Chip technology, geopolitics and the CAD industry

The global semiconductor industry is going through historic tidal shifts that will impact all software. We review the scene and its implications on platforms and devices and possible impacts on the CAD industry.

Editor’s note: This feature was first published in our September newsletter, Xpresso, released in the fall. For early access to our best content, subscribe to Xpresso now. It’s free!

Additionally, since this article was published in September, two notable chip industry events have taken place. Apple finally released its M1 Pro and M1 Max SoCs ship for macOS computers, providing a clearer picture of what’s possible for ARM-architecture chips for personal computers. At the same time, Intel releases its 12th generation Range of Intel Core processors, bringing substantial improvements from generation to generation. The functionality below is shown as it was executed in September.

Intro and geopolitics

RISE IN TENSIONS BETWEEN THE WEST (particularly the United States) and China is ushering in large-scale change in the semiconductor industry. With nearly all of the leading-edge semiconductor manufacturing in Taiwan or South Korea – not here in the US with Intel in its typical leadership position – the US government has stepped in to help a new era of industrial nationalism. In early June of this year, the US Senate passed legislation (the United States Innovation and Competition Act, USICA) which includes $52 billion in federal funding to accelerate domestic semiconductor research, design and manufacturing in what is known as the Chips Act for America.

Considered a matter of national security, semiconductors power almost every type of digital device and certainly every type of computer system running an operating system, including military systems. The United States has traditionally led the world in top-tier chip design and manufacturing. While design leadership remains in the hands of the United States, its national manufacturing champion, Intel, has erred. (see bottom half of article, below).

In addition to security concerns, the global semiconductor industry is lagging behind and unable to keep up with demand. There is a ripe economic opportunity, and all major global economies – from the US, EU, China and Japan – want more of the booming action.

Following U.S. government investment in the domestic semiconductor industry, new Intel CEO Pat Gelsinger announced a $20 billion expansion of two new factories at Intel’s plant in Chandler, Arizona , where Fab 42 is fully operational and producing 10nm node chips. (Picture: Intel Corporation) – Click to enlarge the images

The new US Chips Act could boost the development of up to 10 new chip factories. Intel has promised two new ones in Arizona (see above). Similar EU plans call for self-sufficiency in semiconductor design and manufacturing in the EU. In the United States, which once held 37% of semiconductor and microelectronics production in 1990, it now holds only 12%.

As China aspires to self-sufficiency in semiconductors, it lacks local companies capable of developing and manufacturing the equipment, non-wafer materials and wafer materials used in the semiconductor manufacturer. The US and EU dominate the critical equipment market with almost no equipment manufacturers in Taiwan and a single-digit share in China. What China has is a growing “fabless” chip design industry.

It is a moment of overlap between an old paradigm that is slowly being replaced by a new paradigm.

The global democratization of semiconductor industry design, development and manufacturing is changing the possible future of the computer software industry. This may have important implications for engineering software in the coming decade. The once stable digital economy based on “Wintel” has largely decoupled. In December 2020, Microsoft has announced that it is designing its own ARM-based chips for servers and Microsoft Surface devices. The server chips are for the company’s own Microsoft Azure Cloud Services data centers. This game largely mimics rival Amazon who designed its own ARM-based chip (Graviton 2) to power its AWS data centers.

Microsoft Windows PCs are no longer the center of computing but rather exist in our presence, much like gasoline-powered cars in the midst of the EV automotive revolution. Although they remain in the CAD industries as primary equipment, they are complemented by a rapidly changing landscape of new, smaller device types. Promises of cyclical economic improvement from the Wintel hegemony first slowed and then fell apart quite abruptly in recent years with Intel’s manufacturing misfires. (see more on this below).

It is a moment of overlap between an old paradigm that is slowly being replaced by a new paradigm.

Moore’s Law: Then and Now

Since the inception of Intel and the emergence of the x86 CPU chip architecture, Moore’s law largely kept its promise. Specifically, Moore’s Law, named after Intel co-founder Gordon Moore, states that the number of transistors on microchips doubles every two years. This requires a compound annual growth rate of 41%.

Moore’s Law in vivid real data, across decades of chip advancements from Intel, Motorola, ARM, Apple, IBM and others. It very clearly documents the doubling of chip transistors approximately every two years, which is Moore’s Law. This is a compound annual rate of approximately 41%. Still, there is a pattern to the points above that we want to highlight in the next image below. (Image: Wikipedia Commons)

In the 1970s and 1980s, Intel’s major microprocessor competitors, such as Motorola and IBM, largely kept the semiconductor industry abuzz with advances. AMD added competitive pressure in the 1990s and early this century, and competition for servers, in particular, led Intel to push for larger, more powerful server processors.

In the years from 1994 to around 2007 we can see (in the graph) a huge gap between powerful server chips like IBM’s Power6, Intel’s Itanium 2 and AMD’s K10, all basically over 500 million transistors compared to the ARM Cortex A9 with less than 50 million transistors.

Yet something changes from 2008, as ARM is advancing at a faster rate than everyone else in the chip industry. (See the green line in the table below.)

The Moore’s Law chart is highlighted again this time. The red line is the progress of the Intel X86 chip. The green line is the progress of the ARM chip. As we can see from around 2008, ARM’s progress is accelerating while Intel is barely keeping pace with Moore’s Law. Then, in 2016, things for Intel slowed further, while ARM’s progress accelerated again (steeper line).

Suddenly, ARM licensee Apple releases the A7, a landmark 64-bit SoC with over a billion transistors. At this point, Apple’s iPhone chip now has more transistors than IBM’s Power6 from 2007 when the iPhone was introduced. In just six years, a chip in a phone had more processors than one of IBM’s most powerful server chips in the world.

ARM’s ancestry is just one factor in the global semiconductor tide shift. Intel’s missteps in missing Moore’s Law are another – and we’ll get to that in a moment. No other company in the industry represents the forces of democratization of the global semiconductor industry as much as ARMwhich licenses its chip designs to anyone wishing to work in the ARM ecosystem.

next page: ARM, Apple and Intel