UltraRAM’s Quantum Leap: How Quinas Is Turning Universal Memory Into Commercial Silicon

UltraRAM Just Made Every Other Memory Type Look Outdated

Estimated reading time: 10 minutes

Key Takeaways

• UltraRAM is designed as universal memory that combines RAM-like speed with storage-like persistence.
• Quinas Technology has prototypes, government funding, a major manufacturing partner, and a roadmap to commercial silicon.
• UltraRAM’s TBRT structure enables extreme retention, endurance, and low energy switching compared to DRAM and NAND.
• Key near-term focus areas include industrial-scale wafer production, CMOS integration, and targeting high-performance markets first.

Memory technology has not changed much in decades. You have got RAM, which is fast but forgets everything the moment you cut the power. You have got flash storage, which holds data but wears out quickly and runs slow. For years, engineers have dreamed of a single chip that does both jobs perfectly. That chip is called UltraRAM, and it is closer to reality than most people realize.

This is not a lab rumor or a speculative white paper. A real company, Quinas Technology, has prototype chips, government funding, a major manufacturing partner, and a clear roadmap to commercial silicon. What started as a physics project at Lancaster University in the UK is now moving through formal industrialization programs backed by millions in public and private investment.

In this article, we break down exactly what UltraRAM is, how it works, what has been built so far, and why it matters for the future of AI, neuromorphic computing, and every device that relies on memory chips.

What UltraRAM Actually Is and Why It Matters for Universal Memory

The term universal memory gets thrown around a lot. What it means, in plain English, is a single type of memory chip that replaces both RAM and storage. Right now, your computer uses at least two separate memory systems because no one technology does everything well. DRAM is fast but volatile — it loses data when power stops. NAND flash holds data permanently but is slow, and it degrades after a limited number of writes.

UltraRAM is designed to solve both problems at once. Here is how it stacks up against what we use today:

• Data retention: UltraRAM can hold data for an extrapolated 1,000+ years without power. Some Lancaster University calculations suggest the underlying physics could theoretically retain data longer than the current age of the universe.
• Write endurance: It handles at least 10 million program-erase cycles. Standard TLC 3D NAND flash starts degrading after just a few thousand writes.
• Speed: UltraRAM is designed for 1-nanosecond write operations, which is approximately 10 times faster than DRAM.
• Energy use: Switching energy is 100 times lower than DRAM fabricated on the same node, and about 1,000 times lower than NAND.
• Latency: The target is one-tenth the latency of current DRAM.

These numbers are not minor improvements. They represent a different category of device entirely.

The technology behind this performance is a structure called a Triple-Barrier Resonant Tunneling (TBRT) layer. Like flash memory, UltraRAM uses a floating gate to store charge. But instead of the silicon oxide layer that flash uses — which gradually degrades every time data is written — UltraRAM uses this TBRT structure. The key difference is that this barrier does not wear out during write and erase cycles, which is the root cause of the extreme endurance numbers.

The quantum mechanical process called resonant tunneling allows electrons to pass through this barrier extremely quickly and with very little energy. That is where the speed and efficiency advantages come from. It is not a clever engineering trick layered on top of old physics. It is a fundamentally different operating principle.

Lancaster University's research team progressed from 4-bit ULTRARAM arrays to a patent-pending memory architecture over the period from 2018 to 2022. Their simulations and physical tests confirmed a highly compact, high-bit-density design with ultra-low energy per unit area. That foundation is now what Quinas Technology is working to commercialize.

Lancaster University invents universal memory technology

Simulations of resonant tunnelling through InAs/AlSb heterostructure

How Quinas Technology Is Building the Path From Lab to Commercial Silicon

Quinas Technology was spun out of Lancaster University's Physics Department with one specific mission: take UltraRAM from academic prototype to sellable chip. Since its formation, the company has moved through several formal funding and validation programs that tell a clear story of progress.

The Innovate UK Awards

The UK's national innovation agency, Innovate UK, runs a rigorous program called ICURe (Innovation to Commercialisation of University Research). Quinas graduated from this program and then secured a £300,000 Innovate UK award specifically to commercialize ULTRARAM as "universal computer memory." This funding also included sponsorship for a Department for Business and Trade mission to Taiwan to identify potential manufacturing partners.

This is significant. Innovate UK does not hand out money for ideas. The award is given for demonstrating commercial viability and leading-edge science. Getting this grant means Quinas had to prove, in front of experienced evaluators, that the technology had a realistic path to market.

On top of that, Quinas was also supported under UKRI's "Developing Semiconductor Hardware for Critical Technologies" scheme. Under this program, Quinas explored ULTRARAM for neuromorphic computing and secure, energy-efficient AI systems. The fact that it was classified under "Critical Technologies" places it alongside other hardware the UK government considers strategically essential for national capabilities.

The IQE Partnership and the £1.1 Million Industrialisation Project

The single most important development in recent UltraRAM history is the partnership with IQE plc, one of the world's leading compound semiconductor wafer suppliers. Together, Quinas and IQE completed a £1.1 million Innovate UK-funded industrialisation project with Lancaster University and Cardiff University as partners.

The goal of this project was specific and practical: develop a scalable production method for ULTRARAM wafers. Most of that £1.1 million was deployed at IQE specifically to advance epitaxial wafer production at industrial scale.

This is not a research grant. This is a manufacturing development project. The difference matters enormously. You can have a perfect chip design, but if no one can produce the wafers reliably and cheaply enough to sell, the technology goes nowhere. This project directly attacks that challenge.

UltraRAM requires very precise semiconductor layers — specifically GaSb, InAs, and AlSb — deposited using a process called Molecular Beam Epitaxy (MBE). This is the same process the Lancaster lab uses in its cleanroom equipment. Scaling MBE to industrial production is one of the key technical challenges the IQE partnership is tackling.

Why Taiwan Matters

Tom's Hardware's original lab visit report noted that QuInAs had been at SEMICON Taiwan 2023 as part of a UK Pavilion exhibition. The signals from those discussions suggested that manufacturing partners may be found in Taiwan rather than Europe.

This is a strategic choice. Taiwan's semiconductor ecosystem — anchored by TSMC and a deep network of specialized suppliers — is the most advanced in the world. If Quinas can find integration pathways there, it dramatically increases the chance of getting UltraRAM into real products at volume.

Innovate UK award for Lancaster spinout Quinas Technology

IQE and Quinas Technology complete £1.1m UltraRAM industrialisation project

Quinas - Semiconductor Today

Quinas Technology Ltd case study

Stop Running Your Business on Manual Processes

Most businesses waste hours every week on tasks that could run automatically — data entry, notifications, approvals, follow-ups. While you're reading about the future of computing hardware, the smartest operators are already automating the repetitive work that slows their teams down.

Flowgrammer builds custom business automation systems that remove manual bottlenecks, free up your team, and let your operations scale without adding overhead.

👉 See How Automation Can Free Your Team

UltraRAM's Role in Neuromorphic Computing and AI Hardware

The most forward-looking application of UltraRAM is not just making your laptop faster. It is enabling a new class of computing entirely: neuromorphic computing.

Neuromorphic chips are designed to mimic the structure of the human brain. Instead of shuttling data back and forth between separate memory and processing units — which wastes enormous amounts of energy — neuromorphic systems process data where it is stored. This "in-memory computing" approach is far more efficient for AI workloads like pattern recognition, sensor processing, and real-time inference.

The problem with building neuromorphic hardware today is that conventional memory technologies do not have the right properties. DRAM is too energy-hungry. Flash is too slow and wears out. What neuromorphic chips need is memory that is:

• Non-volatile (holds state without power)
• Extremely fast (sub-nanosecond switching)
• Ultra-low energy (to support dense, always-on operation)
• Highly durable (millions of write cycles per cell)

UltraRAM checks every one of those boxes. This is exactly why Quinas pursued funding under the UKRI "Critical Technologies" scheme specifically for neuromorphic and secure AI applications.

The Memory Guy, an independent memory industry analyst, has noted that UltraRAM's use of III-V materials — specifically InAs and AlSb — gives it extraordinary intrinsic retention. Lancaster's team has extrapolated that for their InAs/AlSb system, data storage times could theoretically exceed the age of the universe. However, the same analyst also flags the real challenge: III-V wafer costs are significantly higher than silicon, and manufacturing scalability is harder to achieve compared with conventional NAND or DRAM production lines.

That concern is exactly what the IQE partnership is designed to address.

For secure AI workloads, UltraRAM offers another attractive property: because data persists without power, systems can be shut down and restarted without reloading data from external storage. This reduces the attack surface for certain types of side-channel attacks and simplifies secure enclave design. These properties are why Quinas' work under the Critical Technologies program specifically highlighted secure AI systems alongside neuromorphic computing.

IIT Roorkee in India has also been brought in as a modeling partner. Their role is to simulate UltraRAM performance in wider application contexts — essentially helping map the technology onto real-world system architectures so that engineers designing AI and neuromorphic chips know exactly how to integrate it.

Quinas - Semiconductor Today

University of Lancaster invents yet another memory

Simulations of resonant tunnelling through InAs/AlSb heterostructure

The Real Obstacles Still Standing Between UltraRAM and Your Devices

Let's be honest about the challenges. UltraRAM is genuinely impressive on paper and in the lab. But there is a significant distance between a working prototype and a chip in a consumer device or a data center rack.

Manufacturing Cost of III-V Materials

The semiconductor layers required for UltraRAM — gallium antimonide, indium arsenide, aluminium antimonide — are compound semiconductors, often called III-V materials because of their position on the periodic table. These materials have remarkable electronic properties, but they are more expensive and more complex to manufacture than pure silicon.

Current DRAM and NAND fabs are built around silicon processes refined over decades. Transitioning even a portion of that infrastructure to III-V materials requires either converting existing equipment (expensive and risky) or building new fabs from scratch (even more expensive). The IQE partnership is working on epitaxial wafer production, but making those wafers cheap enough for mass-market memory is a multi-year challenge.

Integration With CMOS Logic

Modern chips do not just have memory — they have memory sitting directly alongside logic circuits. For UltraRAM to reach its potential as a universal memory, it needs to integrate cleanly into CMOS manufacturing processes. III-V semiconductors and silicon CMOS do not naturally mix well. There is ongoing research into hybrid integration methods, but this is still an open engineering problem.

Scaling to Smaller Nodes

Lancaster's lab equipment can produce UltraRAM features down to about 20nm. That is a respectable starting point, but leading-edge memory production today operates at 1X nm nodes and below for NAND, and sub-10nm for DRAM. Scaling UltraRAM's TBRT structure to smaller nodes while maintaining its performance properties is unproven at this stage.

The Market Entry Strategy

Quinas has been clear that their initial production runs will target the top end of the memory market — the highest-performance, highest-margin segment. This is the right strategy for a startup with limited production capacity. High-performance computing, AI accelerators, aerospace, defense, and secure government systems can justify premium pricing in ways that consumer electronics cannot.

This also means ordinary consumers will not see UltraRAM in their phones or laptops for quite some time. But proving the technology in demanding professional applications creates the track record needed to eventually drive costs down for broader markets.

UltraRAM

First steps towards revolutionary UltraRAM memory chips

Universal memory research passes new milestone

Where UltraRAM Stands Right Now in 2026

The picture in mid-2026 looks like this:

• Quinas Technology is an active company with multiple completed Innovate UK projects and a clear pipeline of technical milestones.
• The IQE industrialisation project has been completed, delivering a scalable production methodology for ULTRARAM wafers.
• Lancaster University and Cardiff University remain active research partners, feeding technical insights back into the commercial development process.
• IIT Roorkee is modeling UltraRAM in broader system contexts, helping identify target applications.
• Manufacturing partnership discussions in Taiwan have been ongoing, with the Department for Business and Trade actively supporting outreach.
• The technology has been validated under three separate Innovate UK and UKRI programs — a strong signal of credibility within the UK's national semiconductor strategy.
• Interest from major tech players, including Meta, has been publicly noted, particularly around UltraRAM's energy efficiency claims.

The technology is not shipping at volume yet. But it has moved decisively out of the "interesting research" phase and into the "serious industrialization" phase. That is a rare and meaningful transition for any new memory technology.

The next major milestones to watch for are announcements around manufacturing partnerships — particularly in Taiwan — and the first results from small-volume production runs targeting high-performance compute markets.

Innovate UK award for Lancaster spinout Quinas Technology

IQE and Quinas Technology complete £1.1m UltraRAM industrialisation project

Quinas Technology Ltd case study

FAQ

1. What is UltraRAM in simple terms?

UltraRAM is a new type of memory chip designed to combine the best qualities of RAM and flash storage into a single device. It is fast like RAM, holds data permanently like a hard drive, uses very little energy, and lasts far longer than either current technology. It is what engineers call "universal memory" — one chip to replace both.

2. Who invented UltraRAM?

UltraRAM was invented by researchers at Lancaster University's Physics Department in the UK. The technology was later spun out into a commercial company called Quinas Technology, which is now leading efforts to bring it to market.

3. How does the Triple-Barrier Resonant Tunneling (TBRT) structure work?

The TBRT structure is a set of very thin semiconductor layers that acts as a highly resistive barrier. When a voltage is applied, electrons "tunnel" through this barrier via a quantum mechanical process — they pass through without physically going over it. This tunneling is extremely fast and uses very little energy. Because the barrier does not degrade during these operations, the chip can endure millions of write cycles without wearing out.

4. How does UltraRAM compare to NAND flash storage?

UltraRAM is dramatically better on almost every metric. It has at least 1,000 to 10,000 times more write endurance than standard NAND. It operates roughly 1,000 times more efficiently in terms of switching energy. And it retains data for over 1,000 years, compared to the 10 years typically cited for modern NAND flash under normal conditions.

5. Is UltraRAM faster than DRAM?

Yes, according to Lancaster University's research and Quinas Technology's claims. UltraRAM targets 1-nanosecond write operations, which is approximately 10 times faster than DRAM. It also targets one-tenth the latency of DRAM and uses one-hundredth the switching energy of DRAM on a comparable manufacturing node.

6. What is Quinas Technology and how is it funded?

Quinas Technology is a university spinout company from Lancaster University, formed specifically to commercialize UltraRAM. It has received a £300,000 award from Innovate UK after completing the ICURe commercialisation program. It also co-led a £1.1 million Innovate UK-funded industrialisation project with IQE plc, Lancaster University, and Cardiff University. Additional support came from UKRI's "Developing Semiconductor Hardware for Critical Technologies" scheme.

7. What is IQE's role in UltraRAM development?

IQE plc is a leading compound semiconductor wafer supplier. IQE and Quinas completed a £1.1 million jointly funded industrialisation project to develop a scalable production method for ULTRARAM epitaxial wafers. Most of the project funding was deployed at IQE's facilities, making them central to the effort of making UltraRAM manufacturable at commercial scale.

8. Why is UltraRAM relevant to neuromorphic computing?

Neuromorphic chips mimic the brain's structure and process data in-place rather than shuttling it between separate memory and processor units. This demands memory that is non-volatile, ultra-fast, ultra-low energy, and highly durable — all properties that UltraRAM is designed to deliver. Quinas explicitly explored UltraRAM for neuromorphic and secure AI applications under a UKRI Critical Technologies grant.

9. What are the biggest challenges UltraRAM still faces?

The main challenges are manufacturing cost and scalability. UltraRAM requires III-V compound semiconductor materials (InAs, GaSb, AlSb) that are more expensive to produce than standard silicon. Integrating these materials into conventional CMOS manufacturing flows is technically complex. Scaling the TBRT structure to the smallest modern transistor nodes is also not yet demonstrated. These are significant engineering and cost hurdles, but they are being actively addressed through the IQE partnership and ongoing research.

10. When will UltraRAM be available in consumer products?

There is no confirmed timeline for consumer availability. Quinas has indicated that initial small-volume production will target high-performance, high-margin markets such as AI accelerators, defense, and secure computing — where premium pricing is justified. Broader consumer adoption, if it happens, would follow after the technology's cost-effectiveness is proven at scale. Realistically, consumer-facing products featuring UltraRAM are likely many years away.