- Texas Instruments licensed embedded ReRAM from Weebit Nano, boosting confidence in ReRAM as a flash alternative.
- Embedded flash is harder to scale below 28nm, pushing designers toward external flash and SRAM workarounds.
- Weebit’s ReRAM fits into existing chip flows with minimal changes and claims much faster writes than flash.
- The universal memory race continues, but TI’s move raises the stakes for next generation embedded storage.
A tiny memory startup has just landed a heavyweight backer, and it is not a quiet one. Texas Instruments, a roughly $170 billion US chip giant, has signed a licensing deal for embedded ReRAM technology from Weebit Nano, putting fresh momentum behind the idea that flash memory may be running out of road.
The move matters because TI is not just another name on a partner slide. It is one of the most influential suppliers of chips that power cars, factories, medical devices, and industrial electronics. When a company like that starts aligning with a new kind of non volatile memory, the rest of the embedded ecosystem pays attention.
Weebit Nano’s pitch is direct and bold: resistive RAM can do what embedded flash does today, but faster, simpler, and with fewer headaches as chipmaking pushes into smaller nodes.
Why flash is hitting friction below 28nm
Embedded flash has served the industry well for years, especially in microcontrollers and system on chip designs that need non volatile storage on the same die. But shrinking it further has become increasingly difficult, particularly once designs move below 28nm.
At that point, engineers often end up using workarounds. A common approach is to pair logic with external flash, then stage code into SRAM during boot. It works, but it adds complexity, increases power consumption, and introduces more moving parts into the design. It also creates new security concerns because more data is traveling between components.
Weebit’s CEO, Coby Hanoch, argues that ReRAM sidesteps these compromises by bringing non volatile storage back onto the chip in a form that scales more cleanly. That means quicker boot times, tighter control of sensitive data, and less dependence on external memory.
What makes Weebit’s ReRAM different in manufacturing terms
Weebit is not trying to bulldoze existing chipmaking flows. Instead, it has positioned its ReRAM as a back end of line module. In plain terms, it fits into the later stages of manufacturing without forcing fabs to redesign the front end transistor structures.
That detail is critical because it reduces friction for adoption. According to Weebit, the added wafer cost sits around 5%, which is far more digestible than the overhead typically associated with embedded flash processes.
Each ReRAM cell stores information using resistive switching rather than the floating gate mechanisms used in flash. The practical benefit is that it supports direct bit level access without the block erase behavior that makes flash slower and less flexible for certain workloads.
Weebit’s approach is also intentionally conservative in the best way. The company says it relies on standard materials, standard tools, and standard fabrication flows. The message to manufacturers is simple: you do not need a custom factory or exotic equipment to build this.
Speed, endurance, and the “better than flash” argument
On paper, Weebit’s claims are designed to grab attention. The company reports write speeds up to 100 times faster than embedded flash. It also cites endurance ranging from 100,000 to 1 million write cycles, a range that can suit many embedded applications where frequent updates matter.
Power is another part of the story. Weebit says ReRAM can run at lower voltages and consume less energy thanks to direct access modes. For edge devices and battery powered systems, that can translate into longer life and less thermal stress.
Hanoch’s view is that ReRAM beats flash across the categories designers care about most. That includes speed, endurance, temperature handling, power consumption, and cost. It is a sweeping claim, but it is exactly the kind of confidence that tends to surface when a company believes the timing is finally right.
Weebit also highlights resistance to electromagnetic interference, positioning it as a safer choice in environments where other memory types can be vulnerable. It even points to concerns around MRAM in consumer settings where magnetic exposure could become a real world issue.
A long build, not an overnight breakthrough
The TI agreement did not arrive out of nowhere. Weebit has been stacking manufacturing relationships over time, moving from SkyWater to DB HiTek to Onsemi, and now to Texas Instruments.
Hanoch describes it as a deliberate climb, with each partner representing a jump in process scale and credibility. The message is that this is not a single lucky deal, but a pattern of increasing industry trust.
Even so, the company remains early in revenue terms. Weebit points to institutional caution as the biggest brake on adoption rather than technical readiness. In other words, the technology may work, but the industry takes time to commit when memory is foundational to product reliability.
The universal memory race is still open
ReRAM has long been discussed as a candidate for “universal memory,” the dream technology that combines the speed of SRAM, the density of flash, and the persistence of non volatile storage. This deal pushes that conversation forward, but it does not close it.
There are still other contenders and new experiments entering the ring, including alternatives like ULTRARAM. For now, TI’s decision signals that embedded flash is no longer the only game in town, and that serious companies are preparing for what comes next.
If ReRAM delivers on its promise at scale, it could reshape how embedded devices store code, protect data, and boot instantly. The real test will be whether it moves from qualified silicon to widespread deployment across the products people actually buy.
Follow TechBSB For More Updates
