ARM’s current management recently invited Saxby to a Q&A session with the 250 members of ARM’s Silicon Valley office. “I came out feeling great--that these kids were just as good as we were,” he said.
“We are very much more global than we were, ARM has a lot more partners in every dimension than it used to have, more employees and more things to be concerned about,” he said. “The need for partnership today is greater than ever because the world is more complex and costly--you have to work with more players to get the job done,” he added.
Saxby talked about the historical importance of ARM’s “leading teaching customers” like Nokia that first used ARM in cellphones and Nintendo that put it in its handheld Gameboy Advance. There were failures too, he reminded them, like the 3DO game console and the Apple Newton.
“We nearly went bankrupt because many of the first customers’ products were failures--it took a very long time for royalties to come but we believed in ourselves,” said Saxby, who served as ARM’s first chief executive from 1991-2001, leaving ARM as chairman in 2006.
“The company has done very well since my retirement and that’s a credit to people like [current CEO] Warren East,” he said. “You need different players at different stages, and for this stage the team is very strong,” he added.
When not working with new startup’s Saxby also acts visiting professor at his former university in Liverpool, teaching on the business of engineering. “I am working in practice less hours than I used to, so I get more time for skiing and playing tennis—really, I am speeding up,” he said.
Since hardware is capex intensive compared to a doughnut shop coupon app, hardware will linger behind the needs.
Feature pull app
"Engineers wishlist" to prioritise the hardware needed for the next gen products and features.
Worth investing in?
Bill Tsai once laid out the math very clearly for me.
VCs can spend a few tens of millions on a chip startup and hope it pays off more than $100M at exit. Or they can invest a few single-digit millions in a Web 2.0 startup and have a shot at a billion.
VCs aren't interested in projects that span more that 4 years, but with the complexity of semi industry at this point you hardly can find any semi start-up that can bring ROI in that time frame.
Software and apps could be quite complex too, but the turnaround time is much faster, especially with the app store infrastructure in place.
I'm still not sure, however, how companies that write software and apps manage to pull in hundreds of thousands of dollars in VC funding for what is basically slideware, while engineers with ACTUAL products have to fight for every dollar. The semi space does require more capex, but that's because it produces real things....
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.