TALAVASEK AND EGGER worked out the final Epic frame design in July 2007, but before moving too far into 3-D modeling, they ordered a rideable aluminum prototype from Taiwan. The sample was neither light nor strong, but as the first proof-of-concept, it would allow them to test the geometry and basic suspension performance.

On trail rides in the Santa Cruz Mountains, Talavasek and Sloan could sense side-to-side flex in the rear triangle. "If you can feel it on the bike, then it's really bad," says Sloan. Lab tests showed that the two-piece swing link was a weak point. Talavasek later designed a new single link to create a stiffer rear triangle. Separately, Egger and his industrial-design team used a 3-D plastic printer to spit out a model of the Epic, which they used to try out the paint job, decals, and other surface aesthetics for the frames.

After settling on a basic frame shape, Talavasek moved into 3-D modeling, using a software program called Pro/Engineer. This allowed him to determine, virtually, the critical qualities of his various designs before they were built. He also could assess the torsional and lateral stiffness of various frame shapes, which would dictate how the bike would steer and how efficiently it would pedal. Kinematics software allowed Talavasek to plot the effects of rider movements and changing terrain on the movement, or travel, of the suspension.

His plan was to start with a light, weaker bike and then, through testing, add to its bulk and strength. By the end of two months, Talavasek had a complex electronic rendering of the bike. He sent a CD-ROM to Specialized's manufacturing partner in Xiamen, China, and, digital file in hand, the manufacturer cut clamshell molds from large pieces of steel that were later machined and polished.

Talavasek flew to China in October 2007 to watch the first bikes come out of their molds. Before leaving, he requested a run of Revision One—first samples of manufactured carbon frames—to be sent to the Morgan Hill test lab.

As part of the endless push to create lighter bikes, building materials for high-end racing equipment have evolved from steel to chrome-moly steel, aluminum, titanium, and the current favorite, carbon fiber, which is notable for its high strength-to-weight ratio. Working with carbon fiber increases the complexity of an engineer's job. When you're using steel or aluminum, the tubes have a limited number of attributes, based on their typically oval or round shape. Working with carbon offers an almost infinite number of design possibilities, because the shapes can be more sculptural.

The process of making carbon fiber and fabricating a bike frame is a complex blend of science, industrial process, and artistry. Starting off as threadlike carbon strands, spools of carbon fiber are spun into sheets resembling a tapestry. Those sheets, which look like mosquito netting, are covered in a highly viscous liquid, and the entire gooey fabric is sandwiched between huge pieces of wax paper. To make bike tubes, many specifically shaped carbon-fiber pieces are layered on like papier-mâché around rigid forms. These forms are removed and replaced by plastic bladders, and then the whole enchilada is placed inside a metal mold for final shaping. Air pressure inflates the bladders and forces the carbon mesh against the inside walls of the mold, at which point heat hardens the sticky material into a solid. The mold yields a cured "monocoque," a fancy word for a single-shell construction technique that supports loads with its skin. The carbon-fiber Epic would have two such pieces, which would be fused together to complete the bike's one-piece frame.

Though the mold for the 2009 Epic was now set, the exact thickness of each section had yet to be decided. Confirming the placement, direction, and thickness of the carbon would consume Talavasek's time for several months, starting in November 2007. Computer models could not tell him exactly how to lay up the plies of carbon fiber—a process that requires not just number crunching but experience and feel. As he sought to push weight limits and make the Epic frame as light as possible, the test lab confirmed what his computer models could only predict.

"Watching how a bike behaves during a test, you can learn a lot," Talavasek said. "Often, I just sit in front of the machine for 15 minutes and watch the frame. You get a sense for where it flexes, whether you need to add or take out material, if there is room for saving weight or making it stronger."

By March 2008, with the world championships only three months away, Talavasek still hadn't decided whether to use heavier aluminum in the rear triangle of the bike or to go with an all-carbon frame. Lab tests showed him that he could cut weight from the chainstay and still get the stiffness he needed with carbon. Testing would help him remove almost half a pound and produce a 3.64-pound frame. In the end, with all its parts, the 2009 Epic would weigh just 21.7 pounds, nearly a pound and a half lighter than the 2008 model.