February / March 2011 Volume 9 Number 1

Better Not Bet Against Steve Perlman

“As an inventor, it’s a compliment to have someone say an invention is impossible. But as a CEO of a startup, it’s frustrating.” —Steve Perlman

On  Steve Perlman’s latest venture, the skeptics were out in force. In March, 2009, a day after he announced his startup, OnLive, game developers studied his idea and said that what he was planning to do couldn’t be done.  He  orchestrated a live demo of his video-on-demand service, which he said would allow a gamer to play a high-end game on a low-end computer, at the Game Developers Conference in San Francisco. At his company’s booth,  computers connected to the internet and accessed games that were stored  and computed in servers in Santa Clara, Calif. But the skeptics said it  was fake. They asked the show’s broadband provider if there was an internet connection at the booth, and the provider said there wasn’t one. Game developers criticized the technology, saying that it was impossible to play fast-action games across the internet because of the latency of the network. It was too slow to accommodate action games and server costs would bankrupt the company.

As  for the trade show connection, Perlman explained that he had to go through another provider to get faster bandwidth into his booth, which he divided into a bunch of different lines for various PCs. Nevertheless, rumor started spreading that OnLive was nothing more than a  scam. The scam rumors gained currency because Perlman’s idea seemed crazy. At the time, Perlman said, “As an inventor, it’s a compliment to have someone say an invention is impossible. But as a CEO of a startup, it’s frustrating.”

Today,  he doesn’t seem so crazy anymore and a lot of people in the game industry are coming around to recognize that. The Trademark and Patent office awarded OnLive a fundamental patent on what has come to be  called “cloud gaming,” where a game is executed in an internet-connected data center and then the data is compressed and pass through the internet to a user’s machine. Richard Doherty, an analyst at  the Envisioneering Group, said that the OnLive patent was a “landmarket  patent.”

The  game is instantly playable, allows the user to log on from anywhere to access a saved game, and it bypasses retail game sellers. In a business akin to Netflix’s streamed movies, OnLive launched its games-on-demand service in June, 2010 and it disrupted the retail game stores such as GameStop. In November, it launched a small $99 adapter that can bring cloud games to big-screen TVs, disrupting the game consoles. And now Perlman says that in 2011, OnLive will use the same technology to create  a movie streaming service.

“When  you first wrote about us, people were saying how it would never work,” Perlman said in a recent interview. “You can never time when a patent will arrive. But it’s gratifying to get the recognition.”

OnLive is gathering momentum. In 2010, the Palo Alto-based company raised $60 million from British Telecommunications and Belgacom. The valuation was $1.1 billion. In a year when 91 game companies raised $1.05 billion, Perlman’s funding was one of the biggest. The company has  gotten lots of players signing up for both one-time game purchases, rentals, and monthly subscriptions for its service, which has been played for millions of hours. OnLive launched an iPad app and it is preparing one for Android smartphones and tablets. The iPad app will let  users spectate online game matches and it will even allow Windows 7 apps to run on the iPad, using OnLive’s cloud application technology.

And  at the Consumer Electronics Show in Las Vegas in January, Vizio, the nation’s largest TV distributor, said it would build OnLive’s service directly into its internet-connected TVs. The company now has 200 employees and it recently hired Etienne Handman, who built Pandora’s cloud-music service into a 75 million user business, as its chief operating officer. OnLive has learned to thrive in a business that is dominated by multibillion-dollar game companies. It can do so by tapping  partners such as Equinix, which allows OnLive to access as many as five different data centers across the U.S. and more in Europe and Asia.

OnLive  can charge a lower price for its games because it cuts out the retailer  and distributor. It shares revenues with retailers, but it can provide games for a much lower cost of goods. At first, OnLive required a subscription fee for its service and users had to pay for games on top of that. But as it neared launch, it decided its costs had dropped enough so it no longer had to require a subscription fee. So users can sign up for free and try game demos. When they want to get a full game, they have the option of purchasing, renting for a few days, or getting a  bunch of games for a $10 monthly subscription fee. Game publishers have  given the company a mix of new games and old games. Mafia II, published  in August by Take-Two Interactive, arrived on OnLive on the same day that it was launched in stores.

Because  of this progress, Perlman is getting the last laugh on the skeptics. But the bar for success is still very high. Lots of threats remain. In game retailing, GameStop is as strong as ever with billions of dollars in retail sales. In consoles, Microsoft, Sony and Nintendo are still selling millions of consoles. OnLive has 40 games available on its games-on-demand service, but there are hundreds or thousands on the other platforms. And the biggest game of the year, Activision Blizzard’s  Call of Duty Black Ops, isn’t yet available on the OnLive service. OnLive has a long way to go before it makes a dent in the market share of its rivals.

If it can’t get to critical mass, it won’t win over more game publishers, who so far have experimented with OnLive. But Perlman isn’t worried because the gamers love the service. They’re talking it up and playing lots of games. PC gamers like it because they don’t have to pay $1,500 to $3,000 on a new hardware system just to be able to play the newest games. OnLive can continuously upgrade the hardware in its servers, but users don’t have to pay a dime for hardware, since OnLive will run on netbooks, laptops, Macs, and older PCs as well as new ones. The users can play on their computers or get the $99 adapter to play on TVs.

The  TV strategy in particular is a great trojan horse. The main chip in the  adapter is a decompression chip made by Marvell Semiconductor. The same  chip is frequently used in TVs that connect to the web. So OnLive’s service can be added to a TV set with a simple software upgrade. To make  a TV into an OnLive game machine requires no extra hardware. And OnLive  can sign up lots of TV makers, who would be more than happy to sell web-connected TVs that take business away from the game console makers. OnLive also has a pretty cool-looking user interface, which allows a user to look at lots of games being played live at the same time. That allows a user to see a game in action before deciding to try out a demo or buy the game. When OnLive starts its movie service, it can do the same for different movie channels, showing scenes from a dozen movies at  the same time.

Perlman  is sitting pretty now because he started with technology and a lot of research. With more than 100 patents in his career, Perlman is something  of an engineering genius. He is also persistent and addicted to being an entrepreneur.

He  joined Apple in 1985. There, he led the development of QuickTime, the online video format that Apple still uses today to play digital movies.  He left in 1990 to join General Magic, a promising startup that failed  in its mission. In 1994, he co-founded Catapult Entertainment, which made Xband modems for early game consoles.

In  1995, he co-founded WebTV Networks, which sought to create a simple appliance that could hook the TV up to the internet. It saw some early success and was acquired by Microsoft for $503 million. Microsoft discontinued WebTV, but it used the technology it acquired to create Ultimate TV, a satellite TV box, and the Xbox video game console.

But  Perlman left in 1999 to start his own business incubator, Rearden Steel, which he used to do fundamental research and then spin out ideas into startups. He used the money he made from WebTV to create a variety  of research-backed start-ups. One of them was Moxi Digital, which created a digital video recorder set-top box. Moxi got a lot of coverage  when it came out of hiding, but it was sold to Paul Allen’s Digeo in 2002. In 2004, Perlman created Mova, which created a face-capture system  that allowed movie makers and game developers to create 3D images of faces that could be manipulated for digital special effects. It was used  to capture the shape of Brad Pitt’s face in the film, The Curious Case of Benjamin Button, which won an Oscar for visual effects. He also helped fund Android, a company headed by entrepreneur Andy Rubin, who sold the company to Google. Today, the Android software is used in a wide range of smartphones and it represents Google’s spearhead into mobile phones and tablets. Almost all of those startups were related to  ideas Perlman had about online entertainment.

In  2007, he spun out OnLive as a separate company, assigning to it patents  and technologies that he had been working on since 2002. If there was a  criticism of Perlman’s earlier companies, it was that they didn’t have legs.
Perlman created cool ideas and sold them off for good valuations, but they rarely turned out to be huge successes. It seemed as if Perlman  specialized in arranging good exits for his companies, but he rarely changed the world. Some of that criticism is unfair, since QuickTime and  Android are still thriving and the Xbox and Xbox 360 were built by teams that Perlman had pulled together.

Still,  OnLive is his chance to build something that has lasting value. He started with a technology that could compress video and send it from one  place to another. OnLive is built on that technology. It uses hefty graphics-based servers to compute a game’s operations. Then it takes video imagery, compresses it, and sends it as fast as it can over the internet. Within a matter of milliseconds, the compressed data arrives at the user’s machine. There, it is decompressed so that it can be displayed on the user’s screen. The images can be displayed in 1080p, or  full high-definition. But there is no huge download of files or data.

The  technology requires a somewhat fast pipe, such as a cable modem or digital subscriber line (DSL) service that can move data at 6 megabits a  second. Fortunately, most homes with broadband—particularly cable users—have broadband speeds that are adequate. OnLive has devised a bunch of tricks to make the delivery of the game smooth, and it has collected more than 100 patents related to its inventions. With years of  inventions, testing and patience, Perlman’s team made the impossible into the possible. Perlman himself did a lot of the fundamental work to figure out how to make the service work. And now there are lots of rivals—Gaikai, OnLive, Playcast -- who are all trying to do the same thing. But OnLive has years of work already done, giving it a big head start.

With OnLive,  Perlman wants to hit the ball out of the park. He has signed up investors such as Warner Bros., Autodesk, Maverick Capital, and AT&T. Game publisher partners include Electronic Arts, Ubisoft, Take-Two Interactive and THQ. He recruited a team of executives, such as  former game executive John Spinale, who had a lot of friends at the big  game companies. All of the evangelism and development time is finally starting to pay off. OnLive’s success will give Perlman some real credibility, even if it is coming very late in the start-up’s life.

He  was able to make it this far because he had a huge bankroll from his WebTV sale. He had an innate curiosity in how to make something that was  technically impossible work. He had a vision to create a new way of delivering games and entertainment that could potentially disrupt multiple industries. And he had the patience to see it through, with already nine years of work behind him.

If OnLive succeeds, it will be because Perlman created it from scratch, he pursued a big vision,  and cared for it like it was his own baby. And that’s a lesson for every entrepreneur.

Dean Takahashi reports from Silicon Valley for Innovation.