Moore’s Law, as defined by Intel’s Gordon Moore in 1965, is the predictive business model that “computing would dramatically increase in power, and decrease in relative cost, at an exponential pace.” And for decades, this prediction held true, empowering innovators and setting the pace in the industry.
Due to the fast-turning cycles that doubled the number of components on integrated circuits, increasing computing power along with it, companies did not necessarily need to think of new solutions to improve semiconductors. Moore’s Law’s impact of continuous improved efficiency was simply an industry target everyone worked toward.
Today, it’s no secret Moore’s Law is becoming more and more difficult to accomplish. For many, 28nm was the last node reflective of Moore’s Law—and while transistors may continue to shrink, costs will rise and performance will suffer. This is prompting chipmakers around the world to reconsider the economics of scaling. The end of Moore’s Law has another detrimental effect on the industry: As the costs required to bring new designs to market rise exponentially, it has become nearly impossible for smaller system designers and innovators to get the attention of the shrinking number of semiconductor companies capable of providing modern, high-performance chips.
So we find that the semiconductor industry is in a state of flux and consolidation as chipmakers search for new revenue streams and efficiencies. Last year alone we saw more than $100 billion worth of mergers and acquisition in the semiconductor space, as companies sought economies of efficiency. And the trend shows little sign of slowing — take this recent New York Times’ headline for example: “Semiconductor Industry Shrinks Some More With Latest Deal.” This summer alone: Intel opened its fabs to the ARM ecosystem; Analog Devices snapped up Linear Technology; Cypress closed its acquisition of Broadcom’s IoT assets; Infineon bought Wolfspeed; and perhaps the biggest deal of them all – SoftBank acquired ARM.
Let’s consider the impact the Softbank / ARM deal may have on access to chips (not to mention the cost)? International Business Strategies analyst Handel Jones was quoted in The Wall Street Journal as saying it’s likely that ARM “is not going to be quite as transparent. (Companies) would be prudent to look at different options.” Patrick Moorhead, an analyst at Moor Insights & Strategy said in the same article that “one of the biggest questions about the deal is whether SoftBank would raise ARM’s royalty rates,” adding that it might do this at it introduces new chip designs
Why Alternatives Really Matter
This consolidation has made it even more difficult for system designers seeking custom silicon to get the attention of the companies that remain standing. In fact, it is nearly impossible for small and independent teams to compete for chipmakers’ attention – even if they had the eight- to nine-figure budgets required.
So, the semiconductor industry is at an important crossroads. It’s clear that the old approach to SoC development – with investments in R&D so intense that every project has become a “make or break” proposition for design teams – has been broken for some time. We need to democratize access to custom silicon— and there aren’t many ways to do that.
One growing consensus is to begin harnessing the power of open-source, through the free and open RISC-V instruction set architecture and the power of the crowd.
Many of the industry’s largest names – Google, HP, IBM, Microsoft, Qualcomm, NVIDIA, AMD – more than 50 in all – are exploring how open source fits into their roadmaps, intrigued by the flexibility and customization made possible through open source and are poised to create a disruptive change in the ecosystem.
For established system designers looking for Plan B as they wait to see how the industry dynamic will shake out, it’s worth evaluating the growing number of open source designs based on RISC-V. Open source has the power to reach underserved markets — fragmented sectors like the Internet of Things or emerging markets that do not yet have the volume or revenue required to engage with traditional silicon suppliers. Through its lower cost of entry, RISC-V can democratize access to for inventors, startups, makers and companies alike.
While it is still early in the evolution of open source hardware, the conditions seem right to cultivate a community around RISC-V. After all, when Linux was first introduced as an open source software model, the company was told that open-source wouldn’t succeed — but at its 25-year anniversary, the Linux movement is stronger than ever.