The make-or-break project kept engineers just out of college working around the clock hunting down bugs. The product had so much buzz that speculators bought up units to resell later for a profit. The company invested so much in development that its future was riding on success.
This wasn't a phone or an app or even a new crypto-payments platform. It was 1964, and IBM was about to launch the System/360 mainframe.
"You felt like you were in the right place," Pat Toole, Sr., an engineer on the team, recalled recently.
The System/360 was the first in a family of mainframes that would come to dominate enterprise computing for the next 20 years. The 360 celebrates its 50th birthday on April 7, and Toole's story offers an inside look at how it began.
Toole graduated from the University of Detroit in 1960 with a degree in electrical engineering. Sputnik had launched a few years earlier and the space race was underway. Jet planes were replacing propeller-driven aircraft and the nuclear power industry was about to be born.
"It was a fantastic time to be an engineer," he said.
It was also when vacuum tubes were giving way to transistors, and engineers like Toole, who had studied metallurgy and solid-state design, were in high demand. He took a job at IBM -- as prestigious then as landing a job at Apple or Google today -- where he helped to retrain engineers on the new transistors.
The System/360 broke new ground in many ways, including the use of the first rudimentary integrated circuits. Until then, transistors and other components were soldered directly onto circuit boards and then wired together after the fact. For the 360, IBM designed a module on which components were pre-integrated on a ceramic substrate, covered with a cap and stamped with a part number.
The chip module was a precursor to today's highly integrated microprocessors and offered several benefits, including greater reliability and durability. IBM also created some of the first automated design tools to build the chips. It was a huge undertaking, and it would turn IBM into the world's largest chip maker when mainframe production ramped up a few years later.
There was just one snag with the chip modules: IBM couldn't figure out how to successfully manufacture them.
Toole was the quality control manager for the circuit boards and on the front lines of trying to make the production process work. Still in his mid-twenties, he held a job at IBM that would previously have gone to someone much older, a measure of how fast things were changing.
"The company was growing so rapidly with the new technology that literally hundreds of people like me got thrown into different pieces of the 360. You were expected to perform and everything had to happen on time," he said.
The chip modules had to be mounted on circuit boards and run through an automated solder process, which meant high temperatures and exposure to chemicals. The first modules were covered with a varnish, but the varnish became brittle and flaked. The engineers switched to a rubberized material, but the rubber expanded and caused electrical failures.
IBM's management in upstate New York became concerned. "Obviously this thing went screaming up to the top of Endicott and Poughkeepsie because [the problems] basically shut the line down," Toole recalled.