feature article
Subscribe Now

Intel’s Blue Genie Pat Gelsinger Wants to Bring Andy Grove’s Spirit Back from the Dead

Can He Do That?

Pat Gelsinger officially became the new Intel CEO on February 15, 2021. He arrives at a time when Intel is facing numerous monumental challenges from multiple competitors in several markets. The world is watching to see how Gelsinger plans to guide Intel, and he’s been very direct about his plans, saying that he intends to bring “Grovian Culture” back to Intel. That’s a reference to Intel’s former CEO Andy Grove, who led Intel to great technical and business achievements over more than three decades. There’s a lot of meaning and Intel history packed into those two words.

Gelsinger grew up on a farm in rural Robesonia, Pennsylvania, southwest of Allentown. In a 2016 interview, he told Forbes that his family raised dairy cows, pigs, soybeans, and sorghum. Gelsinger joined Intel right out of high school, sort of. He actually skipped his last year of high school to earn an associate’s degree in electronics from nearby Lincoln Technical Institute in 1979, according to an official Intel timeline. Gelsinger explained that Intel recruited him right out of Lincoln Tech for a technician’s position in California. He has said repeatedly (and with some pride) that the Intel recruiting report described him as “smart, aggressive, and arrogant.”

At Intel, Gelsinger first worked as a quality control technician. He then worked as a design engineer on the 16-bit 80286 – introduced in 1982 – a microprocessor that Bill Gates reportedly called “brain dead” because of its inability to run multiple copies of MS-DOS simultaneously. (That’s a complicated and unfortunate story involving bit swizzling of the address-extension register when switching between the processor’s real and protected modes.) Gelsinger then worked on Intel’s follow-on product, the breakthrough 32-bit 80386 processor, which Intel introduced in 1985.

Gelsinger attended Santa Clara University while working on the 80286 project and got a BSEE in 1983. He then attended Stanford while working on the 80386 and received his MSEE degree the same year that the processor was introduced. Gelsinger caught Andy Grove’s attention during the 80386 project. Then Intel’s President but not yet CEO, Andy Grove had noticed the young and very energetic Gelsinger’s technical work, saw something he liked, and started mentoring him. He then made Gelsinger the first Chief Architect of the flagship 80486 processor project. Not bad for someone aged just 25 years old. Gelsinger subsequently became Intel’s youngest Vice President in 1992 at the age of 31 and its first official CTO in 2001.

All told, Gelsinger’s first stint at Intel lasted nearly 30 years, and Andy Grove mentored him for nearly all of that time. Gelsinger left Intel in 2009 to become EMC’s President and COO and then became CEO of VMware in 2012. Just remember the adjectives in the original Intel recruiting report: smart and aggressive.

But Intel Wasn’t a Microprocessor Company, at First

From the first paragraphs of this article, you might conclude that Intel was founded to design and make microprocessors. Nothing could be further from the truth. When Bob Noyce and Gordon Moore abandoned Fairchild Semiconductor in 1968 to found Intel, they had semiconductor memory dead in their sights. Microprocessors didn’t even exist back then.

The great thing about semiconductor memory is that your design task is greatly reduced because the memory cells are made using a step-and-repeat process, which greatly reduces the design cycle. Design one good memory cell and repeat, repeat, repeat. Gordon Moore was counting on that fact to speed Intel’s first products to market.

Intel set to work on designing two types of semiconductor memory: SRAM and DRAM. The first product out of the door was the 64-bit (not Gbit, not Mbit, and not even Kbit) 3101 bipolar SRAM, which Intel announced in 1969. Even though Intel was founded to design MOS chips, the 3101 SRAM employed a bipolar process because that was the fast path to a finished product and Intel needed to ship products quickly so it would have an ongoing revenue stream.

Because it was relatively easy to copy, the 3101 SRAM’s initial market success drew a slew of competitors, and Intel lost ground in the revenue department. As Intel’s May 17, 2018 Newsbyte about the 3101 SRAM explains, “…imitators followed suit and siphoned Intel’s profits.” That wasn’t in Noyce’s and Moore’s business plan, and it would not be the last time that Intel had this experience in the hotly competitive semiconductor memory market.

DRAM proved harder to design than SRAM because it used the less developed MOS process technology, not bipolar, and because storing bits as electrical charge on capacitors embedded into a chip was a fairly new idea, developed at IBM by Robert Dennard in 1966, but not commercialized. (Yes that Robert Dennard, who would later introduce the world to Dennard Scaling.) Lots more R&D and more time was needed to bring the world’s first DRAM to market.

Intel introduced the world’s first commercial DRAM, the 1103, in 1970. Initially, the device didn’t work very well and required several time-consuming redesigns to get it working. Eventually, Rev E of the mask design yielded working parts, but that’s a whole ‘nother story. Consequently, Intel wasn’t bringing in much money during those early years. Noyce and Moore needed another revenue stream to keep the doors open until they could get profitable standard parts into the market.

Intel Accidentally Created the Microprocessor Market

And so Intel turned to designing customer chips for other companies. These days, we call those custom-designed chips “ASICs” (a market that Intel has just re-entered in a big way under Gelsinger with the announcement of Intel Foundry Services). A Japanese calculator company named the Business Computer Corporation, nicknamed Busicom, was one of the customers that came knocking on Intel’s door to design and build some ASICs.

The mechanical calculator market, Busicom’s traditional market, was cratering, and Busicom had decided to get into the electronic calculator business. The company contracted with Mostek to develop an LSI chip for its simpler calculators, but Busicom’s engineers also had plans to develop a more complicated calculator. Not your typical four-banger, mind you, but a fire-breathing programmable calculator with lots of profitable options like a built-in printer. This calculator would be called the Busicom 141-PF. Masatoshi Shima at Busicom had developed specifications for 12 ASICs needed to implement the fully optioned calculator. Some fortuitous circumstances brought Busicom to Intel with a contract to design and fabricate the ASICs.

Intel assigned Ted Hoff to study the proposal and make recommendations. Eventually, Hoff realized that Intel would never be able to design and deliver that many unique ASIC designs in the allotted time or budget, so he developed an alternative based on his prior experience with minicomputers. Hoff proposed a 4-chip set including a 4-bit programmable processor, a 2-Kbit ROM, a 320-bit RAM, and a 10-bit shift register that serves as an output expander for the processor but curiously connects to the processor through a serial port on the RAM chip. (At least Intel already knew how to design RAMs even if it had little experience with the other devices.)

Busicom execs agreed to Hoff’s proposal and, after some months’ delay, Intel assigned Federico Faggin the task of designing the 4-bit processor. (Faggin also architected the follow-on processor, the Intel 4040, and led the development teams for the 8-bit Intel 8008 and 8080 microprocessors. He left Intel in 1974 to found Zilog, where he created the world’s most elegant 8-bit microprocessor, in my opinion, the Z80, but that’s yet another complicated story.)

Intel’s 4-bit processor, the 4004, would become the world’s first commercial microprocessor to hit the market, but that was not the original plan for this device. Because Busicom was underwriting the development of these ASICs, Intel would not have marketing rights. It was a proprietary chip, sold only to Busicom for the purpose of building calculators. However, by mid-1971, Busicom was having financial problems and petitioned Intel to cut the cost of the chip set to increase the profit on its calculator. Business-savvy Noyce saw an opportunity and cut a deal with Busicom that permitted Intel to sell the Busicom chip set, including the microprocessor, as a standard Intel product.

Intel thus created the microprocessor market through a series of unanticipated but fortunate events. The company started with the 4004 and speedily followed it with the 8-bit 8008, 8080, and 8085 microprocessors. Competitors quickly followed on the heels of the 4004’s introduction. They knew a good market when they saw one. By the end of the 1970s, Intel had also introduced the 16-bit 8086 microprocessor, which IBM designed into the IBM PC. (That too is another long and complicated story based on a series of events, both fortunate and unfortunate.)

Where’s Andy?

So where’s Andy Grove in all of this? Grove was working at Fairchild Semiconductor when Noyce and Moore were there. Noyce and Moore, along with the rest of the Traitorous Eight, had walked out of the Shockley Semiconductor Laboratory and founded Fairchild Semiconductor about a quarter of a mile down the road in 1957. Grove joined Fairchild in 1963 and became Fairchild’s Assistant Director of Development in 1967. When Noyce and Moore left Fairchild in 1968 and founded Intel, Grove was very, very close behind. Officially, Grove was Intel’s first hired employee, joining on Day One. However, Grove always considered himself an Intel founder. Justifiably so.

By the mid-1980s, an army of Japanese DRAM vendors were knocking the stuffing out of Intel in the semiconductor memory market with the sneaky dual-pronged strategy of offering better prices and higher quality parts. By then, Andy Grove had been Intel’s President for several years. In a 1985 meeting with CEO Gordon Moore, he reportedly asked a question about Intel’s increasingly problematic DRAM situation: “If we got kicked out and the board brought in a new CEO, what do you think he would do?”

“He would get us out of memories,” replied Intel’s CEO.

Grove and Moore then made the really difficult choice to cut their losses and exit the DRAM market, a market that Intel had literally created. Great companies are forged by making tough decisions like this one. Moore and Grove refocused Intel solely on microprocessors and closely associated logic chips. The rest is history. Microprocessors are a huge business now and, as of February 15, they’re officially Gelsinger’s business once again.

There are books aplenty written by and about Andy Grove, but if you want a top-10 list that gives you a two-minute sketch of the man, I recommend Betsy Corcoran’s article “Ten Lessons Learned from Andy Grove.” The picture Corcoran’s article paints suggests how uncomfortable Grove’s mentoring of Gelsinger might have been and provides some strong clues as to how Gelsinger plans to operate as Intel’s newly minted CEO.

Corcoran’s ten lessons learned from Andy Grove are:

  1. It’s OK to be scared.
  2. Relish reality.
  3. Discipline, discipline, discipline.
  4. When the going gets tough, suck it up.
  5. Teach with humility.
  6. Learn with passion.
  7. Did you screw up? Admit it.
  8. Science the shit out of your problems.
  9. Be real.
  10. Love fiercely.

To my eyes and ears, everything Gelsinger’s done to date at Intel follows these lessons closely. After all, Gelsinger had three decades to absorb these lessons from Grove, plus a lot more. Grove put a lot of energy and effort into teaching and growing his people, and it was not always fun being on the receiving end of that energy. Gelsinger confirmed the discomfort in a Tweet he posted just before stepping into the role as Intel’s newest CEO: “Mentoring with Andy Grove was like going to the dentist without Novocain.” Pretty clearly stated, I’d say.

Dealing with Intel’s widely publicized trials and tribulations due to the late development of its 14nm and 10nm process technologies and the late introductions of the x86 processors built on these process technologies are Gelsinger’s first big mountain to climb. That’s doubly true given the punishing blows that AMD’s latest x86 processors are delivering to Intel in the marketplace. Like Gelsinger, AMD’s current President and CEO Lisa Su is no pushover.

This isn’t the first time AMD has mounted an effective attack on Intel. It just hadn’t happened for many years. However, Intel is many times AMD’s size and has considerable resources that Gelsinger can bring to bear on this challenge. But he can’t afford to just beat back AMD once more. Gelsinger’s got to think of the long game too. He needs an effective strategy that will continue to work over the long haul – to permanently put Intel atop the mountain’s peak, if possible. To do that, he’ll need to bring out some heavy guns. If you look at what Gelsinger is doing at the moment, he’s strategically moving tens of billions of dollars into position for a variety of counter offensives.

My pretty obvious guess: He’s going to channel Andy Grove and that’s going to be a shock for the more than 100,000 people working at Intel. Andy Grove’s management style can be most genteelly described as “honest, deep, and tough,” clearly and colorfully described by Gelsinger’s Tweet, as quoted above.

Let Chaos Reign and then Rein in Chaos

In his book, “Only the Paranoid Survive,” Grove wrote: “…let chaos reign and then rein in chaos.”

Although Grove wasn’t technically an Intel founder, he always acted like one. Grove served as Intel’s President from 1979 to 1997, CEO from 1987 to 1998, and Chairman of the Board from 1997 to 2005. He died in 2016. No one has ever fought longer or more fiercely for Intel than Grove. No one worked harder to make Intel prosper. There was no bigger Intel booster than Grove. As people inside the company like to say, Grove clearly bled Intel blue.

By repeatedly saying that he wants to bring “Grovian Culture” back to Intel, Gelsinger’s intent could not be more clear, but as Robin Williams playing the Blue Genie in Disney’s “Aladdin” explained, bringing people (or at least their spirit) back from the dead isn’t done lightly. In the Disney version of the story, the Blue Genie grants Aladdin three wishes and then enumerates the three rules for making these wishes: 

“Rule #1: I can’t kill anybody, so don’t ask.”

“Rule #2: I can’t make anybody fall in love with anybody else.”

“Rule #3: I can’t bring people back from the dead. It’s not a pretty picture. I don’t like doing it!” 

If Gelsinger wants a role model for the Intel CEO position, Grove is clearly the right choice. It’s readily apparent that Gelsinger fully assimilated the ten Andy Grove rules listed above. What’s not quite so clear is whether or not he’s also familiar with the Blue Genie’s three rules – especially that last rule – but we’re sure going to find out. So will Intel’s competition.

Can Pat Gelsinger kill Intel’s competition?

Can Pat Gelsinger get the market to fall back in love with Intel?

Can Pat Gelsinger bring Andy Grove back from the dead, in spirit? 

Here’s how Gelsinger concluded the Intel Q2 2021 quarterly financial call on July 22:

“And before we sign off, one last opportunity to say that it is an honor to be back for my dream job, to run this iconic company at this pivotal time in the history of the semiconductor industry. We’re rebuilding our heritage of execution, innovation, and growth.”

Grab the popcorn. This is going to be a fun show to watch, even if it wasn’t made by Disney.

Leave a Reply

featured blogs
Sep 16, 2021
I was quite happy with the static platform I'd created for my pseudo robot heads, and then some mad impetuous fool suggested servos. Oh no! Here we go again......
Sep 16, 2021
CadenceLIVE, Cadence's annual user conference, has been a great platform for Cadence technology users, developers, and industry experts to connect, share ideas and best practices solve design... [[ Click on the title to access the full blog on the Cadence Community site. ]]...
Sep 15, 2021
Learn how chiplets form the basis of multi-die HPC processor architectures, fueling modern HPC applications and scaling performance & power beyond Moore's Law. The post What's Driving the Demand for Chiplets? appeared first on From Silicon To Software....
Aug 5, 2021
Megh Computing's Video Analytics Solution (VAS) portfolio implements a flexible and scalable video analytics pipeline consisting of the following elements: Video Ingestion Video Transformation Object Detection and Inference Video Analytics Visualization   Because Megh's ...

featured video

ARC® Processor Virtual Summit 2021

Sponsored by Synopsys

Designing an embedded SoC? Attend the ARC Processor Virtual Summit on Sept 21-22 to get in-depth information from industry leaders on the latest ARC processor IP and related hardware and software technologies that enable you to achieve differentiation in your chip or system design.

Click to read more

featured paper

An Engineer's Guide to Designing with Precision Amplifiers

Sponsored by Texas Instruments

This e-book contains years of circuit design recommendations and insights from Texas Instruments industry experts and covers many common topics and questions you may encounter while designing with precision amplifiers.

Click to read more

featured chalk talk

Power Profiler II

Sponsored by Mouser Electronics and Nordic Semiconductor

If you are working on a low-power IoT design, you are going to face power issues that can get quite complicated. Addressing these issues earlier in your design process can save you a lot of time, effort, and frustration. In this episode of Chalk Talk, Amelia Dalton chats with Kristian Sæther from Nordic Semiconductor about the details of the new Nordic Power Profiler Kit II - including how it can measure actual current, help you configure the right design settings, and show you a visualized power profile for your next design.

Click here for more information about the Nordic Semiconductor Power Profiler Kit II (PPK2)