In January 2007, Gentoo’s bug tracker had become part of the problem it was supposed to solve. Slow performance wasted developers’ time and encouraged duplicate reports. GNI helped provide replacement infrastructure. For a company selling hosting, this was an unusually good demonstration: the useful thing about a server is the work people can do when they stop thinking about it.
- The offer: dedicated servers with people to operate and support them.
- The niche: online games, SaaS and digital media with demanding workloads.
- The outcome: a 2010 acquisition by FusionStorm, recorded at $18.265 million.
GNI, the San Jose business also known as Global Netoptex, occupied that unglamorous but consequential territory between having software and keeping it available. Its customers made things people wanted to use. GNI supplied the physical infrastructure and operational attention needed to make using them possible. In online games, the distinction becomes particularly vivid. A beautifully designed world has limited appeal when nobody can connect.
Peace of mind had a maintenance bill
In a July 2006 announcement, founder Derek Wise described the proposition with admirable economy: “We sell total peace of mind.” GNI’s services covered managed hosting, security, cabinets, cabling, load balancers and round-the-clock technical support. Named customers included Symantec/Brightmail and the Oakland Raiders. The product was a bundle of equipment, expertise and responsibility, rather than a machine delivered with best wishes.
“We sell total peace of mind.”
Derek Wise · July 2006
The arrangement with datacenter operator 365 Main shows how the bundle worked. GNI established operations at its El Segundo facility and provided onsite technical assistance. The building supplied a protected physical environment; GNI supplied services and people. Those are separate jobs. A locked room with reliable power cannot, by itself, diagnose what is happening inside a customer’s application stack.
Wise’s account of the company’s beginning adds a rather less comfortable detail. He dated its origin to December 2001, after his employer, Enron Broadband Services, collapsed and left California customers behind. Public company profiles use 2002. It is an origin in which the appeal of having someone remain responsible requires little embellishment.

The dragons needed somewhere to live
The gaming opportunity gave this general hosting business a sharper identity. Online Game Services, or OGSi, was already GNI’s gaming partner before GNI bought it in early 2007 for more than $3 million. The combination brought together managed-services experience and a business focused on the needs of multiplayer-game publishers. Specialization came through a relationship that had already been doing the work.
The scale was tangible. In November 2006, OGSi announced more than 1,000 IBM BladeCenter servers at 365 Main’s San Francisco datacenter. Customers included GoPets and Ping0, the latter associated with the online component of Hellgate: London. This was computing packed into a dense physical installation, with power and cooling requirements that did not become less physical because the customers sold imaginary adventures.
Its commercial answer was “pay-as-you-grow.” A publisher could obtain managed dedicated hosting without first constructing its own infrastructure. The appeal was especially clear for smaller and midsize publishers: operating an online game did not have to begin with becoming a datacenter operator. Growth still required resources. The service changed who assembled and managed them.
A customer wanted more than hardware
The customer announcements make the differentiation more specific. In October 2007, Icarus Studios entered a partnership with GNI Games and Digital Media. Icarus brought services for developing games and virtual worlds; GNI brought infrastructure. Their first announced project was Fallen Earth, with hosting planned through beta testing and commercial operation. The partnership addressed the passage from making an online product to running one.
A year later, Athens-based Aventurine chose GNI for Darkfall Online. Aventurine’s stated requirements included hardware, bandwidth, competitive pricing, flexible terms and knowledge of the games industry. That list is instructive. Technical capacity mattered, but so did the provider’s understanding of its customer’s business. A studio was buying a working relationship alongside the computing equipment.
GNI reported that concurrency figure alongside more than 12.8 million users. The two measures answer different questions. Registered users describe an audience accumulated over time. Concurrent connections describe people arriving together. The latter is the more revealing number when considering simultaneous demand, though it does not tell us how much capacity any individual game needed.
The acquisition price was mostly equity
The business also grew beyond its memorable games. Its 2007 Inc. list appearance put reported revenue at about $904,000 in 2003 and $8.1 million in 2006, with 14 service locations across the United States, Europe and Asia. It ranked 349th. These are historical snapshots of a company expanding its infrastructure business, rather than measurements of an operation today.
FusionStorm acquired substantially all GNI’s assets and liabilities in 2010. Its subsequent financial statements date the transaction to April and record a purchase price of $18.265 million: $811,000 cash, $1.432 million in vested stock options and $16.022 million in common stock. Equity supplied most of the consideration. The headline price was considerably larger than the cash cheque.
Copy the division of responsibility
The practical lesson is a purchasing question: which operational burden will the supplier actually take away? GNI’s case suggests looking beyond hardware specifications to support, deployment, commercial flexibility and familiarity with the workload. The principle travels well beyond games. A service earns its place when it removes work that distracts customers from the thing they are trying to make.
That arrangement has limits. Buying hosting cannot make an unpopular game popular, or relieve a studio of responsibility for its software. A company with the expertise and appetite to operate its own infrastructure may prefer to keep control. Outsourcing is useful when the duties transferred are clear and the provider can perform them. Even fantasy worlds benefit from a decidedly literal contract.