Double the memory, half the space

Your alarm clock goes off. You stagger to the kitchen and switch on your Nespresso® machine. You turn on the radio and quickly check your emails. You grab some milk out of the fridge for your cereal. You put your empty bowl into the dishwasher. You get your towel out of the dryer, take a hot shower, get dressed, and jump in the car to go to work. You may not realise it, but nearly every part of your daily life is dominated by integrated circuits – or chips.

Chips are the core to making electrical devices – from your phone to the thermostat on your shower – work. If you’ve ever suffered from a slow computer, you’ll know that one of the most important components of a chip is its memory. As devices evolve, they are using more data, and the need for more memory grows. But memory takes up space and chips need to be small to be produced at low cost. 

We spoke to Dr Robert Giterman, CEO, and Professor Andreas Burg, a co-founder, to find out a little more about what their technology is and its potential uses. 

How did RAAAM Technologies start? 

Andreas: My lab at EPFL is the Telecommunications Circuit Lab. We spent a lot of time working on the pieces that go into mobile phones – looking at how we could improve how phones compute things. We would make a piece half the size and be very proud of it, but it would sit next to a huge memory element. It didn’t make sense. So we thought, let’s look at the biggest piece in the chip – the memory. If we can save maybe 10% of the size of the memory component, it’s better than making the data part 99% smaller. 

What exactly does your technology do?

Robert: Our technology enables us to reduce the size of the memory block in a chip. By doing this, we can reduce the overall size of the chip and therefore reduce the price for the companies who make them. Our research, which has been ongoing for nearly a decade, allows us to store data using far fewer components compared to what’s currently on the market. This means two things. First, chip design companies can save money on manufacturing. Secondly, they can increase the amount of memory on a chip, increasing the capacity of the device and leading to better performance. Your laptop or phone will be faster and use less power, which leads to longer battery life. 

Andreas: One of the problems we’re seeing at the moment is as technology evolves, it needs more memory. There are so many great ideas, applications, out there which can’t work simply because chips can’t hold enough memory. Our technology is an enabler – it means new ideas and tech, such as AI, are no longer blocked by memory issues. 

How does your technology work?

Robert: At the moment, there are two main memory technologies available. First, there’s SRAM – or Static Random Access Memory. This is the technology which is most widely used. It means that data stored in the memory will remain there as long as the battery works. Our technology is based on embedded DRAM – Dynamic Random Access Memory. This means that from the point the data is written to the memory, it starts to degrade. After time, it disappears. In order to stop the data disappearing, the chip performs a refresh, it rewrites the data back into the memory. This technique of storing data has the benefit of using about half the space of SRAM. Furthermore, unlike traditional embedded DRAM solutions, our technology is fully compatible with the standard process used to make integrated circuits, therefore there’s no extra cost or complexity to the manufacturing process.

Does the refresh process have a risk of data loss, or is the risk worth it because of the size implications?

Andreas: Reliability is a key concern and it’s something a lot of people ask, all the time. You don’t want to make a billion chips and then some of them don’t work. The simple answer is no, there’s no risk of data loss. We’ve been working on this project for nearly 10 years and during that time, the technology we’ve been working with has progressed. Now we can show manufacturers the results from our test chips and the data which proves our method works.

Robert: That’s part of the reason this journey has taken the time it has. We have concentrated on being able to prove our technology works, and that it works at scale. 

How will the Innogrant help you?

Robert: We’ve already got about 10 prototypes we’ve developed in the course of our research. These are our test chips – the ones we rely on to prove our technology works. We’ve started talking to a number of companies we hope will become our future customers. One of the key ways the Innogrant will help us is that it’ll allow us to continue talking to these companies and address any concerns they may have. We’ll be able to show the benefits of our technology in their existing systems. Our goal by the end of the grant is to have two or three firm customer commitments. We’ll then work with them to tie up all the loose ends - designing and finalising the technology and integrating it into the existing systems. Eventually, we hope to sell them the license to include our technology in their products! 

Andreas: RAAAM is at an exciting stage of maturity and we’re also looking for people to join us on our journey. Research and development is at the core of our business and we’re currently searching for highly qualified people to join us, particularly if their expertise is in design and R&D. 

What’s next for RAAAM Technologies?

Robert: Ultimately, our product is not a separate physical chip. We won’t be looking to manufacture chips ourselves. We are going to license the technology – a chip company will buy rights to use our black box of files and then be able to integrate our technology into their chips and sell them on. 

Andreas: Our technology isn’t highly visible – it’s not an app, it’s not a medical device, it’s not a robot. But our technology has the potential to be everywhere, in all these applications. From simple washing machines to Google’s data centre. In a couple of years we hope our technology will be part of every mobile phone, every laptop, every device that needs memory.