Samueli: As Moore’s Law Slows, We Can’t Waste Transistors and Must Be More Creative

2012 Marconi Fellow Henry Samueli told the NY Times “It’s graying, it’s aging, ,,, It’s not dead, but you’re going to have to sign Moore’s Law up for AARP.” He added, “Humans will just have to get more creative.”

I asked him, “Which areas do you think most promising for improvements when scaling is limited?” He emailed:

More efficient designs will be required. We tend to waste a lot of transistors in current chips to accelerate time to market. We won’t have that luxury in the future.
3D packaging becomes more important as you run out of 2D scaling.
IoT devices don’t require the most advanced technologies so this is a fertile ground for R&D, especially with regard to sensors.
More efficient designs will require more research. Samueli’s company, Broadcom, spent $2.37B dollars supporting approximately 8,000 research and development employees, including over 800 employees with Ph.D.s. He also believes “Government funding is always needed (especially for universities) to push the frontiers of technology.”

At the 2012 Marconi Symposium, Samueli warned about the inevitable end of Moore’s Law. Not yet, though. “There are two to three more generations left, Broadcom will probably do test chips in 10nm FinFETs next year, and there’s talk about how to get to 7nm, so its highly likely we’ll get there — maybe we’ll even get to 5nm.

Beyond 5nm is where a lot of skepticism comes into play. The barriers beyond 5nm look pretty tough. Moore’s law has been an amazing phenomenon for almost 40 years, … It’s almost a perfect exponential. Exponentials can’t continue forever, The dimensions are approaching atomic limits.”