2021 MAY 8 ISSUE: INTEL’S $3.5 BILLION WILL SHIFT TO NEW MEMORY CHIPS

With a $3.5 billion investment, Intel’s research and development at its campus on the escarpment above Corrales is transitioning to full-scale production of laser-based, three dimensional computer chips. During an announcement May 3, Intel officials said Fab 11X and other parts of the complex will be re-built and reconfigured to mass produce the innovative 3-D, layered chips referred to as Foveros technology that allows customized mix-and-match assemblages, expected to be most useful for advanced artificial intelligence and 5G applications. Long-time Intel New Mexico manager Mindy Koch told Corrales Comment at the news conference that new chip-making equipment will be added although exactly what and where are still being evaluated.

Having joined Intel as a summer intern in 1988, Koch, a University of New Mexico trained electrical engineer, is now corporate services site manager. Around 1,000 construction jobs will begin this year, and around 700 permanent jobs will be added over the next five years to the current Intel workforce of some 1,800. Changes to Intel’s air pollution permit from the N.M. Air Quality Bureau are anticipated, she said, although the changes needed would likely be achieved through yet another modification or revision. As of last year, Intel’s permit had undergone 10 revisions to 11 modifications (Intel’s permit number is 325-M11-R10).

When Intel began its R&D on the new photonic devices here, Koch said the new process uses essentially the same industrial chemicals as other microchips produced here in the past. For decades, Intel’s permit from the Air Quality Bureau has allowed use of more than 80 chemicals federally classified as hazardous air pollutants. (See Corrales Comment Vol.XXVII No. 23 January 23, 2010 “Dallas Region EPA Stages Surprise Inspection at Intel.”)

The announcement that Intel will invest $3.5 billion on the project was applauded by the following who attended: Governor Michelle Lujan Grishan, U.S. Senators Martin Heinrich and Ben Ray Lujan, Congresswoman Theresa Leger Fernandez and the mayors of Corrales and Rio Rancho. Corrales’ Erika Edgerly was master of ceremonies, introducing all of the above and the main speakers: Intel Senior Vice-president Keyvan Esfarjani who is general manager for manufacturing and operations as well as Intel New Mexico Site Manager Katie Prouty.

The $3.5 billion project apparently will not need yet another round of industrial revenue bonds (IRBs) since proceeds remain from earlier bonds. Intel New Mexico began in 1980 with an IRB offering of $2 billion; another $8 billion was provided in 1993, and then $16 billion in 2004. In June 2019, the Sandoval County Commission extended Intel’s time frame for using the last bonds.

At the May 3 announcement, Intel’s Esfarjani offered a relatively non-technical explanation of how the new computer chips work. Intel had partnered with another chipmaker, Utah-based Micron Technology, in 2015 to develop the devices that stack layers of data storing units with a precision not previously possible. The stacked technique was touted at the time as yielding three times more capacity than similar products. “This enables more storage in a smaller space, bringing significant cost savings, low power usage and high performance to a range of mobile consumer devices as well as the most demanding enterprise deployments,” the partners claimed.

When Intel announced it would conduct R&D on the breakthrough at the facility here in 2016, the concept was described as “moving data with light,” a marketing motto accompanied by the following description. “Silicon photonics is a combination of two of the most important inventions of the 20th century: the silicon integrated circuit and the semiconductor laser. It enables faster data transfer over longer distances compared to traditional electronics, while utilizing the efficiencies of Intel’s high-volume silicon manufacturing.”

Intel and other firms in the high-tech world had been working with photonics for more than a decade, but engineers at the Rio Rancho plant were credited with having moved the technology into production mode. “Intel has always been at the forefront of this exciting new technology, announcing the world’s first hybrid silicon laser nearly a decade ago,” Intel’s general manager for data centers, Diane Bryant, said at the time.

“We are the first to light up silicon. We integrate the laser light-emitting material, which is indium phosphide, onto the silicon and use lithography to define the laser, to align it with precision.” Bryant’s remarks came at the August 2016 Intel Developer Forum in San Francisco. Products for data centers were expected to be Intel’s biggest emphasis in years ahead. But, for some reason, Intel’s partner, Micron, grew disillusioned.

In March of this year, Bloomberg News reported that Micron was abandoning the technique. Its headline said “Micron Give Up on Intel-Championed Memory, To Sell Factory.” The news agency reported that Micron “no longer sees a business for the memory chips that would justify the expense of continued development, and will invest elsewhere.” Intel’s leadership thinks otherwise.

In his remarks May 3, Esfarjani expressed total confidence in the “integrated device manufacturing (IDM) approach. “A key differentiator for our IDM 2.0 strategy is our unquestioned leadership in advanced packaging, which allows us to mix and match computer tiles to deliver the best products. We’re seeing tremendous interest in these capabilities from the industry, especially following the introduction of our new Intel Foundry Services. “We are proud to have invested in New Mexico for more than 40 years, and we see our Rio Rancho campus continuing to play a critical role in Intel’s global manufacturing network in our new era of IDM 2.0.” At the May 3 event, no one mentioned Micron’s decision, at least not near a live mic.

Late last month, an Intel official offered a little background on its differences with Micron. As reported by TechTarget.com, Intel’s director of technology initiatives, Jim Pappas, said during a panel discussion on persistent memory techniques that “We were as surprised as anyone when Micron made their announcement” to close the factory that the two chip giants had jointly funded and operated to produce the 3-D chips, referred to as “3D XPoint.”

The chips are used to store “non-volatile memory,” in that they retain stored information even when the power is off. Intel’s plan is based on its innovative a “phase change memory” approach, the market for which has been predicted to reach $46.5 billion by 2026.

Intel uses its Optane brand chips for solid state drives that have the same function as a computer hard drive, but use flash-based memory. Micron Senior Vice-president Rajeeb Hazra last month explained his firm’s decision to abandon the approach on which Intel now focuses this way. “The question really is how much value did we see in attaching and focusing on 3D XPoint versus other alternatives that CXL [compute express link] enabled.

“As we looked at our technology capabilities, talked to customers and looked at multiple options, we came to the conclusion that it is better for our customers, and better for us, to go down a path of different kinds of things attached to CXL in the memory space than 3D XPoint, and therefore we terminated our 3D XPoint program.”

Or maybe Micron knew by then that Intel would be ramping up 3D XPoint products at the New Mexico factory. The two chipmakers ended their joint operation of the plant in Utah last year, but Micron continued to supply Intel with 3D XPoint wafers on which Optane units are embedded. So when Micron closed of that plant, basically Intel had few options besides making its own if it intended to move ahead with the new product. And Intel New Mexico was chosen to accomplish that.