Allen, and other managers have believed in passionately that have really guided our view of what we need to do as a company, and what we need to do as an industry. The company was founded on this notion that the PC model was very different than the mainframe and the minicomputer model. Hardware and software and services would be separate businesses. There would be specialists. And so the whole shape of the industry changed from one of vertical integration to specialization.

About 10 years ago, we started talking about something called Windows Everywhere. I want to give you an update on that today. But Windows Everywhere, when you really stop and think about it, was pretty silly in 1987, 1986, 1985. We talked about it in there a lot of times. But here, by the year 1997, I think we can really talk in quite an important way about this notion of Windows Everywhere. Softer software, the notion that software will adapt and shape and customize itself to you. We’re not going to talk about that today, but it’s another one of these guiding principles. Information at Your Fingertips, a vision that Bill Gates talked about starting about seven years ago has been very much a guiding force for us, and we think for others. And of late, we’re really pursuing this notion that IT can form almost a digital nervous system, acting as the backbone of the way corporations operate.

For today, I really want to dig into Windows Everywhere. I want to talk about it. I want to update you on it. And I want to give you a little bit of a perspective of some of the things that we’re trying to continue to invest in to make that possible.

Windows everywhere is an investment that we started really with the notion that said, if we can take the Windows platform and deploy it quite broadly across a very broad set of devices, we would give developers a platform that was very rich and very appropriate for the broadest set of applications. The initial investment, of course, was in PCs, then in laptops; with Windows NT it’s graduated to become in servers. And today I want to update you on that investment. I want to talk a little bit about where we’re taking this notion of Windows Everywhere in terms of hand-held devices, high-end workstations, servers. Particularly here today at PC Expo, I want to talk about manageable PCs, because we have a lot of industry partners introducing the first NetPC, central administration, Windows support for really dumb terminal devices. And Windows on what I’ll call Windows, Windows on UNIX and Windows on Mac. And I think that will be clearer to you as we go through the talk.

When we’re finished with the kind of work that we embarked upon, really starting 10 years ago, we hoped to fill out this picture quite broadly. From low-end devices, Web TV-style devices at the very low-end, consumer devices which you will find attached to every television set, which you will find attached to every DVD player, every low-end consumer electronics device, on up through hand-held PCs that will be functional both at home and in business, to Windows terminals, NetPCs, portables, desktops, workstations, servers.

We had a chance about a month ago to talk about some of our initiatives to continue to push Windows to the super high-end with our scalability efforts behind the Windows NT server. Filling this picture out, I think, is very important for us, for our partners, and for the customers who want to mobilize this platform in the broadest set of ways. It’s very important largely because of what it means to software developers. Software developers in many, many ways are really the lifeblood of our industry. They provide the innovation, the excitement, the applications, and I’m not talking today just about the developers that live in third-party software companies. But really the developers that live also inside large companies, small ISVs, custom application development houses. These people are creating the value that lets businesses run more effectively.

Developers need two things, or at least this is what we hear loudly and clearly. They want the best platform in a technical sense to build their applications, and I want to talk about that. And they want the widest set and the easiest set of deployment options for those applications. Without a lot of machines to run applications, software developers have a very hard time. When it is harder to deploy applications, software developers have a harder time.

But the first thing they need is the tool set. And at Microsoft we put a lot of efforts in Windows to creating the richest, most mature set of services for developers. We talked a lot about the Win 32 API, the Windows 16-bit API, both of which are still supported. On top of that today, there’s literally been thousands of companies and hundreds of thousands of developers across the world who have built third-party libraries and objects which essentially extend the Windows core service set. If it’s not built into Windows, you can go get a third-party library that will put up a grid of numbers very conveniently that will give you access to transaction processing services, that will give you access to groupware services.

And so you amalgamate all of the exciting foundation that has been built both at Microsoft and in small third-party companies, and in custom application development companies, and in companies like IBM and Oracle and many, many others. It’s really a phenomenal, phenomenal amount of innovation.

So this rich set of services is incredibly important. And you couple this rich set of system services with the most modern things, the most interesting things going on. There’s a lot of discussion in the industry today about Java. So we’ve taken that to heart and integrated into Windows fully system support for so-called “100 percent pure Java” in the Windows products. But with the announcements that we’ve made over the last several months, and particularly yesterday of this thing we call J Direct, we also allow developers who want to work in Java to take advantage of not only the full Windows set of system services, but the full set of services that have been built up in third-party libraries and objects, et cetera. Developers need to write applications quickly. They need to write applications that are as good as the best of the best. And we’re allowing people to mix the best of what is coming with all of this huge body of existing Windows system services. The fastest application, the best application then can be built right away, no need to restart.

The Windows platform supports a very wide set of development languages, BASIC, COBOL, FORTRAN, SmallTalk, C, Java. We think that’s very important because it is not going to be the case that one language will fit all application development scenarios. We put a lot of effort to make sure that we provided the best environment for developing Java applications including our own Visual J++ product, because we want in no way the Windows platform to be anything but the best platform for writing applications in absolutely any language. Already we’ve found incredible interest in our Visual J++ product. We think we sell well over half now of all of the Java development tools sold in the market, and we’ll just continue to push that product forward and enhance that product in a broad set of ways.

So the goal is to build up this set of services, and then give developers the chance to deploy these applications very broadly. Windows is a very popular platform. Over 90 million units of Windows will be sold in the next year. The installed base of Windows machines is close now to 400 million. The installed base of Windows 95 and Windows NT Workstation is over 200 million. We need to, in our Windows Everywhere initiative, continue to broaden out the places in which these Windows applications can be delivered, more platforms serving more market opportunities, more niches. We need to provide facilities that make it easier and easier for developers to deploy these applications, and we need to make it possible for people to target less and less expensive hardware that is less and less expensive to manage.

And so I want to focus in really on that aspect of Windows Everywhere, how we let these developers, these over four million developers around the world today, take advantage of Windows Everywhere to take their best work, deploy it cheaply, quickly and in more and more places.

The most recent member of the Windows family of products of this Windows Everywhere vision is something that we call the Windows CE operating system. It’s a small portable, real-time operating system, runs on multiple chip sets already today. And it is Windows compatible. It is a subset. It doesn’t support the whole Windows API, but it is Windows compatible. There have already been devices introduced from the people on this list, and now a number of others, that deliver the Windows CE operating system in the context of hand-held machines. There are people working on quite a variety of new Windows CE form factors, and Windows CE applications. People want to do these hand-helds, but they also want to do consumer devices. We’re taking a look at our acquisition of Web TV for using Windows CE in that device. I’ll talk later about the use of Windows CE in so-called “Windows Terminals. “We see being able to use it quite broadly in a variety of low-end devices.

There are over 3,000 independent software vendors today building for Windows CE. We thought this thing would start out as just a little pocket-held device that sort of mobile professionals would carry, but the level of interest that we’ve seen in Windows CE from people in vertical markets is, frankly, phenomenal. People in the retail community who want to take the little devices that the retail sales clerks carry with them to monitor inventory and convert those to be Windows CE based. We’ve had ISVs in the health-care field talk about wanting to do Windows CE devices for doctors and nurses. We’ve had people in Japan talk to us about wanting to use Windows CE in hand-held devices that would be carried by utility linemen and repair people as they go out and work physically in installing electrical lines, et cetera, et cetera.

So the variety of applications and interest is quite broad and, frankly, opens up to the Windows developer a whole new community of applications, of interest, and market opportunity. You will see us not only continue to bring out new versions of Windows CE in the hand-held form factor, but really push into these verticals and really push more extensively Windows CE into quite a variety of consumer electronic devices.

I’d now like to invite Jon Roberts to come on stage and join me. Jon runs Windows marketing, and he’s going to show you one of the applications that we’ve been working on with Canada Trust in the on-line banking field with Windows CE.


MR. ROBERTS: Well, as Steve mentioned, this is actually a very small device. It only has two megabytes of RAM, four megabytes of ROM. I’m going to turn it on real quick. You’ll notice that it comes on instantly. If you look on the screen, you see a lot of familiar applications. We have about seven applications on this. Each of these applications only takes about 100 to 200 K. So these are really cut-down applications. And what I’m going to do is launch the Internet Explorer real quick.

[Mr. Roberts inadvertently launches the application several times, creating a memory conflict. Given limited time to restart the demo, he simply explains how the application works. ]

MR. ROBERTS: This is an app that Canadian Trust is using. It’s called Easy Web. And essentially from any Windows application, any Windows browser, including this browser, people can go up and do on-line banking. And so this is just one example of a companion device that they’re using. A couple of others, as Steve mentioned, one is Renault in France, and they’re using it as a pocket device to take orders of cars, upload it so that they can process how much it would cost and process the order.

You mentioned baseball earlier. In fact, the Seattle Mariners are giving it to their talent agents to go out and take notes in terms of who they want to recruit and this and that. So you’re really finding it to be an excellent companion device and I might add, on a personal note, a difficult device to demo. Thank you.

MR. BALLMER [to Mr. Roberts]:Strike one.

[To audience:]A particular area in which this Windows Everywhere strategy has taken us, and really one, frankly, that has progressed far more quickly than I think we had anticipated, is the penetration of Windows into the workstation space. And that was really precipitated by the launch of Windows NT Workstation and the incredible price-performance dynamics that a high-end PC running NT Workstation provides, versus what people can offer in the UNIX workstation space. There are some numbers here that give you a sense of the price-performance comparison of a Compaq workstation, versus a Sun workstation. [Slide shows a three-to-one price-performance advantage of Windows NT over Sun. ]

But, what we’ve seen come with that is just an incredible surge in application developer interest, an ability to target Windows applications to the high-end workstation. If you take a look, for example, at the top ten CAD/CAM applications, graphical information systems, those applications are all now today targeted and available on the Windows NT Workstation product.

Other important CAD products, for example used in the design of most major aircraft around the world, is just now in the process of being delivered on the Windows NT system. If you take a look at the visualization space, the kinds of work that has been done on Silicon Graphics Workstation, digital image editing, the kind of work that’s been done for visualization here on Wall Street for financial trading applications, that application set has now moved over.

If you take a look at some of the other applications that have been done in the manufacturing space, they’ve moved over. Oil and gas exploration and geographical mapping in that industry, those applications have moved over. And from the standpoint of both the developer and the person deploying these systems, you now can have the price-performance of the PC, you can have this full range of Windows applications, and you can continue to run the hundreds of thousands of existing Windows applications. So we see people now able to target a broader set of applications and reach out to new places with the Windows API set.

This has happened, as I said, far more quickly than we ever anticipated. The UNIX workstation market, amazingly, has actually declined, actually declined in 1996 versus 1995. We don’t think that trend will stop. The price-performance economics of Windows NT Workstations versus what’s available for UNIX and RISC is too compelling. And Windows NT Workstation now has the full set of capabilities that developers in these areas need. So this notion of Windows Everywhere is starting to extend, or has extended itself, on up into the workstation business, where we think Windows NT Workstation will outsell UNIX workstations now by over a million units a year. As I said, the UNIX workstation market actually declined last year.

We talked a little bit about the server in an event that we did here in New York a month ago scalability day. I’m going to talk a little bit more about the server market. We’ll sell over a million Windows NT Servers this year. And we’ve seen a phenomenal surge of interest in developers who want to have access or sell in new places Windows server-based applications.

If you take a look at the work going on at companies like BAAN or SAP or PeopleSoft, there has just been an incredible shift to the Windows NT platform. Almost 50 percent of new SAP installations today go on top of Windows NT as opposed to UNIX. A few weeks ago at the Supercom show we talked about the growing range of interest in Windows NT Server from telecommunications software vendors.

We see people do SS7 switching on top of Windows NT Servers. We have a project that was announced with Octel, where they’ve moved their voice-messaging system over on top of Windows NT Server. We have projects going on with people like Ericsson, where they’re bringing their PBXes to use Windows NT Server. These are markets typically that were targeted before either by UNIX or proprietary operating systems. But, with the growth in capabilities in the Windows NT Server platform, people are moving over.

Nine of the top ten database price-performance records have been set on top of Windows NT Server. In the intranet space, over 60 percent of the new intranet servers and about 55 percent of the new Internet commercial servers are going in on Windows NT. We’re working hard to continue to improve the scalability and availability of Windows NT, so that people can target an even broader set of deployment possibilities with their applications. In order to do that, if people want to configure increasingly complex and large data centers with Windows NT, we’re working with third parties who have data center management tools. And we continue to add features to the core set of services provided by Windows NT at the server level.

We’ve announced routing and RAS capabilities recently, directory, security options, transaction services, message queuing services, active web page services. Again, we’re trying to give developers a richer set of services, and give them the ability to target an even broader set of deployment possibilities.

The NetPC, in some senses, at least at Microsoft, we’ve joked this ought to be called the NetPC Expo, with the range of support that we see from vendors, hardware vendors, at the show for the NetPC. People often ask, though, what is a NetPC?A NetPC, in some senses, is just a PC which has been done, some people might say right first and then we’ve also tried to take out a set of features which are simply overkill for a wide variety of users. So we’ve taken out some of the things which have provided complexity and management overhead. We and our industry partners, Intel, Compaq, Hewlett Packard, who have been our lead partners in this effort, we’ve taken things out and we’ve tried to add simplicity.

In addition we’ve added some things here in the short term to try to make Windows systems in general more manageable. Not only NetPCs, but mobile PCs, regular PCs, and those thingsgo under the name of our Zero Administration Initiative. And specifically we want to show you some things that we already have available today to make NetPCs and other PCs more manageable. So there’s a variety of things that we’ve done to simply do the PC right and scale it down for a variety of low cost deployments, that we think will be incredibly valuable.

We’re joined in that belief by a number of hardware vendors. In our partner pavilion here at PC Expo you can see NetPCs, from Compaq, Dell, Hewlett Packard, IBM, Gateway, NEC, Seimens Nixdorf, Digital, Unisys, Hitachi, are all here in our booths showing NetPCs. It’s not an exotic, sexy, wild demonstration. It’s just a nice, simple, low-cost PC. And we think that that is a valuable addition to the set of deployment options that developers can see for their applications. It’s a lower-cost device, in terms of acquisition cost. And it’s a lower-cost device in terms of the management overhead in supporting users on those PCs.

We’re working very hard on this broader initiative that we’ve called our Zero Admin Initiative. And at the heart and soul of what we have to deliver, we will deliver concurrently with Windows NT version 5. Windows NT version 5 goes to beta here in the fall and we hope to ship that product roughly a year from now. And the vision of our zero admin option is to simply create a world in which administrators who want to have no ongoing need to touch clients. That is to take the cost, all of this management and people overhead cost in deploying applications, out of the system.

The core breakthrough that we’re focused in on, in addition to some of the things we’ll show you already available today, allows the administrator to lock down and secure from the user’s wandering hands, all of the resources of the system. The key initiative is the ability to store all of the state, all of the user’s configuration information, data, et cetera, on a central server and have that automatically in a smart way get mirrored out to every client. We sometimes joke around, it’s kind of an intelli-mirror kind of concept. I don’t know what name we’ll wind up with, but to mirror things out to all of these clients automatically.

It’s facilitates, if the design is right, roaming. So users can move from machine to machine. Very important, for example, in a customer-service environment, or a retail environment. It makes it very easy to replace PCs. We’ve got a customer we’ve worked with, a bank in Australia, whose number one goal in their mortgage processing center is when a customer-service rep’s system goes down, to just be able to shuttle in a new system and have the rep back and running in 25 minutes. That is a key target, and it’s only possible if all of the information, all of the data, all of the profile information about that user and what he’s doing, is cached up on the server and then just replicated out. You can’t be installing things on the client. So it’s a zero install world.

We’re trying to make sure we permit disconnected operations. So if you do have users who want to take a mobile machine with them, we can intelligently cache everything that’s necessary on the laptop before it gets undocked and try to provide them the benefits of the PC, with the benefits of centralized computing. One model will not fit all. I don’t think all applications in all corporations will be centrally administered. But, we do know that this is an opportunity for us to give developers and the IT professionals that they’re working with a lower-cost way of deploying these applications that add so much value.

I’m going to ask Jon to come back on. And with a little bit of luck, do a demonstration for you of a NetPC, using some of our Windows NT 4 and Windows 95-based zero administration options. And then we’ll show you a little bit of Windows NT 5, and what we hope to do in terms of this smart caching or smart mirroring. We have here a NetPC from Compaq, who along with Hewlett Packard are the two lead system vendors working with us on the NetPC initiative. But, I’ll let Jon go ahead and show you the machine and the demonstration.

MR. ROBERTS: A little bit of Vanna White first. First you’ll notice on this machine there is actually no floppy drive, no CD, there is no way that a user could put any software into it or take anything out. By design it’s a very sealed case that’s trying to cut down on the cost of user futzing, if you will. Inside, as Steve mentioned, we’ve made it a simple, more manageable PC. There’s no isobus. It’s going to support ACPI, it’s going to support wake-on-demand LAN, so that you’ll be able to administer it, even when the machine is shut down.

[Mr. Roberts shows a simple browser-based interface. Even if the user shuts it off, the browser will come back on, centrally controlled by Windows NT 4. 0. ]

MR. ROBERTS: We’ve got Java scripting up here, active objects, et cetera. This is actually an application that Turner Broadcasting is using, and they do it for order processing. I’m simply going to select order entry. It’s just a very, very simple app. It’s going to come up in the browser, and it’s just an example of a sort of single-task function that a company might want to do in a very, very locked down configuration.

So I’ll go to agency, select an agency, go to buyer and select a buyer, advertiser, et cetera, click on what it is I want to do, sports, ad, submit. You know, very, very basic stuff, simple tasks.

But what’s interesting is because this is an IE 3 robust browser, you can actually do a fair amount within the browser. I can go and allow this user to reference and access various information resources. This could be an internal home page so they could get access to different applications, different documents. I’m simply going to launch this, and you’ll see that I’m going to be going out and actually launching PowerPoint. It’s there, I can click on it, go through whatever I want. And, in fact, you’ve noticed that the menu has changed, edit, delete, et cetera. So these users have some capabilities. But, as Steve mentioned —

MR. BALLMER:This was all set up by the administrator. If the administrator didn’t want to let them play around, he wouldn’t have to.

[Mr. Roberts shows a Task Station view, again using the browser, to launch Microsoft Word from a central server. The only user option is to store the file on a server. Then he logs in as an administrator and shows how additional applications can be added for users. These could be applications on the server or on the user’s machine, depending on the administrator’s wishes. ]

MR. BALLMER:Okay. So all of the things you’ve shown us in terms of the ability to lock down the desktop, disable the start button, force saves to a given location, disable or enable just a specific set of applications to run, that’s all part of this Zero Admin Kit, which people can get today for Windows NT 4, or for Windows 95?

MR. ROBERTS: Yes. Within about a week we’re going to release it for Windows NT 4, and then about a week or two after that for Windows 95 on the client.

MR. BALLMER:And, of course, that runs on NetPC.

MR. ROBERTS: NetPC or any PC. Okay, and now, joy of joys, we’re going to the complicated demos. What I want to move is fast forward to Windows NT 5. And I will say that this is the first time this has been demonstrated publicly, and I’m pretty excited about that.

As Steve said, we’re going to take what we’ve done with Windows NT 4 a dramatic step, and really give you the best of the central server and try to maintain the flexibility of the desktop. So all that you see here is a desktop. And we simply want to show how we can add more control to that.

I’ve now switched to the server machine. I’m going to grab an object that I want to place on that desktop, drag it down to the server, and then what I want to do is switch back to the client, and you’ll see that I dragged the handbook down, and it’s now been added to the desktop. So I’ve just done that centrally. So I could do more than simply add things to the program menu. I can place anything on the desktop. And if the user were to delete that handbook, the user doesn’t have the option, because just like that program, it is centrally controlled. So the handbook comes back up. So it’s more extensible control over the whole system.

[Mr. Roberts shows an upcoming feature that will prevent users from installing applications on client machines, installing applications on servers and downloading to client machines only the minimum code necessary to run the application. ]

MR. BALLMER: Project is still entirely, including my profile information, my preferences, it’s all still up on the server. All it did was build a configuration for me, and then stream down appropriate bits for enough cache in memory on the machine?

MR. ROBERTS: Exactly. And so now I can click this, and I’m going to be going into installing Project now on this machine if I wanted to. And the administrator could do it so that I would simply be accessing it up on the network. So this is streamlining network install of applications. Part of what we’re trying to get with the sort of thinking of intelli-mirror.

[Mr. Roberts shows that he can take files on his network drive and create a mirrored copy for the client machine and work on all the files offline. ]

MR. BALLMER:So what you showed here was, despite the fact that all of the data and all of the state is up on the server, you can replicate it down to a laptop when you want to disconnect, and it just uses all of the smart caching and replication in Windows NT 5?

MR. ROBERTS: Right, exactly. And, as Steve mentioned, this is very, very exciting, when you think about a mobile user. You’ll be able to roam from machine to machine, dock, and get your user state. We would be able to destroy this machine, which I’ve been inclined to do at recent points in the past, and plug it in and get it automatically returned to you. So this is what we hope to reduce costs with Windows NT 5.


MR. BALLMER:Not all administrators want to configure everything in the same way. And so one of the key goals has got to be for us to allow not only a rich set of applications to be built, but we have to have a wide set of deployment possibilities. Everything stored on the server, things cached down on the workstation, things locked down, things not locked down. If you talk to people in a variety of industries, even take something like telecommunications. They have these huge customer-service centers, where people really want to lock down exactly what the user can do, no access to applications, no ability to get themselves in trouble, because the customer-service center has to always be up and running. That same telecommunications company, though, quite likely has a group, a very large group, of engineers who absolutely wants to be able to twiddle and touch and control and launch and do everything with any application on the system. And we’re trying to provide the full range of tools to allow deployment in an organization that is as heterogeneous as almost every corporation in this room I’m sure is.

The next thing I want to talk about in this Windows Everywhere strategy is something that we call the Windows Terminal. And the Windows Terminal is truly a thin client. This is something we’ve just recently started talking about. We hope to have partners introducing Windows Terminals the first half of next year. But a Windows Terminal is truly a thin, thin, thin, thin, thin client. Not just a thin client, it’s thinner than that. And it is in some important ways technically. Our goal with the Windows Terminal is to have it work essentially like a dumb terminal. That is, the server simply sends it strings of information that it puts up. And not only is all of the storage central, as we saw in the Windows NT 5 demo, but all of the processing is done centrally.

It’s actually a little different, for example, than the model of the network computer where, in fact, you have to run a big old browser, and Java applications, and everything down on the client. It’s actually a less thin device. This device is trying to do a lot less than even the so-called NCs by recentralizing a lot of computing. You could have access on the thin, which we hope will be about $500 or less, terminal devices. They’ll run full Windows applications. They’d run Java applications. They’d run any Web content. But all the processing happens at the server, and the screens of information are then painted on these $500 clients.

We’re working with a broad set of suppliers. Current PC companies, as well as companies that are in the terminal business today. We expect that there will be some software that lives permanently down in these terminals, and that that will be a Windows CE-based client, so that you have all of the architecture to download these screens being based and running off of Windows CE. And these things will require an add-on to Windows NT 5 that we call Hydra, as a code name. But it essentially builds from the technology that we’ve licensed from Citrix and from Prologue to bring multi-user, dumb terminal support capability to the Windows NT Server environment.

Do I think everybody in the world will need Windows-based terminals?No, I do not. I think it is more a replacement for today’s terminal market, which has shrunk over time. Perhaps some applications which are on PCs today will be deployed on Windows-based terminals. But the vast majority of workers will still want to have and need to have local intelligence, local processing power, in a way that makes this, I think, an important entry to the Windows Everywhere strategy but doesn’t wipe out the rest of the PC market.

Also with our acquisition of this multi-user technology from Prologue and Citrix, we’re embarking on something — a part of the Windows Everywhere strategy that I call Windows on Windows. That is very important to developers. How does a developer who wants to target the most modern set of Windows services target old Windows 3. 1 machines, old perhaps NT 3. 51 machines?In the future, how will somebody who wants to target Windows NT 5 get those applications to run on Windows 95?

The answer builds from this same Hydra technology that we’ll use to support dumb terminals. So that if you have a system that’s doing a perfectly good job supporting a user, but you want that user to run a new application that is targeted at a higher end version of the operating system, or higher end hardware, they can through this multi-user emulation.

This is incredibly important to the IT people I talk to. This essentially allows people to think about how they extend the life of a PC. I think, over time, people will want to upgrade PCs, as they do today. But if part of what we get as a result of this Windows on Windows initiative extends the replacement cycle on PCs for some people from two years to three years, or three years to three-and-a-half years, or three years to four years, not every PC, but extends the life of a number of PCs, this will be another valuable element in letting developers continue to target and take advantage of the most modern technology, but still have a low-cost deployment option back into the install base of Windows systems.

If you couple those two things, Windows on Windows and Windows Terminals, along with what I’ll call Windows on UNIX or Mac, through this multi-user technology, we’ll be able to let people who do own Macintoshes and UNIX workstations also run Windows applications. You really do get a world that is more heterogeneous in terms of the deployment of Windows applications, and lower cost for management in some cases, and certain can extend the lives of existing PCs.

I want to invite Jon back for one last time to just give you a little bit of feel, maybe not even in this case for Windows on Windows, but Windows on DOS. And for those of you who have seen the demonstration from Citrix or Prologue, it’s not going to be that amazing to you. But for people who haven’t, you’ll see how an existing, what is it, 386SX system, two-megabyte MS-DOS machine can run Windows 95 applications.

MR. ROBERTS: And this has got a 40-megabyte hard disk. And, in fact, we have thousands of these, so if you want a deal, we can negotiate afterwards. But what I want to show, actually what I’m running here, for the first time showing publicly, is Citrix’s new ICA protocol, which they’re code naming Picasso. They announced this morning, which is allowing us to use this Windows NT 4 configuration.

[Demonstration of older PC running up-to-date versions of Microsoft Word and Excel. ]

MR. BALLMER:But you can see, we actually get not as fast performance as many users would want. But if you’re simply trying to give access to a user to one new application who is perfectly content with everything else going on on their system, the terminal emulation capability, really multi-user capability, does really give you the ability to extend the life of old PCs, which frankly is probably the number one thing that IT people talk to me about — how do they extend the lifecycle of existing systems and existing hardware.

I want to just make a contrast a little bit here at the end of what we’re trying to do with Windows Everywhere versus what Sun, Oracle, and IBM are trying to do with the Network Computer. The Network Computer, I would say, is part of a whole theme from these guys to try to undo what the PC revolution has done. The PC model, this notion of hardware and software being separate businesses. The notion of building up libraries of compatible software which can be deployed very broadly. In some senses, people from Sun, Oracle, and IBM are trying to say, no, let’s start again. Let’s go back and re-pursue the UNIX myth. Let’s try to undo what’s already happened in the PC world.

But this NC thing smells a lot like UNIX to me. The NCs from Sun, Oracle and IBM are simply not compatible. They’re not. These are companies that, in some senses, all compete on being different, and their machines are not compatible. A browser in a Java run-time, it’s not thin. A browser is one of the largest applications being written today, and the notion that that is a thin client, for people who really want thin, is simply misshapen. A terminal-based approach gives you a much thinner client.

There’s been a proposal essentially to rewrite all applications for this world. To not use the applications and libraries of codes that have been built up. But there’s no obvious new benefit in terms of applications that you can’t write today using Windows, the Java run-time in Windows, the Java VM, the Java development tools, anything you can do today in the Windows environment.

The whole approach here overloads servers and networks. So, by the way, do Windows Terminals, but that’s another option in the Windows family, it’s not the only option in the Windows family. There’s no NC support for portable, printing, and expandability. There are different programming models. There’s the Java model at the client, there’s the NC model, there’s the HTML model, and there’s still good old UNIX running up on the server, and it’s fairly poor coexistence with existing PCs without technology like the terminal emulation that we saw today in terms of Windows clients.

Now, in some senses, I joke that this Windows Everywhere is part of our
“write once, run everywhere”
strategy. It’s a very different kind of strategy than the one Sun is pursuing, but it certainly talks to the same core idea. And that idea is letting developers get the broadest set of deployment opportunities for their applications, but we’re not starting from scratch in Windows. As I said, there’s a huge installed base and a huge Windows run rate. There’s lots of code libraries. The glass here is half full. Windows has this rich set of services, and we’re extending into new markets, and we’re trying to take out some of the prohibitive costs and a problem for people in terms of deploying Windows applications.

We’re pretty excited about Windows Everywhere. We think it was the right strategy in 1987, and with all of the momentum behind Windows today, all of the progress we’ve made in hand-held workstations, terminals, servers, et cetera, we think this is a glass that’s, if not half full, a lot more than half full. And we’re excited to be continuing to pursue Windows Everywhere, the richest platform for developers with the broadest set of deployment options.

I thank you very much for your time this morning. We certainly enjoyed it.


MR. BALLMER:I’d like now to invite on stage with me David Kirkpatrick from Fortune Magazine. David tells me he’s got a number of questions.

MR. KIRKPATRICK: Thanks, Steve. That was really amazingly interesting, as usual. How thin are those Windows terminals now?

MR. BALLMER:I think probably the best way to measure their thinness will be in their cost. And we anticipate with the ROM that they need, they just have some ROM and a very little bit of RAM, they should be under $500.

MR. KIRKPATRICK: Well, you know, Andy Grove was at Fortune recently and he said to us that if that happens everything Intel is planning, you know, is in trouble. Do you see it that way?

MR. BALLMER:No, I don’t. I don’t think these terminals replace every PC. Some people have that view of the world. The terminal market worldwide is still over 2 million units a year, dumb terminals. I’d like those all to be Windows dumb terminals. That’s 2 million people who, for whatever set of reasons, still haven’t chosen to go with the PC approach. And we think there is an opportunity to extend the Windows architecture into those terminals, without materially affecting the PC business.

MR. KIRKPATRICK: So the growth rate of the PC business is probably likely to remain more or less as it would have been otherwise, in your opinion?

MR. BALLMER:Yes, I don’t think Windows terminals will effect that materially.

MR. KIRKPATRICK: Okay. Let me ask about another thing that might effect it, which was news to me, was this whole Windows on Windows approach. I can’t believe the PC companies that are all represented here today are totally thrilled about the notion of new applications running on old machines. How do you think about that and tell me a little bit about what your discussions with the PC companies have been like over that?

MR. BALLMER:Well, certainly there’s a possibility that that would have some impact on replacement rates in the ways that we described. I think there is so much great innovation and relatively, I’m much more excited about what we call knowledge workers than most people are. I think most U. S. PC users are knowledge workers. They are people whose companies value their creativity, their insight, and their wisdom. Even a customer-service rep, you want that person thinking, alert, aware. These are not so-called “task workers” even though people treat them that way. And there will be an ongoing investment in PCs for the lion’s share of workers.

There are cases where people say to themselves today, I’d like to deploy a new application, but I can’t because I can’t afford to upgrade this PC. If people can afford to upgrade, they will. But, for all the IT people out there who have an application that they’d love to deploy but can’t think about it, because they don’t have the budget to buy a new PC, this sort of gives another option. And I’m more excited in terms of what it gives developers than what it takes away from PC manufacturers.

MR. KIRKPATRICK: So the PC makers are buying that view?

MR. BALLMER:What shall I say. I won’t say everybody is turning cartwheels and saying, this is the best thing since sliced bread. But, I think certainly the technologies are available in the marketplace today from guys like Citrix and Prologue, and yet the PC market continues to grow.

MR. KIRKPATRICK: So you see it as just one more way to expand the market, in effect?

MR. BALLMER:The opportunity for Windows application developers, that’s right.

MR. KIRKPATRICK: Could it potentially have a role even in the Third World, I mean, when you have all these countries that simply can’t afford the latest Intel hardware. Is that something you’ve given any thought to?

MR. BALLMER:Yes, actually Third World countries don’t think that way at all. They don’t ever understand why people in the U. S. think they should get the cheap, low-end technology. They want top of the line Pentium 2. I mean, that’s all you sell in China these days.

MR. KIRKPATRICK: China is Intel’s highest-quality processor market they say, the highest-speed market.

MR. BALLMER:There is no doubt about that — especially in countries that didn’t make investments in mainframes and minicomputers. In some senses their investment in PCs tends to serve higher-end needs and they make investment in high-end gear.

MR. KIRKPATRICK: So this will not help them particularly. It’s really more for the corporate —

MR. BALLMER:No more and no less than anybody else in the world.

MR. KIRKPATRICK: Okay. Talk to me a little bit about this Comcast deal. I think a lot of people were really surprised to see Microsoft buying into the cable industry. What are you thinking is the goal there?

MR. BALLMER:Well, certainly one of the big opportunities and frontiers for PCs is increasing household penetration. Only about 40 percent of U. S. households have PCs today. We think that number will ramp up some. But, there’s two things, how do you ramp it up even further and get it more into what I’ll call the consumer space?Part of that will come with low-end devices, like the devices from Web TV, connected to TVs. But, the only way that will make maximum sense is if the cable companies and the phone companies are investing enough in their digital service infrastructure.

We thought our investment in Comcast would help, what shall I say, re-energize investment in cable infrastructure and phone infrastructure. And certainly if you’ve seen the surge in cable company stock since your investment, it seems to be working about as we had hoped. And that should help bootstrap the consumer system software market for us.

MR. KIRKPATRICK: We were talking about that backstage. I guess you weren’t amazed, but I was amazed to see cable company values, across the board, even those of my own employer, Time-Warner, go up a considerable percentage, simply because Microsoft invested a relatively small amount of money in the industry. But, you were saying that it was analogous in your mind to the investment you made in UUNet some years back, where you felt that that investment kind of helped bootstrap the entire ISP market.

MR. BALLMER:Right, exactly right.

MR. KIRKPATRICK: You know, I guess it goes to the next question, which is a little bit ticklish, if you’ll forgive me. You know, when you talk about Windows everywhere and then you look at something like the kind of power you’re just acknowledging Microsoft has to affect values in industries and the entire way we perceive of potential of certain industries, some people find that a little bit scary, to be frank. I mean, I’m sure you’ve heard this.

MR. BALLMER:All news to me — no, go ahead.

MR. KIRKPATRICK: I’m curious, you know, some people say, well, benign dictatorship, I mean, talk to me about this issue — you’re going for such a broad set of influence and power and market penetration. What should lead us to be confident that once Microsoft gets this kind of influence that it will continue to work for our benefit for the long-term?

MR. BALLMER:Well, I mean, the market is the ultimate pressure point. Seriously, if there are holes in the Windows product line, there’s always a competitor just jumping up to jam home a new device or a new approach into that. Whether it’s some of the things people tried to do with hand-helds, whether it’s the things people are trying to do today with NCs, you know, there are plenty of fine companies out there, whether it’s startup companies like Citrix and Prologue, anytime we leave space there is somebody who is going to really feed off of that.

And the key thing for us to do is to understand what’s going on, understand where the holes are and the holes aren’t there because people — you know, sometimes we miss things, sometimes even with our investment budget we can’t get to everything. But, the fact of the matter is, those holes are important, they’re competitive opportunity and we do tend to respond. The competition very much, I think, does what it’s supposed to do and keeps us working hard and investing aggressively and extending the range of opportunities for Windows developers.

MR. KIRKPATRICK: I’m not saying that — I think you are getting that market penetration because you’re providing products that people want, no doubt about it. I was talking to someone at Microsoft not too long ago and asking how Microsoft viewed this question of, kind of, you’re just amazingly a continual ramp of success. And this person was saying that, you know, it isn’t always that Microsoft is so incredible, but that a lot of our competitors have really sort of stepped on their own feet. I mean, when I think of Novell and you could say Lotus in some ways, Borland, do you think of it that way, that in a way you’re a beneficiary of some pretty dumb moves on the part of your competitors?

MR. BALLMER:Dumb, sure, sure, dumb. You know, any success has three aspects. You’ve got to do some good work, you’ve got to be lucky and to some extent the other guy has got to do either bad work or less good work than you do. So there really are three elements and we’ve been the beneficiary of probably some wrong moves by competitors, as well as our own good work and just some good fortune.

MR. KIRKPATRICK: One sort of general question, too, you know there’s so much potential in the Windows business that you’ve described today. And yet at the same time Microsoft is very aggressively moving into the media space and some articles recently have said, this is really the fundamental direction you want to take, to become a “media company,” is that correct?

MR. BALLMER:Couldn’t be more incorrect. Hard direction, I mean, the next three years, the notion that we’re a media company, four years, five years, we are a company that makes system software. You go out four years from now, I’ll tell you what our most popular products are Windows, Office, Windows NT Server, BackOffice, and Windows might have four or five new flavors, the CE flavors that we talked about, the
flavors. But, that’s the backbone of our business this year, next year, the year after, and the year after that.

Yes, we’re making some other investments and we’ll be in a couple of other businesses, but don’t get confused, Windows, Office, Windows NT Server and BackOffice, that’s what this company is about.

MR. KIRKPATRICK: And that will be the dominance of your revenues —

MR. BALLMER:Preponderance of our R & D, preponderance of our revenue, preponderance of our investment.

MR. KIRKPATRICK:You’ve been really stepping up your lobbying effort and you’ve got a little more presence in Washington these days. I mean, I suppose some of that is because there have been a lot of agencies looking at you. But, I’m not thinking of that so much here. But, I guess I really just wanted to ask you, and I didn’t mean to make it a non-softball question, I really didn’t, but what will we see from Microsoft in government, and why are you stepping up your lobbying?

MR. BALLMER:Well, there are a number of issues, which I think are very important issues, where government has to — the governments, not only in this country, but abroad, have to make a set of decisions, so that software developers can move forward. Take privacy concerns on the Internet, take the export of encryption and security technology. These are aspects of our market, which the government has an appropriate role in.

And we have a distinct point of view on what’s the right thing to do, and we’d like the government to do that, because we think it leads to growth in our business. And that’s why not only individually, but as part of the Business Software Alliance, we’ve been fairly active in trying to communicate the software industry viewpoint to the government.

MR. KIRKPATRICK: How would you summarize the current state of government’s understanding of your business?

MR. BALLMER:I think government understands the software industry pretty well. But, then government also has to sort of put that in the context of other social concerns. We have a fairly public disagreement about what should be exportable. For example, for encryption technologies with the U. S. government, but the interest of the software industry and a view, at least by some government agencies on how you stop dictators and crooks, sort of have come into conflict and that’s an appropriate role of government to try to go through those issues.

MR. KIRKPATRICK: Okay. I always can come up with a lot of questions for Microsoft, but I think we’re going to try to keep things rolling and I thank you so much, Steve.

MR. BALLMER:Thank you.