Weebit Nano advances second customer toward ReRAM production

Weebit Nano has begun silicon validation with a second customer whose chip is running software from integrated ReRAM, marking another step towards potential production in 2027. The company also added two AI customers, advanced qualification work and ended the quarter with A$164.2 million in cash.

  • Second customer begins silicon validation with working ReRAM samples
  • Customer targets production in calendar 2027, subject to validation
  • onsemi starts qualification while DB HiTek and Texas Instruments programs progress
  • Two new AI customer agreements signed during the quarter
  • A$6.1 million in customer receipts and A$164.2 million closing cash
An image related to Weebit Nano Ltd Add us as a preferred source on Google
Image © middle. Logo © respective owner.

Second customer reaches silicon validation

Weebit Nano Ltd (ASX:WBT) has moved a second customer into silicon validation, with initial samples successfully running software from the company’s integrated resistive RAM. The customer is targeting production in calendar 2027, although that timetable remains subject to completing validation.

The milestone gives Weebit two customers testing working silicon, according to the company. It follows the earlier stage at which two customers had taped out chip designs incorporating Weebit ReRAM, with one functional prototype already reported in May through the two customer tape-outs. The identity of the newly validating customer was not disclosed.

Qualification work expands across major chipmakers

Weebit said onsemi began qualification of its ReRAM during the September quarter, while work with Texas Instruments continued as planned. Collaboration with DB HiTek also progressed after qualification, with the parties preparing the technology for high-volume manufacturing as customer interest increases.

Evaluations with several potential new fab customers continued, but the timing of those programs is being shaped by fab capacity and engineering-resource availability. Texas Instruments was identified as Weebit’s largest licensee in the company’s FY26 results, when multiple product tape-outs and qualification milestones were also reported through the Texas Instruments licensing program.

AI pipeline gains customers but remains developmental

Weebit signed agreements with two new AI customers during the quarter to explore different uses for in-memory computing, meeting one of its key calendar 2026 targets. It also received positive analysis results from a US-based customer studying ReRAM for an in-memory-computing accelerator aimed at harsh-environment applications.

For the specific AI task assessed, the study projected more than 80 times lower power consumption than a GPU-based implementation, roughly 280 times higher compute efficiency than an FPGA-based accelerator and inference latency below one millisecond, compared with tens of milliseconds. Those are performance projections from a study, not commercial operating results; Weebit said significant research and development remains before a fully commercial in-memory-computing solution is available.

Customer receipts support a well-funded balance sheet

Customer cash receipts were A$6.1 million for the quarter, comprising licensing and non-recurring engineering fees. Weebit also received A$3.7 million from an R&D-related rebate for its French subsidiary, while R&D cash payments were A$3.7 million.

The company finished September with A$164.2 million in cash and cash equivalents, after using A$3.17 million in operating cash during the quarter. Weebit reaffirmed revenue guidance of at least A$7.1 million for the first half of calendar 2027, while cautioning that customer receipts can vary between quarters according to contractual milestones and may not align with revenue recognition.

Bottom Line?

The next decisive evidence will be whether silicon validation converts into production and whether the expanding AI evaluations become repeatable commercial programs.

Questions in the middle?

  • Can the second customer complete validation and maintain its calendar 2027 production target?
  • How quickly will onsemi, DB HiTek and Texas Instruments move from qualification or evaluation into volume designs?
  • Will the projected AI power and latency advantages translate into paid commercial deployments beyond the current studies?

Sources

1