How Adesto’s CBRAM surfs the wearable and Internet-of-Things wave
New technologies for non-volatile memories are being developed, with market momentum growing over the last three years (Source, Yole Développement report Emerging Non-Volatile Memory 2016). Adesto Technologies has developed CBRAM, a resistive RAM technology.
As the company is publicly listed, it is much easier to understand the pros and cons of this technology and what is really happening, especially with the growth of wearable applications and Internet-of-Things (IoT) trends. Yole Développement interviewed Ishai Naveh, Vice President of Marketing and Business Development at Adesto to discover more.
Yole Développement: Can you briefly introduce Adesto, its history and current activities? What have you been doing in non-volatile memory (NVM) in the last year?
Ishai Naveh: Adesto is mainly a product company; we develop and ship non-volatile memory. We are shipping CBRAM products, but they’re on a longer growth trajectory, so most of our current activities are around floating gate memory technology. We have both CBRAM and Flash products. We are open to licensing CBRAM but right now this is a small portion of our business. Licensing at some point might become a larger focus but right now I think most companies are taking a wait-and- see approach. Other companies will say that they have a foundry-like agreement that includes licensing, but actually the foundry agreement is the main thing that people do and the license is a small part. We have several years’ experience producing CBRAM at 130nm and we have been selling product since we ramped to high-volume. Last February, we signed an agreement with TowerJazz Panasonic in Japan to port the technology to 45nm.
(Source: Emerging Non-Volatile Memory report, Yole Développement)
YD: What products are available now from Adesto?
IN: In CBRAM, we ship products that compete with EEPROM. We also brought out a product that’s low density like EEPROM but more aimed at energy harvesting, or battery powered devices. It’s extremely low power, both in standby and when operating. We have more than 500 customers worldwide using Adesto Flash and CBRAM products, and we continue to supply them with memories that meet the specific needs of their applications in consumer, industrial, communications, medical and computing markets focusing on low-power and the IoT.
Adesto Operating Principle (Courtesy of Adesto Technologies)
YD: What are the different low power IoT applications targeted by Adesto?
IN: We’re targeting IoT applications with both Flash and CBRAM. We just issued a big press release and gave a presentation at the Linley Group Processor Conference introducing our new Flash product called EcoXiP for eXecute-in-Place systems that more than doubles processor performance, lowers system power consumption and reduces system cost compared to standard serial peripheral interface (SPI) devices. In CBRAM, we’re going to the extreme low end of power consumption.
Adesto EcoXiP on board (Courtesy of Adesto Technologies)
Applications such as energy harvesting or execute-in-place can be almost anywhere. Systems are emerging that can’t afford DRAM from a cost or power consumption point-of-view, but still need more memory than on-chip Flash or SRAM. Many systems also want batteries to last a year or more. This is a whole new paradigm. There are many applications here, going from switches on the wall to all kinds of wearables where people want the battery to last one year to five years. Agricultural is an area that I now think will ramp very strongly; it’s definitely a market where you want a battery for several years. These types of systems require an order of magnitude lower energy consumption than traditional systems. Of course, people on the system side are working on this, while we are working on the memory side, but we see memory as part of the system. It’s not a boot memory any more that stores code or data, but it’s part of the system. DRAM and even SRAM are power-hungry in the IoT era. In many applications you might only operate 1% of the time, so standby current is a killer. Active current should also be controlled. But standby like DRAM or SRAM, you just cannot live with that. Many of these applications are still being developed and customers don’t want their information given out, so that’s why I’m hesitant here, but the whole idea is you need to get the power consumption to a different level.
YD: So low power is the market driver for your solutions?
IN: Yes, for the whole company, it’s not just CBRAM, this is the company’s strategy. This is what we do best and the area that we go after.
YD: What is your technological roadmap for your emerging NVM CBRAM technology?
IN: The roadmap for CBRAM is that we’re going from 130nm to 45nm first of all. 130nm didn’t allow us to do much in the sense of density, so we were very limited. 45nm will allow us to go up to probably several tens of megabits. Since we focus on application-specific memory, we’re going to do memories with unique features in power and performance, like the product that we launched last month using Flash, in the future we may launch a similar one in CBRAM which will be even lower power and higher performance.
YD: What is your production infrastructure? Are Altis and TowerJazz Panasonic your two main foundry partners? IN: For CBRAM, TowerJazz Panasonic is still in development, but yes, these are our main partners. YD: How do you think a fabless company can be successful in the highly concentrated memory business?
IN: The foundry market today is concentrated. It can be a challenge to find a partner, but with our technology’s lower power consumption we did find common ground. Part of the value that you bring to a foundry is that they will be able to use it for embedded applications. After we complete the 45nm project, it will be much easier because it will be qualified as a mainstream technology and the value will be much easier to see. YD: We see an increase in new technologies related to this non-volatile memory, CBRAM, which you’re developing, but also magnetic RAM, on phase-shift memory and so on, there are multiple activities around that and you’re benefiting from that. IN: We see the emergence of other memories but it’s not very strong; it’s not like foundries feel they must have it. If you remember 15 years ago when phase change just started, everybody had to have phase change. It’s not the same situation now. Yes, there are some, but they’re pickier and it’s an uphill battle. If you look at the market there are only a few companies that really have new memory in volume production like Adesto and Everspin in MRAM. Everybody else is talk and presentations. Noise from these other companies that don’t have products yet is enough to muddy the water, and so people say why should I choose you when I can choose them? You can look at what kind of great numbers they show, but they show it on a single cell, they don’t have a product. It makes things more difficult, but fine. We have our foundries so it’s less important now.
YD: What are the objectives and challenges of your company in the next three years?
IN: Our objective is to grow with the products, both Flash and CBRAM. To be successful in the market you need to show continued growth. Then of course there is the technology, to bring it up in 45nm and qualify it. In 45nm, we already have the architecture and design mature enough to make interesting products. And our new EcoXiP product represents a new architecture and a big change for the whole industry. The big challenge for us will be driving adoption over the next year.
YD: Is there anything else you want to say?
IN: I want to emphasise that strategically we’re not trying to compete with Samsung and Toshiba on cost per bit of mass storage. If they’re interested in CBRAM, we’d be happy to license it to them and have them bring up the terabit products. This is not a game for a new company. But we made the technology, and we are bringing it to the level of tens of megabits, and that will allow them to take it to the next level. But overall we think that the point of our technology is to compete on low power, high performance, and for embedded applications, the simplicity of integrating CBRAM into the manufacturing process. We think that this is the future. Battery-powered systems, systems that operate a small proportion of the time, and systems that are battery-less, that harvest energy. As a company and as a technology, this is where we can win.
Ishai Naveh, Vice President, Marketing
Ishai Naveh is a co-founder and the Vice President of Marketing and Business Development for Adesto Technologies. In his role, he is responsible for forging partnerships as well as positioning the company in growing, lucrative markets. Previously Mr. Naveh, was the VP of Marketing for Non-volatile Memories and Mixed Signal Technologies at Tower Semiconductor USA. Prior to arriving in the USA, he was Senior Director for foundry technologies at Tower. Mr. Naveh began his career at National Semiconductor, holding several positions in the USA and Israel. Mr. Naveh holds a B.Sc. from the Hebrew University and MBA from Heriott-Watt University. He has published several papers and has been issued more than 10 patents.
Related report:
Emerging Non-Volatile Memory 2016 report
The new storage-class memory (SCM) category will be the biggest emerging NVM market, with embedded MCU seeking emerging NVM for strategic differentiation
More information here
Source: www.yole.fr / www.adestotech.com/