Grid computing can trace its intellectual roots back to the operating system Multics, but its immediate ancestor is "metacomputing". Metacomputing dates back to around 1990 and was used to describe efforts to connect US supercomputing centers. Larry Smarr, a former director of the National Center for Supercomputing Applications in the US, is generally credited with popularizing the term.

FAFNER and I-WAY were cutting-edge metacomputing projects in the US, both conceived in 1995. Each influenced the evolution of key grid technologies.

FAFNER (Factoring via Network-Enabled Recursion)
FAFNER aimed to factorize very large numbers, a challenge very relevant to digital security. Since this is a challenge which can be broken into small parts, even fairly modest computers can contribute useful power. Many FAFNER techniques for dividing and distributing computational problems were forerunners of technology used for SETI@home and other "cycle scavenging" software.

I-WAY (Information Wide Area Year)
I-WAY aimed to link supercomputers using existing networks. One of I-WAY's innovations was a computational resource broker, conceptually similar to those being developed for grid computing today. I-WAY strongly influenced the development of the Globus Project, which is at the core of many grid activities, as well as the LEGION project, an alternative approach to distributed supercomputing.

Birth of grid computing
Grid computing was born at a workshop called "Building a Computational Grid", held at Argonne National Laboratory in September 1997.

Following this, in 1998, Ian Foster of Argonne National Laboratory and Carl Kesselman of the University of Southern California published "The Grid: Blueprint for a New Computing Infrastructure", often called "the Grid bible". Ian Foster had previously been involved in the I-WAY project, and the Foster-Kesselman duo had published a paper in 1997, called "Globus: a Metacomputing Infrastructure Toolkit", clearly linking the Globus Toolkit with its predecessor, metacomputing.