The following was provided while participating in the 2023 Destination Startup® Showcase.
Return to Destination Startup® Information Page
What We Do
Codelucida provides disruptive error-correction solutions supporting the next generation of densest NAND flash memory chips. As the memory densities are increased to enable higher storage capacities, the memories become more prone to errors, and their performance degrades. Codelucida’s solutions provide the strongest error correction with highly scalable performance and reduced power.
Affiliated Institution: University of Arizona
Have we formed a company? Yes
Funding: Grant Funding, Angel Funding (including Self or Friends/Family), Venture Capital
Our Company and the Problem We Are Trying To Solve
Driven by AI, data analytics, IoT and 5G, the amount of data that needs to be stored is increasing at an exponential rate, and flash memories are the core chips used for data storage in those applications (including consumer, data center, embedded, edge, and industrial storage applications). In response, the industry is rapidly increasing the densities of these chips so that they can store more data. These increases however are accompanied by severe degradation in reliability and performance. At the same time, the performance requirements for storage are doubling approximately every two years. Therefore, stronger, more efficient, error correction is critical in order to enable denser flash memories to meet the reliability and performance targets. Further, conventional error-correction solutions also consume too much power and silicon area, posing a major challenge for the industry. With a founding team that has decades of experience in error-correction, we are addressing this challenge by utilizing a combination of unique patented algorithms (protected by multiple issued US patents) and proprietary know-how that were developed based on years of research by the founding team. Our solutions provide the strongest error-correction capability with more than 2X savings in area and power while also greatly simplifying the memory management. Our technology has already been adopted by a leading US SSD manufacturer for defense (that has gone to mass production) and we recently secured multiple ASIC chip design wins adopting our solution for a embedded memory chip (used in industrial/automotive), and a high-volume consumer storage (SSD and microSD).
Our Go-to-Market Strategy
Codelucida's technology is licensed as custom-designed semiconductor intellectual property (IP) cores that are integrated into flash controller chips. Therefore, our target customers include flash controller chip makers, solid-state drive makers building in-house flash controllers, and data storage system builders that directly control/manage the data access from the flash memories. The global flash controller chip market into which Codelucida's solutions will be licensed into was estimated to be $3B in 2021 and expected to grow with a CAGR of 7.7% until 2028. Initial target customers will be the tier-3 customer prospects who don’t have in-house expertise and are in need of a solution that will take our solution in silicon to the market. As our IP penetrates the market, this validation will help us gain traction with the tier-2 customer prospects (leading flash controller suppliers) and eventually the tier-1 prospects (NAND flash memory makers and hyper scalers). We will also engage with key partners in the eco-system including the foundries and memory chip makers For example, we are already an IP partner for AMD-Xilinx, a leading supplier of FPGA chips (programmable chips) for FPGA-based storage applications (our first customer used an FPGA-based flash controller). The key differentiators of our technology, which is based on a completely different underlying algorithm called FAID® , include: 10%-15% increase in error-correction capability which translates to 4 orders of magnitude reduction in error rates, highly scalable performance achievable by a single core, at least 2X reduction in power and resource usage, and simplifying media management.
How We Will Generate Revenue
Our business model is a licensing model where our IP is licensed to customers in the form of encrypted IP cores to be integrated into the flash controller chips and eventually become part of the silicon once the chips are taped out. The structure of the licensing fees will depend on the customer, the nature and scope of the application as well as the customer’s volumes. Licensing fees may include upfront fee, NRE fees, one-time project fees, and per unit royalty fees. For the customer prospects with lower volumes, we license on a flat fee per project basis with the license (and scope of use) being limited to the project, with the expectation of receiving additional projects, whereas for customers with high volumes, royalties are charged. Ultimately, the objective for licensing with each customer to ensure that there is a recurring component to the revenue received, either through recurring projects, or through royalties.
How We Will Benefit From Destination Startup®
We anticipate a $2M fundraise in Q2 of 2023. Based on the milestones that have been achieved this past year, we would aim for some level of participation from strategic investors in the round that can help us further penetrate the market. Given the timing of the showcase and our own timeline for a fundraise, the showcase seems like a good opportunity for us to be able to engage and seek interested investors (angel, venture, or corporate capital) as well as strategic partners who may be interested in our technology either as a direct customer or as a key partner to enable their customers or suppliers.
Our Team
Codelucida’s management team includes the co-founders and co-inventors of the company’s proprietary FAID algorithm listed below. The founders are widely regarded experts for error-correction in the industry. – Dr. Shiva Planjery, as part of his doctoral research, co-invented novel decoding algorithms that laid the foundations for the company’s patented technology. – Dr. David Declercq, was a fully tenured professor at ENSEA in Cergy-Pontoise, France, having published more than 150 papers in the area of LDPC code and decoder design. Several of his contributions towards error correction algorithm design have been employed by the industry as well as adopted in several standards. Dr. Declercq resigned from his academic position in France to relocate to Tucson, Arizonan July 2017 to be full-time CTO of the company. Since then, he has continued to drive the innovation in the company leading to rapid advancements in the technology. – Dr. Bane Vasic is a professor of Electrical and Computer Engineering at the University of Arizona and an IEEE fellow. In addition, Codelucida has an experienced board of directors that includes: - Wayne Rickard, Member of Tech Coast Angels and Arizona Tech Investors, former VP of Seagate, storage industry veteran of 20+ years - Gopi Sirineni, CEO of hardware-based security startup Axiado, former VP for Qualcomm-Atheros, former executive at AMCC and Marvell, semiconductor industry veteran of 20+ years - Tom Lathrop, former Investment director at Intel Capital for 11+ years where he led many investments in the flash memory/controller space that had successful exits.