Renesas is introducing a new microcontroller (MCU) group, called the RA8P1, that is designed for artificial intelligence (AI), machine learning (ML), and real-time analytics applications. The devices, produced at TSMC's 22 nm ulta-low leakage process, target edge and endpoint AI applications, supporting convolutional and recurrent neural networks with up to 256 MACs per cycle

The RA8P1 combines a 1GHz Arm Cortex-M85 core, a 250MHz Cortex-M33 core, and an Arm Ethos-U55 Neural Processing Unit (NPU). The MCU use a 1 Mb on-chip embedded MRAM memory.

TSMC is launching its first design center in Europe, with a focus on memory technologies for automotive applications. The company announced that it will start developing 5nm MRAM technologies, to complement its 22-nm (in production), 16-nm (ready for customers verification) and 12-nm (already in development) technologies. 

Scaling MRAM technology to 5nm will be a significant step, required, according to TSMC, to implement AI automotive chips and technologies. The company will also develop 6 nm RRAM memory in the new design center.

The MRAM industry is slowing growing, as adoption of MRAM memory solutions is on the rise, and new technologies are being developed to increase capacity and performance. Here are some recent and popular news that you may find of interest:

• Numem to start sampling MRAM chiplets by the end of 2025
• Mycronic acquires MRAM test equipment developer Hprobe
• Researchers demonstrate the world's lowest write power SOT-MRAM device
• Vertical Compute raises 20 million Euro to develop MRAM-based solutions for AI
• Researchers from UNIST design a high-efficiency graphene-based MRAM device
• Samsung details its eMRAM roadmap with 5 nm expected by 2027
• MRAM-based memory architecture could accelerate AI by a factor of 1000
• Everspin awarded a $14.55 million project to provide stable MRAM manufacturing
• Leading the Charge in MRAM: An Interview with Everspin's CEO
• Researchers from IST and Western Digital develop a promising new SOT-MRAM material
• Purdue University team to develop energy efficiency MRAM-based AI
• Kioxia and Hynix co-develop the world's smallest 1Selector-1MTJ cell
• Renesas achieves the world's fastest STT-MRAM random access speed
• IBM's next-gen FlashCore modules to feature Everspin's 1 Gbg STT-MRAM

A team of researchers at Tohoku University have achieved the world's lowest write power SOT-MRAM device, that is also offering an extremely fast write time.

The researchers focused on the tilt angle of the canted SOT device and the magnetic anisotropy of the free layer, using micro-magnetic simulation to reduce the write power. The have achieved a write power of 156 fJ in 75° canted SOT devices fabricated using a 300mm wafer process. 

Everspin Technologies reported its financial results for Q1 2025, with revenues of $13.1 million, above the company's previous guidance range, and  net loss of  of $1.2 million. At the end of the quarter, Everspin had $42.2 million in cash and equivalents, up slightly from the previous quarter.

The company did not change its guidance in 2025, saying that it expects a ramp up in revenue in the second half of the year. 

Researchers from Beihang University and Truth Memory Corporation have fabricated a 1 Kbit SOT-MRAM chip using a 180 nm complementary metal oxide semiconductor (CMOS) process, and implemented a physical unclonable function (PUF) on it. This research represents a significant step forward in the field of PUFs.

The newly developed SOT-MRAM sr-PUF achieves a strong, highly reliable, and reconfigurable PUF that can resist machine-learning attacks. The performance of the SOT-MRAM sr-PUF is remarkable - it has a challenge-response pair (CRP) capacity of 109. Its uniformity is 50.07%, diffuseness is 50%, uniqueness is 49.89%, and the bit error rate is 0%, even in a 375 K environment. These values are near-ideal, indicating excellent randomness and reliability.

Hprobe, a developer of testing equipment for magnetic devices,  has been acquired by Mycronic, a Swedish high-tech company specializing in precision production equipment for the electronics industry. 

Hprobe says that the integration represents a strategic move to enhance Hprobe’s technological and market leadership in MRAM and magnetic sensor testing, supporting the transition to high-volume manufacturing, and enhancing reliability and performance for end-users. by leveraging Mycronic’s global infrastructure, Hprobe will expand its presence across Asia, North America, and Europe while accelerating its technology and product roadmaps to meet growing demand for wafer testing of advanced MRAM and next-generation TMR sensors. 

Everspin Technologies announced new additiontions to its PERSYST EMxxLX family, the EM064LX HR and EM128LX HR, targeting automotive applications. These new chips will operate at temperatures from -40°C to +125°C, meeting the AEC-Q100 Grade 1 standard.

The new chips offer capacities of 64Mb and 128Mb, Quad SPI Interface, fast write speeds and a simple architecture. Everspin will start offering engineering samples in June 2025, with mass production scheduled by the end of the year.

Last month, during Everspin's conference call, the company updated that it has signed a contract with Purdue University to advance AI hardware based on its MRAM technology, valued at $10.5 million over four years. Now the company has officially announced this new contract. 

Everspin says that this is a strategic award to work with a consortium led by Purdue University to utilize Everspin MRAM as the underpinning in a program called CHEETA (CMOS+MRAM Hardware for Energy Efficient AI). The contract allows for phases totaling up to $10.5 million for Everspin Technologies over four years. The current phase with Everspin Technologies is valued at approximately $4 million.

Everspin Technologies reported its financial results for Q4 2024, with revenues of $13.2 million, and net income of $1.2 million. 

Reflecting on the whole year, revenues in 2024 reached $50.4 million (down 21% from 2023) and net profit  was $0.8 million. At the end of Q4 2024, the company had 42.1 million in cash and equivalents, up $3.8 million from the previous quarter.