Q: As Europe's chip industry strives to achieve profitability, many European companies are implementing fab-lite models, consolidating and exiting the memory sector. As a European-based memory integrated device manufacturer (IDM), what has given Numonyx the rare competitive edge to remain successful in the devastated memory market?

A: Numonyx's competitive edge is due to the company's strong, unwavering market and technology leadership in today's memory industry. We are currently one of the top non-volatile memory (NVM) companies in the world and the number one supplier of NOR solutions for wireless and embedded applications. Our NOR market share has experienced remarkable growth, increasing 4 percent quarter-over-quarter (QoQ) in Q3 2009, while the number two supplier's share decreased 6 percent QoQ. In NOR's two focused segments, wireless and embedded, Numonyx is currently the number one and number two leader, respectively.

Q: In 2009, various analysts proclaimed the memory business model broken and advised memory companies to consolidate. However, in Q3 2009, the memory sector saw companies such as Samsung Electronics and Hynix Semiconductor report double digit growth. Do you see this positive trend continuing, or is consolidation still critical? As a company that was created by two semiconductor powerhouses, Intel and STMicroelectronics, how do you feel this combination of key memory technologies has benefited Numonyx's customers?

A: I would expect that the impressive financials of Samsung Electronics and Hynix Semiconductor are a result of the economy slowly recovering and the recent widespread consolidation. Numonyx experienced double-digit QoQ growth as well in Q3 2009, so we're quite pleased with our results. With that said, I certainly believe that further consolidation within the industry is still needed.

When customers select memory suppliers in today's industry, they are drawn to companies that deliver multiple technologies and invest in revolutionary solutions that will provide them value in the future. Through the combination of Intel's and STMicroelectronics' memory businesses, Numonyx is able to provide customers both of these benefits. The company offers customers a one-stop shop for all their memory solutions (i.e., NOR, NAND, random access memory (RAM) and phase-change memory (PCM)). Furthermore, the teaming of two prominent semiconductor leaders gives Numonyx customers confidence in the supply and product longevity of these technologies.

Regarding investment in innovation, Numonyx has put remarkable effort into the research and development of PCM, a breakthrough technology that promises to deliver solutions to many of the challenges facing the memory industry today. The merging of Intel's and STMicroelectronics' PCM development teams gives Numonyx the scale to significantly increase investment in PCM technology and bring it to the marketplace faster.

Q: Intel and Numonyx have aggressively pursued the development of PCM, a technology that fuses the speed of dynamic RAM (DRAM) with the non-volatility of Flash memory. With the cutting-edge capabilities and recent breakthroughs of PCM, many in the industry believe that it will replace Flash in future years. Do you foresee PCM being a "Flash killer," and what impact will it have on the other established memory technologies? What efforts has Numonyx made to speed the adoption of PCM in the marketplace?

A: I don't believe PCM technology is necessarily a "Flash killer," as NOR Flash and NAND Flash both have a long lifetime ahead of them and are used in many applications. However, PCM does combine the greatest attributes of DRAM, NAND and NOR, enabling more capabilities on a single chip. PCM's high-speed memory and its ability to erase the technology before writing it are certainly similar to DRAM, while its non-volatility parallels NAND and NOR. However, PCM cannot be directly substituted for DRAM, NAND or NOR technology, though it may attract a portion of their business. I strongly believe NAND and NOR customers will see and take advantage of this breakthrough technology's value, helping it grow into a potential multi-billion dollar business. Consequently, there is a lot of market development work involved to ramp PCM technology. We believe there are new applications primary to the computing market that PCM technology has the ability to develop, and we are currently focusing our efforts on this.

We have two different efforts underway to successfully introduce PCM technology to the market. The first effort involves reaching out to our current NOR and NAND customers to identify what PCM's increased functionality can provide them in terms of enhancing applications or reducing bill of materials (BOM). The second effort, as I previously mentioned, is to penetrate the computing market. The memory hierarchy in computing is already being changed by applications such as SSDs, so I believe the time is ripe to analyze alternative ways of architecting memory subsystems in computing platforms. We have received positive feedback from numerous data processing original equipment manufacturers (OEMs) that believe PCM has a strong value proposition in the industry.

Q: The commercial use of NVM technologies (e.g., ferroelectric RAM (FeRAM), PCM and magnetoresistive RAM (MRAM)) has been slow as their long-term reliability and scalability needs to be established. However, these technologies are starting to break ground as developers have made significant improvements in the technologies' level of integration and performance. What customer barriers must be overcome for these alternative technologies to potentially gain market share? What are the main challenges these new technologies face when competing with established memory technologies?

A: Aside from the fundamental technical challenges of new technologies, the barriers to adoption are generally related to cost and where in the system the memory sits. These two factors certainly influence the ability of customers to ramp new technologies.

Regarding cost, to gain market interest, I believe the alternative technology‘s expense must be five to 10 times lower than its competitors' if there is no other value proposition offered. As an example, PCM's cost is certainly not five to 10 times lower than its competitors', so Numonyx focuses on conveying the value of PCM's features to consumers. Compared to DRAM, PCM is much more energy-efficient. When no information is being read or written, there is no power consumption. Further, PCM's write endurance is far greater than NAND's and NOR's. As NAND and NOR scale, their limitations primarily arise from the fact that the number of times you can write to these technologies generation after generation is starting to shrink. We believe these benefits that set PCM apart from other memory technologies will allow it to gain significant market share in today's industry despite its costly price tag.

The time it takes for a new memory technology to be employed depends on if it is on a memory bus or abstracted interface. If a memory technology is on a memory bus (e.g., DRAM bus and NOR bus) that requires software and hardware changes, it generally takes a far greater amount of time to make these modifications. Additionally, multiple costs involving expense, risk and time arise. On the other hand, if a memory technology is on an abstracted interface (e.g., SSD and memory card), it is much easier for a customer to potentially adopt a new technology as they have the ability to test it and make sure it meets their performance and reliability requirements. However, the downside to this is that most abstracted memory devices/technologies are data technologies which are usually measured on a dollar per gigabyte basis.

The main challenge new technologies face when competing with established memory technologies deals with a cost/value tradeoff. As I previously mentioned, if a new technology is more costly than a traditional technology, then it better possess great value that customers are willing to pay for. An alternative, disruptive technology such as FeRAM has never heavily saturated the market because the technology wrestles with a severe cost disadvantage compared to NOR and offers no real value proposition. With PCM, Numonyx has been working with customers over the past several months to get them over the hurdles required to accept a new technology based off of what we think is a reasonable balance of cost and benefit.

Q: Taiwan legislators recently voted against any government funding efforts to restructure the island's embattled DRAM industry, arguing that DRAM prices and demand have recovered, eliminating the need for government aid. Though Taiwan only holds a 15 percent share of the global DRAM market, it does house six of world's top 10 DRAM manufacturers. Do you feel the Taiwan DRAM industry will be able to rebuild itself without government intervention? Does the Taiwan government's decision to not support its DRAM industry have any significant impact on the worldwide memory industry? Please explain why or why not.

A: It is hard to say if Taiwan's DRAM industry will be able to rebuild itself since my insight of this particular market is limited. However, given the current rebound in DRAM pricing, it certainly seems they have a better chance of bouncing back without government intervention.

If Taiwan's DRAM industry is incapable of correcting itself on its own, the potential effect on the global memory market will be quite significant. The impact felt by its competitors and customers will be great, as pricing and supply dynamics will shift. And it will not only be significant to DRAM. It will absolutely have an impact on NVM companies that supply both DRAM and NAND.

Q: Today's industry is seeing many hard disk drive (HDD) manufacturers enter the crowded, competitive SSD market. Gartner estimates unit growth of enterprise SSDs to double and its sales to reach $1.0 billion in 2010. With NAND forecasted to generate more than 20 percent of Numonyx's revenue in 2009, what potential do you believe the SSD market holds in driving NAND demand? Do you believe SSDs will eventually replace HDDs in the PC market?

A: The SSD market is currently in its infancy phase, so its growth potential is great and could go through the roof. I believe the adoption rate of SSDs will dramatically increase in 2H 2010, making it an interesting time for the NAND industry. We are certainly keeping a close eye on the SSD market because if the demand for SSDs exceeds capacity, then the impact on average selling prices (ASPs) will be dramatic.

It has taken the industry longer than expected to fully comprehend the true value of SSDs over HDDs. Over the next couple years, I believe SSDs will gain significant share of the PC market. The cost/value trade-off I discussed earlier comes into play with SSDs. It is well known that SSDs cost more per gigabyte than HDDs. I recently paid 7 cents a gigabyte for 1.5 terabytes, totaling only $105 for a HDD. SSDs are clearly not as cheap, but there is value in its cost of ownership (i.e., cost to build and maintain it and liabilities) and performance. There is something extremely nice about picking up my laptop without shutting the lid and walking around the office, conference room to conference room, without worrying about crashing a hard drive. Over the next couple quarters, I predict SSD builders will create a new price point with a lower density SSD that will have more than enough density for consumers in the workplace.

Q: As 3D technology furthers the advancement of transistors, through-silicon via (TSV) wafers are predicted to represent 25 percent of the memory business by 2015, creating great technological changes within the memory market. How do you see 3D technology developing in the next five years? What challenges arise with 3D TSV technology?

A: When I worked at Intel in the late 1990s, the push we got from customers (primarily wireless customers) to integrate more silicon per square millimeter was incredible, and as a result we pioneered many chip scale and multi chip packages. There will be further requirements to improve upon existing, simple processes such as stacking chips. I cannot forecast what percentage TSV wafers will represent in Numonyx's business or the industry's business in 2015. But I do know that the requirement to adapt and make changes will always be there, and as a memory supplier, we will be ready for it.

One of the main challenges of 3D TSV technology is, of course, cost. In the memory business, we must be able to provide this technology at a low cost. As well, clearly, there are the original technical challenges we faced when we first started stacking chips, and we need to make sure there aren't any new challenges. As engineers, it will be our job to solve that.

Q: The semiconductor initial public offering (IPO) market is showing positive signs of recovery as four semiconductor companies filed in November 2009. As a company that expects to be profitable in the coming quarters and with a healthier IPO environment, has Numonyx pondered growing its business by going public? Looking forward, what other expansion strategies has the company considered implementing?

A: After working at two public companies, I believe there are many benefits available to a private company. As a new company, we are currently enjoying those benefits. Of course, there is always the possibility of becoming a public company in the future, but we have no plans to file an IPO at this time.

I believe that PCM's numerous benefits (i.e., high read/write speeds, low volatility and high storage density) will allow the company to develop and enter new, untapped markets. I strongly believe that using PCM technology to penetrate these markets will help drive Numonyx's expansion and present us with many other future opportunities.

Back to Industry Reflections Home