Kilopass 'Introduction to VLT Technology for Low Power, Reduced Manufacturing Costs, and Improved Performance'
Unlike capacitor-based DRAM technology, it does not require refresh, so power consumption is also improved.

Was it because there wasn't enough time to demonstrate, considering they claimed to be a 'game changer' in the DRAM market?

Kilopass Technology, Inc., which claims to be a leader in the OTP memory market, took the time to introduce its VLT (Vertical Layered Thyristor) technology, stating that it is a technology that will change the landscape of the DRAM market in the future.

Kilopass's VLT technology is based on thyristor technology, a complex structure electrically equivalent to the cross-coupled pairs of bipolar transistors that form a latch. This structure is suitable for memory applications because it stores values, and unlike current capacitor-based DRAM technology, it is characterized by not requiring a refresh. Since thyristors were first developed in the 1950s, there have been several attempts to utilize them in the SRAM market, but none have been successful.

▲ Kilopass CEO Charlie Cheng

In other words, this means that the company has succeeded in applying thyristor technology, which eliminates the need for capacitors, to a standard DRAM structure that uses capacitors and transistors. VLT explained that it realized a DRAM bit cell based on vertically implementing a thyristor structure as an auxiliary device for a 4.5F2 high-density cell structure.

Then, what kind of changes does applying VLT technology bring to DRAM?

First, it was stated that new materials and physics are not required. It was also explained that the advantage is that data such as yield can be identified 15 months before mass production using a simulator that is 1,000 times faster. Since more than 30 test programs are conducted and various voltage ranges are supported, there is no need to add anything entirely new. This implies a 45% reduction in costs and ultimately leads to the ability to apply a safe process.

In particular, VLT added that it can scale directly down to 7nm because it has eliminated complex capacitors. This is particularly noteworthy because while logic foundry companies producing processors, SoCs, and NAND flash memory products are already ahead at process nodes below 10nm, capacitor-based DRAM technology is lagging behind by as much as two generations.

In addition, the absence of capacitors eliminates the need for complex and power-consuming refresh operations. According to data from the manufacturer, VLT-based DDR4 DRAM can reduce standby power consumption by 10 times to 50 fA/bit and increase performance by up to 15%.

The company stated that VLT bitcell operation and silicon measurements were completed in 2015, and the memory macro test chip is currently in the early stages of testing. The full macro-level test chip was tape-out completed in May, and the first chip (silicon) test is currently underway.

"In licensing discussions with major DRAM manufacturers"

Charlie Cheng, CEO of Kilopass, who visited Korea, said, “Kilopass is currently actively discussing business partnerships and licensing with major DRAM manufacturers,” adding that he could not disclose further details due to ongoing discussions with clients.

When asked about his ambition to prioritize the PC and server markets and his view on the stagnating DRAM market, he responded optimistically, “It is true that DRAM in the PC market is stagnating, and I am aware that NAND flash is expanding in the server market. However, as the social media and cloud markets grow, the unique advantages of DRAM will continue.”

At the meeting that day, Kilopass began by stating that innovation usually originates from startups, citing examples such as Qualcomm, SanDisk, and Atheros, which competed against established powerhouses with innovative technology. The implication was that they, too, would change the landscape of the market like those companies; however, the meeting left much to be desired in terms of an explanation regarding their 'strategy,' which was just as innovative as their 'technology,' specifically what they planned to do and how they intended to change the landscape of DRAM.