Coby Hanoch comes to Weebit Nano with 15 years’ experience in engineering and engineering management and 26 years’ experience in sales management and executive roles. Coby was Vice President Worldwide Sales at Verisity where he was part of the founding team and grew the company to over $100M in annual sales which facilitated its acquisition by Cadence Design Systems (NASDAQ:CDNS). He was also Vice President Worldwide Sales at Jasper, doubling sales before it was acquired by Cadence. Coby was brought in as CEO to help PacketLight avoid bankruptcy and get it back to a leadership position in its domain. Prior to Weebit, Coby set up his own consulting company, EDAcon Partners, helping startups define their corporate strategies, to set up their worldwide sales channel and raise capital.
What is Weebit Nano’s backstory?
Weebit Nano delivers the industry’s most advanced memory technologies to help semiconductor companies and foundries easily and cost-effectively differentiate their products. We’re bringing to market a new type of non-volatile memory (NVM) called ReRAM (Resistive Random Access Memory) which will be the successor to flash memory for the many applications that need better performance, reliability, power consumption and cost.
Weebit was founded in 2015 and is headquartered in Israel, with R&D teams in Israel and France. Our R&D partner, CEA-Leti, is one of the world’s most advanced microelectronics research institutes, and together we’ve created a truly innovative NVM technology based on over a decade of research.
Weebit is focused on creating advanced technologies that are also economically viable. Even the most cutting-edge technology won’t succeed if it isn’t affordable and easy for customers to integrate and manufacture. Our focus on both innovation and commercial success comes from the deep experience of our executive team and Board. Even just looking at four key people – our Chairman Dadi Perlmutter, Directors Atiq Raza and Yoav Nissan Cohen and myself, together we have over 150 years of combined industry experience with companies including AMD, Intel, National Semi, Tower Semi, Cadence, and others. A similar depth of experience is found across the company.
Weebit is a public company listed on the Australian Stock Exchange (ASX:WBT). Being publicly listed is a great way to accelerate technology development since it gives us immediate access to the financial markets when fund raising is needed, and it also provides transparency that translates to customer and partner confidence.
What makes Weebit ReRAM unique?
Today, flash is the most common NVM, and while it has done a great job to-date, flash has limits in terms of speed, power, cost and endurance. As an embedded technology, it also can’t scale to the most advanced process nodes with the rest of a chip. So, for these and other reasons, the industry needs a new NVM. Of course, it has to be a technology that doesn’t require heavy investments and also one that can use existing manufacturing techniques.
Weebit ReRAM can scale well below 28nm, and it has much faster memory access time and higher endurance than flash, as well as lower power consumption and operating voltage. It can also maintain data at high temperatures for many years, a requirement of many companies we’re talking to. Our ReRAM is also based on the most common materials used in fabs, and uses standard tools and processes, so it can be easily integrated into any standard CMOS flow. It needs only two additional masks, versus around 10 for flash, so added wafer cost is minimal.
ReRAM is also a back-end-of-line (BEOL) technology, an advantage over flash which is a front-end-of line (FEOL) technology that often requires designers to make compromises with analog components and devices. ReRAM doesn’t have this problem, and you can also adapt ReRAM once for a technology node and it works for all its variants.
So that’s all compared to flash, but Weebit ReRAM also wins on almost every parameter compared to other emerging NVMs like MRAM. This is true on tech specs, and more importantly when you look at the simplicity of our technology and how easy it is to integrate into existing processes, translating to lower cost and lower risk.
What market segments are you targeting?
Nearly every electronic device in the world is a potential target for our ReRAM. Digital products that wake up, run code, sense the environment, process and store data need NVM. Adoption timelines in different applications vary, due both to end market requirements and ReRAM’s rollout, which is starting in small densities. Our first offering will be ReRAM IP that customers will embed in their designs. Discrete chips will come later.
Embedded applications for ReRAM are fairly endless and span different process nodes, memory sizes and usage profiles. There are short-term opportunities in areas like power management ICs and other analog designs where BEOL NVM technology is a true advantage, and in areas like IoT and MCUs where ReRAM checks all the boxes for low power, low cost, a high level of integration, plus endurance in harsh conditions. Over time, we’ll see ReRAM in areas like edge AI, industrial and automotive, and there are longer-term opportunities in neuromorphic computing, where the properties of ReRAM mean it can efficiently emulate neural networks in an analog way, versus the simulations you see today.
What keeps your customers up at night?
Weebit’s customers are fabs and semiconductor companies. Obviously concerns vary, but we know they are looking to deliver designs to spec, on time, and on quality. They are also focused on innovating with technologies that help them differentiate against their competition, while maintaining a competitive price.
As companies look beyond 28nm, scaling challenges for embedded NVM become a real concern. For other designs, it could be the cost of NVM process integration, for example in analog and power flows. Flash is expensive and complicates the design, while forcing design constraints. Of course, power is always a concern, where ReRAM demonstrates an order of magnitude improvement over flash. The list of specific design concerns is long, and ReRAM can help customers solve such challenges.
Of course, selecting a new technology like ReRAM is a strategic decision because customers need to know the IP works. This is where our qualification results are key. We’re already sharing initial results, and potential customers are extremely impressed with what they’re seeing. As we continue towards production and deliver final results, it will give customers the level of confidence they need to integrate Weebit ReRAM into their designs.
On a personal note, what was your path to joining Weebit?
I joined Weebit as CEO in 2017 after spending almost 40 years with global EDA, IP and semiconductor companies, in CEO and founder roles as well as engineering and sales positions. At the time, I had been in discussions with Weebit’s chairman Dadi Perlmutter about a Board position and then, when the previous CEO asked to step down for personal reasons, he asked me to become CEO.
Since joining Weebit, my goal has been to focus our efforts and drive meticulously toward mass production. This meant making the early decision to first concentrate on developing IP solutions for the embedded market, and also focus on standard materials and processes. This doesn’t mean we aren’t continuing our efforts toward discrete products – we are actually making good progress in that area – but the main effort is on completing successive milestones towards a commercial IP solution. If you look back over the last several years, you can see we’re doing just that.
What’s next for Weebit Nano?
Weebit is making great strides toward commercialization. Currently, we are qualifying our ReRAM module with CEA-Leti, and while Leti isn’t a production fab, they have a very advanced R&D fab, so the results are significant. We expect to complete that full qualification before the end of the year. We’ve also finished technology transfer to SkyWater Technology, which is the first time we’ve transferred to a production fab, and we expect those wafers back before the end of the year. We’ll then begin that qualification, with expected results in early 2023. We have many other initiatives underway, including making Weebit ReRAM available in 22nm FD-SOI.
This is all pushing Weebit embedded ReRAM technology toward mass production. As I mentioned, we’re also working toward a mid-term goal of discrete ReRAM chips where the key is novel selector technology, and we’re making solid progress together with Leti.
I am highly confident that 2023 will be a banner year for Weebit. The focus, flexibility and excellence the Weebit team exhibited throughout the pandemic – meeting and in some cases exceeding development milestones – was impressive. This sets the stage for continued stellar execution, and we couldn’t be more excited about the opportunities in front of us.
How do customers engage with you?
We are already in discussions with multiple semiconductor companies and fabs that want to get an early advantage with Weebit ReRAM IP, and of course we’d love to engage with other forward-thinking companies. Contact email@example.com to get started! SkyWater Technology customers can also reach out to their SkyWater representatives directly to begin designs.