Moore innovation needed
The world of semiconductors is almost a paradox. Take its role in the mobile industry, for example; while smartphones are getting progressively bigger – even Apple is rumoured to be taking a bite out of the ‘phablet’ market with its much anticipated iPhone 6 – the chipsets that power these devices are now smaller than they’ve ever been.
This is because of Moore’s Law, a concept that guides the size to performance ratio of computer chipsets, which has been used for almost 50 years. In essence, Moore’s Law dictates that the number of transistors incorporated into chipsets will double, or there abouts, every 24 months. Back in 1965 when this idea was first introduced that probably didn’t seem like such a tall order. But now, nearly half a century on, it’s proving quite a struggle.
In 2013 AMD expressed difficulty in shifting from 28-nanometer to 20-nanometer chips and just last month Intel experienced setbacks in rolling out its new 14-nanometer family (Broadwell), although that’s not to say new developments won’t come to light. Size reductions and performance increases are still a possibility however the two are no longer being delivered together in the way they were before. Interestingly, Intel reckons Moore’s Law will continue to drive innovation and revive the sluggish PC market in the coming years, but perhaps that’s because the man behind Moore’s Law, Gordon Moore, is an Intel co-founder. Who knows?
For most, though, the appeal of Moore’s Law as a benchmark for development in the semiconductor industry is starting to fade. Chipset manufacturers have long competed to make their processors smaller, faster, and cheaper, only to now be faced with a brick wall. Looking at this objectively, it was inevitable that at some point manufacturers would approach the the limit in terms of how small chipsets could get while still governing a performance increase, and more importantly continuing to return a profit. Even Intel’s former chief architect Bob Colwell predicts the maximum extension of Moore’s Law will be reached in 2020, around the 7nm or 5nm mark.
So where does that leave the chipset industry today? Well, at Semicon in San Francisco next week the industry’s best and brightest will get together and discuss ways of keeping Moore’s Law alive, or find a suitable alternative that will continue to drive innovation while providing the necessary performance improvements to meet growing consumer demand. Semicon will also play host to a lot of discussion around growth of the Internet of Things, especially as manufacturers are starting to turn away from chasing dramatic leaps in processing power and focusing instead on making smaller improvements to their existing designs that can support new types of connected devices. And as Qualcomm predicts there will be 25 billion connected devices by 2020, many requiring a cheap yet efficient chipset to perform basic tasks, that looks like quite the money spinner.
The semiconductor industry isn’t giving up the fight though, nor is it resting on its laurels. Even with the potential death of Moore’s Law hanging over our heads, there’s never been a more interesting time to be involved in the industry. There’s a real chance for a new, outspoken company to step into the fray and help carve out the future of chipset development. So if that sounds like you, but you need a kick-ass PR agency to help get the ball rolling, then why not get in touch?