Periodic Reporting for period 1 - Fe-NVRAM (Solving the scaling challenge of the memory industry: high-speed, low-complexity and low-cost non-volatile ferroelectric memory (Fe-NVRAM) made in EU)
Période du rapport: 2024-04-01 au 2024-12-31
The semiconductor industry is a 600B$ industry. The share of semiconductor memories is 150B$. But there are no memory manufacturers in Europe. So, the EU is missing out on a huge share of the semiconductor market!
FMC will change that – by creating fundamentally new memories based on our disruptive material innovation, we will become a major memory player from Europe, reducing the EU’s critical dependence on overseas suppliers, creating jobs, and making semiconductors greener.
Our memory is as fast and low cost as DRAM, while providing non-volatile data storage.
Our memory products will significantly reduce the energy consumption of data centers, enable instant-on capabilities, and double the battery life of smartphones, PCs, and IoT devices. In addition, non-volatile working memory immediately protects against data loss due to power outages.
Our vision is to disrupt the memory industry and become the first EU memory company, set up production facilities in Europe to become the next EIC unicorn.
FMC will change that – by creating fundamentally new memories based on our disruptive material innovation, we will become a major memory player from Europe, reducing the EU’s critical dependence on overseas suppliers, creating jobs, and making semiconductors greener.
Our memory is as fast and low cost as DRAM, while providing non-volatile data storage.
Our memory products will significantly reduce the energy consumption of data centers, enable instant-on capabilities, and double the battery life of smartphones, PCs, and IoT devices. In addition, non-volatile working memory immediately protects against data loss due to power outages.
Our vision is to disrupt the memory industry and become the first EU memory company, set up production facilities in Europe to become the next EIC unicorn.
A DRAM DDR-4 design is in development for a 25nm DRAM technology, which uses ferroelectric layers in the DRAM capacitor instead of dielectric layers. This FE-vnRAM is non-volatile, this nonvolatility saves energy 1) during stand-by (refresh every 64ms is not needed, 2) during operations (refresh is not needed as well) and especially in AI-centers, due to the fact that data-traffic between fast DRAM and non-volatile NAND can be avoided. This is possible because the FE-nvRAM combines DRAM features like speed and low-cost with non-volatility.
While the advantages at the memory component level are significant, the benefits that can be realized at the system level are even more obvious. FMC evaluated several use cases from the server and datacenter segments and quantified the respective benefits. With FE-nvRAM, in-memory databases performance can be boosted by 2x, memory-buffer database performance will improve by 1.9x the speed of random writes in storage applications can be performed at maximum speed, and AI training loads can run more reliable and faster - yielding a customer return-on-investment of more than 1000%. Therefore FE-nvRAM will have a significant impact on the performance, cost, and energy efficiency of datacenters soon.
While the advantages at the memory component level are significant, the benefits that can be realized at the system level are even more obvious. FMC evaluated several use cases from the server and datacenter segments and quantified the respective benefits. With FE-nvRAM, in-memory databases performance can be boosted by 2x, memory-buffer database performance will improve by 1.9x the speed of random writes in storage applications can be performed at maximum speed, and AI training loads can run more reliable and faster - yielding a customer return-on-investment of more than 1000%. Therefore FE-nvRAM will have a significant impact on the performance, cost, and energy efficiency of datacenters soon.
Replacing a dielectric capacitor in a DRAM technology by a ferroelectric one leads to massive innovation, if not disruption beyond global state of the art on every single level of the DRAM-technology: e.g. with respect to 1) underlying physics: ions in a crystal, flipping back & forth are the two bits of the capacitor, "1" and "0". In contrast to a capacitor filled with a dielectric, the ferroelectric capacitor is leakage immune. 2) with respect to process/ technology-level, the ferroelectric layer of HfO2 or HZO is "doped" with some percent foreign atoms to generate & stabilize ferroelectricity. 3) on device level the ferroelectric capacitor increases retention of a 1T/1C DRAM from 64ms to years, furthermore ions in a crystal are immune to radiation. 4) On design-level , the ferrolectric layers make the DRAM a non-volatile FE-nvRAM, well suited for most advanced non-volatile DDR4 and LPDDR4 which need no refresh every 64ms and no refresh circuitry at all. This saves power and cost massively and allows for further disruption on product and system level: a DRAM is a fast & high density memory at the price of volatility (that means the stored information gets lost if the power is turned off). FE-nvRAM is fast and of high-density and at the same time non-volatile (that means it retains the information for months or years). In consequence, a FE-nvRAM can be a one-chip solution to DRAM/NAND or DRAM/NOR combinations, especially in AI datacenters, where up to 50.000 GPUs and 2 GB/ GPU of memory chips are needed in a system.