The Rise and Fall of a Console King

    If Shakespeare was a playwright in Japan in the 1990s, he might have devoted a minor tragedy to Ken Kutaragi (below).

The Japanese engineer came up the ranks at Sony to become a father of the Playstation. It was the first videogame console to sell more than 100 million units, surpassing Nintendo and Sega who had been the leaders in video gaming, but he nearly lost the franchise with a risky bet on technology.

I interviewed Kutaragi twice. The first time was in March 1999, right after he presented at a top semiconductor conference details of the chips that would power the Playstation 2, eighteen months before the consoles would actually be available in U.S. stores.

By then, Kutaragi was already a media darling, a crowned prince of consoles holding court at a major confab of chip designers. He saw himself as an alpha engineer, so he granted interviews to deep tech press like EE Times.

I suspect he had a love/hate relationship with the limelight. His English was passable, but he was not fluent, a somewhat awkward situation for someone positioned as a tech visionary. His features were plain; he was not as handsome or charismatic as a Steve Jobs. And he lacked the laser focus on the product experience that helped folks like Jobs and Jeff Hawkins succeed. 

Sony’s PS2 strategy was similar to what worked for the original system: Beat sleepier rivals to the market with leading-edge tech and undercut them on price. Like the original Playstation, the first to use CD-ROMs, the PS2 would be the first to embrace the latest media, video-capable DVDs. Yet both consoles debuted for only $299.

People like Brian Halla were bullish. Halla led the consumer division at LSI Logic Corp. that helped design the main chips inside the original Playstation.

Kutaragi “had more impact on my career than any other customer I've ever known. I tell him I have a piece of my home named the Kutaragi wing,” Halla told me.

 

A Spark that Launched a Star

In our 1999 interview, Kutaragi recalled his early days as a humble Sony engineer, fresh out of college in 1975, writing low-level software for puny microcontrollers.  He found his calling a decade later when he was a corporate research engineer and bought one of the first 8-bit Nintendo NES videogame consoles.

“The graphics were very sophisticated if you compared it with one of the computers of that time,” Kutaragi said. “But the sound was terrible. I was frustrated that such a nice machine had such horrible sound.”

Seeing an opportunity, Kutaragi and a Sony salesperson met with executives from Nintendo to propose they use Sony’s audio technology. “We [had] designed a small [audio] chip and made an offer to Nintendo, and they picked it up in their 16-bit system, the Super NES,” said Kutaragi.

Emboldened by his success, Kutaragi made another proposal to Nintendo in 1989: the two should collaborate on the first CD-ROM-based console. A year later, Nintendo agreed and they were off to the races.

But in the summer of 1991, just before the new systems were set to be launched, Nintendo announced it was dropping its partnership with Sony and would work with Philips. Sony was reportedly demanding a cut of Nintendo’s content business while Philips was happy just to sell them the CD-ROM drives.

Sony decided it would build its own console “to beat Nintendo,” said Kutaragi.

Besides the first CD-ROM, the Playstation would boast a million-transistor system-on-a-chip to deliver state-of-the-art graphics. Kutaragi talked to every semiconductor company who would meet him in his quest for a partner to help design the mega-chip. Some weren't interested, others said it couldn't be done. Ultimately, Halla seized the opportunity for LSI Logic.

“Almost every night for two-and-a-half years we had conference calls on the project, with Kutaragi in attendance at most of them, mainly to go over engineering trade-offs,” Halla said. “The rest of the meetings were highly animated philosophical discussions about pricing.”

Some five years later, riding the Playstation’s success, Kutaragi raised the stakes. The PS2 chips would run ten times faster and deliver two hundred times more performance – but the box would carry the same $299 price tag. It would not only attack competition from Nintendo and Sega, but be positioned against home PCs selling for nearly $1,000.

Launched in 2000, the Playstation 2 became the biggest selling videogame console of all time. Kutaragi’s confidence swelled.

 

You Say You Want a Revolution

“Immediately after we completed the development of the PS2… I started thinking about what would be next, he told me in a 2005 interview about the ill-fated Playstation 3.

“I have aimed at bringing about a computer revolution, teaming with the world's largest computer company [IBM] and with Toshiba, a longtime partner since the days of the Playstation 2,” he said.

Together they designed a massively parallel microprocessor, the Broadband Cell Engine that would run ten times faster than the PS2’s CPU. The three companies sent dozens of engineers to an IBM facility in Austin, Texas, to design the processor. They promised it would someday power not only the Playstation 3 but all sorts of computers and consumer systems, taking on Intel’s x86 that dominated the burgeoning PC market. 

At a time when cloud computing was still in its infancy, Kutaragi’s vision was expansive.

“The Cell processor assumes the existence of multiple Cells,” he said, imaging an internet-like network of Cell processors.

“Servers around the world will form one virtual computer…One Cell or cluster of Cells can also function as a server; but whereas the present internet mainly handles characters, applications on the Cell network will also handle semantics and reasoning.

“Though sold as a game console, what will in fact enter the home is a Cell-based computer,” he said.

It was an engineer’s dream, one intoxicated on the rising prowess of electronics. “Sony will surely continue heading in a semiconductor-focused direction,” he said.

 

Not the Expected Results

In late 2006, the Playstation 3 (below) debuted with its Cell processor, a separate graphics processor designed by NVIDIA and a Blu-ray disc, bringing high-definition video to the console for the first time. But all that technology had its price -- $599 – more than twice that of rivals, and even that was not nearly enough. 

Analysts estimated it cost Sony $840.35 to build each unit. The Playstation 3 beat its upstart rival, the Microsoft Xbox, in console sales, but within a year Kutaragi’s division reported a nearly $2 billion loss. 

The Nintendo Wii (right), selling for just $249, beat the Playstation 3 in sales. Led by a game-developer-turned-manager, the Wii eschewed big, expensive chips. Instead, it used relatively cheap sensors to detect users’ gestures, enabling a new class of fun and easy-to-play games like bowling and tennis that were bundled with the consoles and got whole families up off the sofa. 

 

By contrast, the Playstation celebrated in Sony’s ads the quartet of confusing icons (a triangle, square, circle and cross) it used to identify the functions of buttons on its controller. It had its tech chic appeal for insiders, but the more intuitive Wii went viral.

Like Jobs and Jeff Hawkins, the Nintendo boss thought about what users would want, not about building the biggest chip.

The pain hit Sony’s partners, too. The massively parallel Cell, while powerful, was complex, and thus difficult to program. Developers shunned it.

The Cell found a handful of uses in arcade games, and a record-breaking supercomputer, but it was not enough to sustain future generations. IBM stopped work on the big chip in 2009.

By then Kutaragi was on the road to an early retirement, letting others drive the franchise forward. Like a tragic hero, he drove his vision beyond his limits and got humbled.  An engineer at heart, he neglected the art of content.

To Kutaragi’s credit, he lived in a crazy-making time.

In the 1990s, big companies and over-funded startups madly pursued a Holy Grail -- a mirage as it turned out -- of a kind of god-box of home electronics. What emerged over time was not one cool thing but four or five—laptops, flat-screen TVs, smartphones, consoles and more -- all capable of handling amazing network services for communications and entertainment with dazzling graphics, audio and video. 

The Cell processor was the poster child for a dozen massively parallel processors that companies, mostly startups, pitched to EE Times reporters. They all went belly up because they were too hard to program and existing PC processors were cheaper and powerful enough.

At least two people parlayed the lessons of the Cell processor into their prosperity.

Jensen Huang, the founder and CEO of NVIDIA, invested patient capital for a decade to build software tools around his company’s massively-parallel gaming GPUs in hopes of finding new, exciting apps for them. One day in 2012 a trio of researchers won a computer vision contest running a complex algorithm for what was then called deep learning on the GPUs. In a blog, Huang later called it "The Big Bang of AI."

Meanwhile, a former IBM exec behind the Cell, Lisa Su, stayed in touch with her Sony collaborators. She joined AMD in 2012 where a relatively small but successful unit wound up designing several of the custom chips used in videogame consoles, including the Playstation 4 and 5. Su, a distant cousin of Huang, became president of AMD and one of NVIDIA’s biggest rivals in AI.

Next: The Engineer Behind My Favorite Gadget