OnLive founder and chief executive Steve Perlman faced his gathered employees last Friday morning and told them that he was sorry for the pending shutdown of the company. He said that he had been a “nonstop fundraising machine,” but he finally hit the end of the runway and couldn’t raise more. His team applauded. But as employees collected their belongings and final paychecks, and as more information filtered out, some of them grew angry and voiced their anger to the media.
Few game companies held as much promise as OnLive had. After years of research at his think tank, Rearden, Perlman formed OnLive in 2007 with the goal of creating cloud-based gaming (a system in which users played games stored and executed in broadband-connected data centers). They could instantly play games from any computer, the OnLive console, or even an iPad, and they didn’t have to wait a long time to download their games. Publishers provided more than 200 hardcore games to run on the “instant gratification” cloud-gaming service, including blockbusters such as Assassin’s Creed: Brotherhood and Batman: Arkham Aslyum. It was a brilliant idea, and OnLive executed its service for two years with a regular cadence of product innovations. But the question was whether it was ahead of its time. OnLive had all of the advantages of a disruptive startup, but it simply ran out of runway.
Silicon Valley has always been a place where failure is considered a learning experience, not a source of shame. But OnLive’s demise and rebirth generated a lot of anger, showing that it’s indeed possible to make the announcement of a bankruptcy even worse. The story of OnLive is a startup tragedy, and its King Lear is Perlman himself. Perlman should have sold his company to a big gaming developer or an enterprise firm like Hewlett-Packard for hundreds of millions of dollars.
Instead, more than 200 employees lost their jobs on Friday without severance pay or advance notice. Their hard-earned stock options, some of them accumulated over more than five years, are worthless. The company filed for a bankruptcy alternative, known as the Assignment of Benefit of Creditors, and its assets were immediately sold to an affiliate of Lauder Partners, headed by OnLive investor Gary Lauder. A larger group of investors, including Perlman, were wiped out. Smartphone and tablet maker HTC, one of the last firms in that invested at an estimated valuation of $1.8 billion, confirmed that it lost $40 million. Broadband Internet and communications provider British Telecom is also expecting a total loss as well.
Perlman proved brilliant at raising money, but that created a vicious cycle. It is because the company had raised money at such a high valuation that it would have been hard for Perlman to retreat and sell out for a low price in the hundreds of millions, particularly if the business was weakening. HP, which loved OnLive’s enterprise desktop virtualization technology, considered buying it for a while in the late spring and even gave $15 million to OnLive for an exclusive acquisition negotiation, where OnLive would agree that the only suitor it would negotiate with would be HP.
But HP, which reported a blockbuster $8.9 billion loss this week, became gun-shy and didn’t cut a deal. Nor did other rumored suitors, like HTC, which evidently decided to cut its losses. OnLive’s enterprise technology was promising but still young. Its consumer business was not gaining traction fast enough. Many blamed the failure to sell squarely on Perlman’s ego. A well-known venture capitalist told me privately that Perlman’s personality is what stopped him from investing. Perlman could have warned employees of the pending collapse, but if that word leaked out, the chances of selling the company would have been even worse.
From that perspective, Perlman is a sympathetic tragic figure, doomed to act out a foregone conclusion. Perlman himself might have put in more money, but he is likely also investing a lot of Rearden’s money into Dido, another potentially revolutionary startup that is using cloud technology to create an extreme high-speed wireless broadband network. If he took Rearden’s money and invested it into OnLive, he might have even put Rearden or Dido at risk. In the last week, Perlman holed up in a conference room dubbed Camelot. He was calling everybody he knew to try to get more money so he could buy time and negotiate the exit. Nvidia CEO Jen-Hsun Huang visited but wasn’t interested. OnLive owed an estimated $30 million to $40 million to various creditors. Perlman evidently waited too long before calling it quits, and the employees suffered immensely for that.
The new OnLive is a leaner company, and it supposedly has a lot of patents and patent applications related to cloud gaming. (Some dispute that OnLive owns the patents, but OnLive says it does.) Offers to join the new company have been made to 60 or 70 employees, while others may find themselves tapped for consulting. OnLive will continue to operate its service for 1.5 million active gamers. It will probably seek a new buyer and settle for a more realistic price.
The question now is whether Perlman still has a role to play in this drama. It’s not clear if he will stay on and try to make the service into a real business or perhaps sell it for a large sum. The facts will come out over time that will determine whether Perlman acted brilliantly or foolishly. Certainly, his vision was bigger than anybody else’s. I do not think it’s fair to say he acted entirely selfishly in selling the assets to the new company, as he got wiped out, too. But he certainly made mistakes and probably underestimated how much fury would result from the perception that OnLive “screwed employees.” Failure is forgivable, but failing to do right for employees is not. OnLive did pay employees their paid time off, and it gave them COBRA health insurance extensions (employees have set up a fund for those who need help with the COBRA payments here). But few considered that enough.
OnLive’s ABC filing was “expedient but very bad publicity,” said Michael Zyda, the founding director of the University of Southern California’s GamePipe Laboratory. “All people are going to remember is that [a company with great technology] treated people badly and is now back alive at half-staff…. OnLive is dead. Most people just don’t know it.”
One employee said that the company was doing well in last fall, but then the momentum began to slow. Some gamers complained about the latency, as you still had to have a good broadband connection in order to have a good OnLive experience. It raises the question about why OnLive didn’t start its service in Korea or Japan, where the broadband connectivity is already much faster than in the U.S. Nvidia says that cloud distribution becomes obvious once users can get 10 megabits a second broadband speeds for handhelds and 30 megabits a second to the home.
Some employees started to leave. Executive turnover included John Spinale, who left to become the head of Disney’s social game business. Publishers never gave OnLive big exclusives, in part because they couldn’t offend the top game store chain, GameStop, and because they didn’t want OnLive to control their relationship with customers. Royalty checks to publishers were getting smaller. They preferred the business-to-business cloud gaming service offered by rival Gaikai, which was a lot less mature than OnLive but offered publishers to operate their own sites. OnLive spent money on server infrastructure and didn’t have a lot left for customer acquisition. OnLive was always a step ahead of Gaikai in deploying real technology that worked, but it also had to pay many more employees than Gaikai had. Perlman kept the state of the finances close to his vest.
Sources say OnLive burned through $200 million, though the actual number has never leaked. At the close, it was reportedly losing around $5 million a month, one source said. It’s not hard to point fingers in this case, as Perlman pointed at himself for the failure. OnLive’s rival, David Perry, managed a beautiful exit for his firm Gaikai, selling it to Sony for $380 million in July. That could have been gasoline on the fire for the attention that suitors were paying to OnLive. But the price for Gaikai was low, and it didn’t help OnLive command a higher price.
The new company has a chance to be profitable and to sell itself and its technology. Some of the employees are choosing to stay aboard. But for how long is the question, since many have seen years of equity evaporate into nothing. The wild card is the patents. Silicon Valley has seen huge battles over patents lately, and giant companies are spending billions to acquire them. If OnLive’s patents and patent applications are truly valuable, they might be used to one day create the service that gamers really want to experience. But rivals say they believe they have their own patents that will protect them and contain OnLive’s legal might. Once again, those patents will determine whether Perlman is brilliant or foolish. They were not so obviously good that they led to an OnLive acquisition…yet. Andre Vrignaud, a gaming expert at Hit Detection (and not a patent expert), observed that, on the surface, it may be hard to interpret OnLive’s patents broadly because similar remote desktop companies have been operating much longer.
OnLive could also have a future filling the role that Gaikai had played, Vrignaud said. Gaikai was a “white label” service that let other companies take the risk and expense of creating a game portal, and all of the customer support issues that came with it. OnLive might be able to serve that role, now that Gaikai is locked up by Sony. Samsung, for one, might very well need a partner, since it announced an alliance with Gaikai just before Sony picked it up. The new OnLive has confirmed that Perlman will remain as CEO. It appears, then, that he will get a second shot at handling this right.
The rest of the game industry may suffer along with OnLive, particularly if funding for daring ventures dries up.
“This, coupled with the incident that occurred at 38 Studios [where the game studio went belly up and 300 employees were surprised] and recent performances by Zynga, has really left the industry with a black eye,” said Jesse Divnich, the vice president and analyst at market researcher EEDAR. “Certainly, investors will become even more reserved on placing large investments into our industry.”
But at the very least, OnLive has something to teach the rest of the cloud gaming business, sort of like passing a torch. Will O’Brien, general manager of casual game cloud service Big Fish Unlimited, said in an interview that the news suggests that OnLive’s model was not sustainable, but that doesn’t mean cloud gaming has collapsed. He says Big Fish has lower server costs because its casual games don’t require high-end graphics. His firm also owns 1,000 games that it will publish on its network.
“There is bad news here, but industry history shows it takes multiple companies to break down industry barriers,” O’Brien said. “The pioneers get the arrows, the settlers get the land.”
Peter Relan, who is chairman of iPad game streaming firm iSwifter, says that in 20-20 hindsight, the business model and the timing were big mistakes at OnLive. He said that the inability to charge subscription fees for high-end games that required expensive servers was a recipe for losses. And it is only recently that graphics-based servers have become standard in many cloud services firms. Server costs are dropping, and Nvidia is launching ways to enhance graphics-based cloud gaming, but it might not be in time to save OnLive.
“Sometimes being two years ahead of the market is the same as being wrong,” Relan said.
Marketing technologist? We're studying the big marketing clouds
Fill out our 5-minute survey
, and we'll share the data with you.