Rethinking The Next Semiconductor Roadmap

Article By : Rick Merritt

The semiconductor industry needs to think differently and find roads Ahead to growth. Trade group SEMI aims to launch think tanks, investment funds...

HALF MOON BAY, Calif. — For a moment, it felt like the semiconductor industry was in the middle of a three-way knife fight.

A former U.S. Senator suggested finding new supply chains because China was no longer safe. A Wall Street banker advised adopting new business models to tap into tech investments pouring into software and services.

Two Silicon Valley executives joined the disruption pile-on. They called for rethinking the way the industry makes chips — even whether it should be making something as passé as wafers.

An attendee on the final day of the Industry Strategy Symposium might have considered leaping off the nearby California cliffs into the ocean. But not the event’s new host.

SEMI hosts its annual ISS event at the Ritz-Carlton, just outside Silicon Valley. (Image: EE Times)

Trade group aims to launch think tanks, investment funds

Ajit Manocha, chief executive of the SEMI trade group, has a strategy. It’s all about collaborating.

Manocha convened at the event the first of several planned think tanks. In April, he will hold his first meeting with investors about setting up a semiconductor fund. And his trade group is lobbying the U.S. government to increase its spending on chip research by an order of magnitude.

Separately, he is exploring a partnership with the Imec institute in Belgium. He sees it as a way to bring together the research center’s constituents with his industry partners in new ways that expand the chip sector’s horizon.

The good news is that “the fundamentals for growth are all there,” Manocha said in an interview with EE Times. “In 40 years in this industry, I’ve never seen so many disruptions at the same time.”

The bad news is that there’s no clear upside to some of the disruptions — like the U.S./China trade war. “It will create an imbalance and fragment the supply chain,” he said. “The business landscape will change drastically in 10 years if we don’t fix this problem.”

In a talk at the event, defense expert and former U.S. Senator James Talent expressed much the same sentiment. “We want to deter [China’s unfair trade policies] without provoking them, but right now, we’re provoking without deterring,” he said.

On the financial side, Manocha aims to create a working group of global venture and corporate investors along with entrepreneurs and government representatives. He will propose that they create a fund for the kind of long-term investments that semiconductors need.

The U.S. Department of Defense ramped up its Electronics Resurgence Initiative last year, spending about $300 million a year on a variety of programs. It was widely praised, but Manocha said the industry needs annual investments in research measured in billions.

“If China is going fast, we need to go faster or the growth will slip away from us,” he said.

The additional spending needs to “start with talent. Government should take a lead in STEM education, which is way down in the U.S. and Europe, too,” he said a day after SEMI gave an update on its own workforce initiative.

Public school investments will build a foundation for spawning talent. But for now, “the U.S. has to depend on imported talent, so if we don’t fix immigration, we will have a major shortage in the West,” he added. “We’re pushing hard for better visa policies for high-skilled immigration.”

The usually upbeat executive is not optimistic about getting a response in the next six months given the Democratic transition in Congress and the general gridlock. Two stints on presidential advisory groups and six years on the board of the Semiconductor Industry Association taught him that “on good days, Washington is slow, and on rough days, it is even slower.”

He remains bullish on the chip industry, noting the potential of IoT to improve everything from agriculture to health care and more. Think tanks that he hopes to start in Europe and Asia will join the nascent U.S. group to sketch out the potential.

Banker tells chip vendors to think differently

The semiconductor industry needs to think differently to get venture and private-equity funding that’s largely passing it by in favor of software and services.

Specifically, it needs to find ways to reduce the cost of EDA tools and IP as well as chip fabrication, then find fresh business models and new vectors for growth. That was the prescription from Marco Chisari, a managing director at Bank of America Merrill Lynch, in a talk at the event.

VCs spent 15% of their tech dollars in semis in 2004 but, these days, have scaled back to just 1%. Private-equity groups have held firm at about 5% to 6% amid booming spend on tech in general from both groups.

Investors naturally favor software and services startups such as AirBnB and Uber that ask for relatively low sums upfront and promise high returns quickly. The proposition for chip startups is usually the opposite.

“If you sat in the chair of the one responsible for the money, you’d make the same decision,” Chisari said, acknowledging that investors could use some education in the economic transformation that semiconductors have helped spawn.

As a result of the dynamics, only one chip company is among a group of 176 unicorns, aka tech startups with a valuation of $1 billion or more — the U.K.’s Graphcore. Public markets flow the same way, he said, noting that in the last seven years, only four chipmakers have had successful IPOs.

“The problem is complex and wide … the dependence on Synopsys and Cadence design flows is critical but expensive,” he said. “Can we reduce design and prototype costs and the cost of [making] silicon?”

Investors “ only live for faster RoI … [but] there won’t be many software investments for them without hardware,” Manocha said. “It’s time to educate them that the future is based on bigger, better hardware.”

Semi Investment Trends
VCs and private-equity investors are largely ignoring semiconductors. Click to enlarge. (Source: Bank of America Merrill Lynch)

Shades of Dr. Strangelove in Silicon Valley

If the picture from Wall Street was dark, the view from Washington, D.C., was coal-black.

“You are a very important piece in a very large game that’s about a lot more than the trade deficit,” said Talent, who served on a number of defense commissions and was a policy advisor to Mitt Romney.

“Economic and military dominance may depend on who controls semiconductors — that’s a growing sense in Washington and, I think, in China as well,” he said.

It’s a message that chip execs have heard before, one that Silicon Valley was largely founded on. The latest twist is Washington’s new foe.

“The [China] regime uses wealth it is creating to build a platform to challenge the norm-based international order which the U.S. and its allies midwifed,” Talent said.

The China government bankrolled at least half of the investments in its now-dominant solar panel industry, restricted market share of foreign companies, and has stolen as much as $320 billion a year in technology via commercial espionage, he charged.

“They’ve taken to new levels the tactics” others have used, he said. “It amounts to an attack on the global trading system.”

Meanwhile, “China has been engaged in a big military buildup,” he added, giving attendees a diagram (below) of the conventional threat in the Pacific and describing emerging tech vulnerabilities in satellites and elsewhere.

China Threat Talent
China is exerting military influence on disputed islands in the Pacific, said a Washington defense expert. (Source: Jim Talent)

“The next major conflict will be fought in the information domain, including space … China has its own space and cyber-force with hypersonic, quantum, and AI weapons,” Talent said. “Who wins the race to operationalize those technologies basically wins.”

Talent predicted that the U.S. and China will strike some sort of a trade deal this summer that will “lessen but not resolve” tensions. “My recommendation is to be flexible” about supply chains. “Don’t put all your eggs in one basket,” and look toward growing “defense budgets as an opportunity for you,” he concluded.

Asked about the long-term outlook, he observed that “hopes for a more liberal China have proven futile. Reality, not hope, should drive U.S. policy.”

It could take several years for a new U.S. policy on China to emerge.

“We got preoccupied with the global war on terror, [but now, we] have to develop doctrines and deterrence mechanisms … to channel [China’s] ambitions into acceptable directions with purposeful measured responses,” he said.

One veteran analyst at the event brushed off the comments as typical of the winds from Washington.

Technologists want to rethink the roadmap

The event concluded with a panel of technologists who generally called for rethinking industry roadmaps.

Federico Faggin, lead designer of the Intel 4004 and Zilog Z80, said that electronics ought to adopt a biological view wherein systems are like organisms, constantly changing. “That is the new frontier,” he said.

Terry Brewer, president and founder of Brewer Science, rejected the bio metaphor but called for radical change. AI could knock the industry out of its humdrum pattern of building separate processors and memories, he said.

“I feel sorry for people in this room who think they will make wafers the rest of their lives,” he added.

An Applied Materials executive did not go quite so far, but he called for a deep rethinking of Von Neumann computing and today’s techniques for making components.

“Things are in trouble going forward, so we have to question things up and down the stack,” said Sanjay Natarajan, who, on LinkedIn, describes his mission as getting Moore’s Law unstuck.

“We still build chips fundamentally the same way with discrete functions and tools like deposition and etch,” he said. “We need to rethink that … Maybe we need an analog switch.”

“We have settled into a paradigm that has been wonderfully successful,” he added. “It’s hard to disrupt that, but it’s not going to get us to the next plateau.”

Intel’s chief technologist, Michael Mayberry, took a more philosophical, tongue-in-cheek view. “One form of psychology says the more ignorant you are, the more sure you are of yourself,” he said. “I don’t know how I’m going to solve these problems, but that’s a start for what to work on.”

Leave a comment