Startup Funding: October 2023

Investors are betting heavily on data center technology, with October funding going to companies developing data processing units (DPUs) to accelerate a variety of tasks, a near-memory distributed dataflow architecture for AI, and liquid cooling technology. Much of this is linked to the build-out of the edge, closer to the source of the data than the cloud but not as compute-intensive.

Other areas drawing interest last month included fab process optimization, analog layout, a single-die reference clock, and assisted and automated driving. This report tracked 58 companies in October that collectively raised $3.2 billion. Of those companies, 24 are described below. The table at the end of this report lists all 58 companies, with links to press releases.

• Summary table
• Chips
• AI hardware
• EDA
• Manufacturing & equipment
• Packaging
• Test, measurement & inspection
• Memory & storage
• Wireless
• Security
• Displays & AR/VR
• Automotive
• Batteries & energy
• Previous reports

Chips

MangoBoost received a $55.0M Series A investment led by IMM Investment and Shinhan Venture Investment, joined by Korea Development Bank, KB Investment, IM Capital, Premier Partners, and all seed investors including DSC Investment, Stonebridge Ventures, and Must Ventures. MangoBoost develops data processing unit (DPU) hardware and software based on its DPU IP to accelerate a variety of data center functions, including virtualization, storage, AI, disaggregation, security, and networking. Funds will be used to enhance off-the-shelf FPGA-based full-stack DPUs, expand custom optimized DPU solutions, and develop DPU-powered AI. Founded in 2022, it is based in Bellevue, Washington, USA, and Seoul, South Korea.

Efabless raised $6.3M in a Series A1 extension that included GlobalFoundries, Synopsys, and New North Ventures. Efabless provides a cloud-based chip design and fabrication platform, an IP marketplace with a focus on analog and mixed signal, and a library of open-source community designs. The platform supports design and fabrication of semi-custom ASICs incorporating a user’s own design within a 10mm² user area on a carrier chip that provides the infrastructure required for building a chip including IO and power, clock, reset, and a management SoC. Designs are fabricated on multi-project wafer shuttles, packaged, and assembled on evaluation boards. Founded in 2014, it is based in Palo Alto, California, USA.

Pearl Semiconductor drew $4.5M in funding from Shorooq Partners, QBN Capital, Sawari Ventures, and individual investors. Pearl Semiconductor designs high-performance reference clocks and timing ICs. Its reference clocks and clock generators are based on piezoelectric MEMS resonators that are built on top of CMOS, which it says enables a single die reference clock solution that includes the MEMS resonator and required circuitry for oscillation, frequency multiplication, and temperature compensation. Its PLL architecture uses DSP algorithms and circuit techniques to continuously suppress any spurs while minimizing the noise contribution of active circuitry within the PLL bandwidth, which it claims gives it the merits of a Sigma-Delta Fractional-N PLL architecture while behaving like an integer-N PLL. Founded in 2020 as a spin off from Si-Ware Systems, it is based in Amsterdam, the Netherlands.

AI hardware

EdgeCortix raised $20.0M in funding led by SBI Investment and Global Hands-On VC, joined by Renesas Electronics, Cycle Group, and Monozukuri Ventures. EdgeCortix provides scalable, run-time reconfigurable neural network processor IP for edge inference. It has implemented the IP in an edge AI co-processor designed using an AI hardware and software co-exploration process. The startup also offers a compiler and software framework for developing edge AI inference applications from modeling to deployment. It targets power and mobility-sensitive devices like aerial, underwater, or ground-based vehicles, smart city, smart manufacturing, visual processing, and 5G-AI integrated systems. Funds will be used to bring its next generation hardware to market and for hiring. Founded in 2019, it is based in Tokyo, Japan.

Lemurian Labs raised $9.0M in seed funding led by Oval Park Capital, with participation from Good Growth Capital, Raptor Group, Alumni Ventures, Untapped Ventures, Plug and Play Ventures, Silicon Catalyst Angels, Blue Lake Capital, Futureland Ventures, AI Operators Fund, and Tola Capital. Lemurian Labs is developing a hardware and software accelerated computing platform tailored for AI applications. The startup is working on a Spatial Processing Unit (SPU) AI processor that uses a near-memory distributed dataflow architecture and parallel adaptive logarithm (PAL) numbering system that it claims expedites AI workloads by up to 20x greater throughput at a 1/10th of the total cost of legacy GPUs. Founded in 2018, it is based in Toronto, Ontario, Canada.

EDA

Astrus drew $2.4M in pre-seed funding led by Khosla Ventures, joined by 1517 Fund, RiSC Capital, HOF Capital, MVP Ventures, Alumni Ventures, and Plug and Play Ventures. Astrus is developing AI-powered analog layout tools. The startup says its product will take a schematic design and generate layouts automatically, with the ability for rapid iteration. Astrus is basing its technology on deep reinforcement learning and planning, similar to what was used in AlphaGo. Founded in 2022, it is based in Toronto, Ontario, Canada.

Manufacturing & equipment

Minds.ai drew $5.3M in seed funding led by Monta Vista Capital, joined by Momenta and others. Minds.ai offers a platform for optimization of semiconductor fab manufacturing processes using supervised learning, reinforcement learning, and generative AI. It aims to improve efficiencies in areas including throughput, fault detection and classification, preventive maintenance planning, reduction of wafer scraps, and prediction of wafer lot processing times, alongside capital investment planning and optimization. Founded in 2014, it is based in Santa Cruz, California, USA.

Packaging

Jiangsu Silicon Integrity Semiconductor Technology (JSSI) received CNY 600.0M (~$82.0M) in financing led by Panorama Capital and KQ Capital. JSSI provides packaging and testing services, including WLCSP, flip chip, QFN, BGA, SIP, SIP-LGA, BGA, FOWLP, 2.5D/3D, and chiplets. Founded in 2020, it is based in Nanjing, China.

Microview Intelligent Packaging Technology drew nearly CNY 100.0M (~$13.7M) in Series A+ financing from Share Capital and others. The company makes die bonder equipment. Founded in 2019, it is based in Shenzhen, China.

Test, measurement & inspection

UniSiC Technology, also known as Chenxin Technology, drew CNY 100.0M (~$13.7M) in funding from Volcanics Venture, CR Capital Management, and Summitview Capital. UniSiC develops automated test and analysis equipment for silicon carbide (SiC) and silicon power devices and modules, including high-frequency power electronics. Its products include dynamic and static characterizers, continuous power characterizers, and reliability test systems, alongside SiC MOSFET known good die and wafer level reliability test systems. It also offers a SiC CT high-voltage generator, integrated X-ray source, and MRI gradient power amplifier. Funds will be used for R&D and mass production. Founded in 2020, it is based in Shanghai, China.

Spacetek Technology raised CHF 5.8M (~$6.4M) in seed funding led by Swisscom Ventures, joined by Spectrum Moonshot Fund and existing investors. Spacetek Technology builds time-of-flight mass spectrometers for real-time analysis of process gases in semiconductor manufacturing. The startup says its instruments measure continuously and deliver calibrated mass spectra of the full mass range up to 10 times per second, enabling users to address yield issues. Founded in 2018, it is based in Gümligen, Switzerland.

Memory & storage

Floadia Corporation raised 1,050.0M yen (~$7.1M) in a Series D round from Inabata, Cypress Capital, and Japan Finance Corporation. Floadia makes embedded non-volatile memory IP cores with a focus on low-power applications such as microcontrollers, power semiconductors, and sensors. It says its flash memory was tested to retain data for ten years at 200°C. The company is also developing ultra-low power compute-in-memory AI accelerator chips using flash memory. Funds will be used for sales and development of AI chips. Founded in 2011, it is based in Tokyo, Japan.

Quinas Technology received a £0.3M (~$0.4M) grant from Innovate UK. Quinas is commercializing UltraRAM, a non-volatile memory technology that uses quantum resonant tunnelling in compound semiconductors. According to the startup, the new memory combines the non-volatility of flash with the speed, energy efficiency, and endurance of DRAM. UltraRAM is a flash-like floating-gate memory that uses atomically-thin layers of InAs/AlSb to create a triple-barrier resonant-tunnelling (TBRT) charge-confining structure. The TBRT switches between a highly-resistive (locked) state (with no bias, ‘store’), to a highly-conductive (unlocked) state on application of just 2.5 V across the gate stack (program/erase). The startup was also selected for the ChipStart UK incubator program. A spin out from Lancaster University founded in 2023, it is based in Lancaster, UK.

Wireless

HaiLa Technologies drew $10.4M in funding from strategic investor Murata Electronics and existing investors including Stanford University, Mika, Ecofuel Fund, Chrysalix Venture Capital, and TandemLaunch. HaiLa develops communication IP for radioless ultra-low-power IoT sensors. Its technology uses a backscattering technique that allows modulation of digital sensor data on top of existing ambient signals of different protocols while maintaining the integrity of the signal to the original protocol and without the need for a dedicated signal generator. Initial products use the Wi-Fi protocol. Funds will be used for R&D, hiring, and bringing its SoC to market. Founded in 2019, it is based in Montreal, Canada.

Qualinx added $10.2M to its Series A round from FORWARD.one, InnovationQuarter Capital, Waterman Ventures, and the Netherlands Enterprise Agency, bringing the round to $20.0M. Qualinx has developed a radio chip that can receive geo-positioning information from Global Navigation Satellite Systems (GNSS). The chip is based on the company’s Digital Radio Frequency (DRF) technology, which makes it possible to transfer a significant part of the chip’s analog area to the digital domain and enables CMOS scaling for GPS-like radios, thus reducing the chip’s power consumption, size, and cost. Funds will be used to complete the development of its second-generation GNSS SoC, which will combine the GNSS sensor with an IoT radio on one chip, and to commence mass production in 2024. Founded as a spin-off from Delft University of Technology in 2015, it is based in Delft, the Netherlands.

Security

zeroRISC launched with a $5.0M seed investment led by Cambridge Angels. zeroRISC is developing a commercial cloud security service for silicon built on the OpenTitan open-source silicon root of trust (RoT) project to provide cloud-based secure device management below the operating system. The startup plans to offer discrete and integrated RoTs from OpenTitan ecosystem components, a proprietary integration kit for the integrated RoT, a secure-by-default and secure-by-design embedded operating system, and a set of proprietary cloud-based services that tie into the secure OS and secure silicon designs. The first discrete open market chip from zeroRISC is hardened against physical attacks and supports post-quantum secure boot. Founded in 2023, it is based in Boston, Massachusetts, USA.

Displays & AR/VR

Xpanceo raised $40.0M in seed funding from Opportunity Ventures. The startup is developing smart contact lenses. Its prototypes include a lens for night vision, a lens that offers health monitoring functions, and a holographic lens for AR. Funding will be used to merge the three into a single lens. Founded in 2021, it is based in Dubai, United Arab Emirates.

Mojo Vision added $21.1M in closing its Series A round led New Enterprise Associates, Khosla Ventures, and Vanedge Capital, with participation from Shanda Grab Ventures, Dolby Family Ventures, Advantech Capital, Liberty Global Ventures, Fusion Fund, Open Field Capital, and Knollwood Investment Fund, bringing the round to $43.5M. Mojo Vision is developing high-performance microLED technology to create dynamic displays with up to 28,000 pixels per inch and high brightness. The technology combines high-efficiency quantum dot ink and a display system that incorporates an optimized CMOS backplane, wafer-to-wafer bonding, and custom micro-lens optics using a high-volume manufacturing process based on 300mm gallium nitride on silicon (GaN-on-Si). Applications include AR/VR, automotive, light field display, and large format displays. Previously, the company was working on a smart contact lens, and is now focused on commercializing the microLED technology behind it. Founded in 2015, it is based in Saratoga, California, USA.

Automotive

Nullmax raised CNY 780.0 (~$106.8M) in Series B financing led by Shanghai Stonehill Technology. Nullmax develops machine learning chips for L2/3 and L4 driving and parking functions in passenger vehicles. It also offers a camera-based perception system; full-stack applications and middleware for perception, planning, and control; and cloud-based vehicle data management. Founded in 2016, it is based in Shanghai, China.

Pony.ai received a $100.0M investment from Saudi Arabia’s NEOM Investment Fund. Pony.ai develops an L4 autonomous driving software and hardware system that uses a combination of lidar, radar, and cameras. It operates self-driving vehicles as a robotaxi service in several Chinese cities and is expanding to trucking and personal vehicles. The company also offers its autonomous driving domain controller and toolchain to auto manufacturers and Tier 1s. NEOM and Pony.ai will establish a joint venture to develop, manufacture, and deliver autonomous vehicles, an autonomous driving service, and smart vehicle infrastructure in the Middle East and North Africa. Based in Fremont, California, USA, and Guangzhou, China, it was founded in 2016.

Leadrive Technology received CNY 600.0M (~$82.3M) in Series D financing led by Legend Capital and Oriza FOFs. Leadrive makes IGBT power modules and main-drive dual-motor controller products for electric vehicles, including heavy-duty vehicles. Funds will be used to expand production capacity and develop next-generation power modules, power bricks, and silicon carbide technology. Based in Shenzhen, China, it was founded in 2017.

Vayu Robotics drew $12.7M in seed funding led by Khosla Ventures, with participation from Lockheed Martin Ventures, ReMY Investors, and others. Vayu Robotics develops AI models for mobility and low-cost bioinspired sensor systems to replace lidar in mid-range applications. It targets last-mile delivery, factory automation, and automotive. Founded in 2021, it is based in Berkeley, California, USA.

Batteries & energy

Skeleton Technologies drew €108.0M (~$113.7M) in Series E investment from Siemens Financial Services, Marubeni Corporation, and others. The company develops supercapacitors using ‘curved graphene,’ a proprietary material it says provides increased energy density and electrical conductivity compared to activated carbon. It also builds what it calls superbatteries, which combine the fast charging of supercapacitors with the high energy and long duration of batteries for high-power applications such as grid energy storage and heavy-duty equipment and vehicles. Funds will be used for development of next-generation products and manufacturing expansion. Founded in 2009, it is based in Tallinn, Estonia.

JetCool drew $17.0M in a Series A round led by Bosch Ventures, with participation from In-Q-Tel, Raptor Group, and Schooner Capital. JetCool develops a microconvective liquid cooling technology that uses arrays of fluid jets to cool high-power electronics. Its offerings range from fully sealed cold plates and direct liquid-to-chip products to hotspot-targeted embedded liquid cooling and complete turnkey cooling systems. The startup claims its liquid-to-chip solution provides a heat transfer coefficient 10x greater than competing approaches and removes the need for thermal pastes and interface materials, while its on-chip cooling solution embeds directly into the chip substrate to circulate coolant and target the die’s hotspots, providing TDP cooling of more than 1,000W. Founded in 2019 as a spin off from MIT, it is based in Littleton, Massachusetts, USA.

Table

Footnote on name inconsistencies

Some Chinese companies inconsistently represent their names in translation or appear to have multiple names, some of which may be artifacts of an Internet translation engine’s best guess in translating the companies’ names from Chinese characters into the Latin alphabet for Western consumption. Semiconductor Engineering cannot independently verify a company’s preferred name in the Latin alphabet, but a logotype on a company building or website is likely the correct company name in English.

For example, Jiangsu Silicon Integrity Semiconductor Technology is the name that appears on the company’s building signage but the name Xinde Semiconductor appears in Google translations of press releases.

Ingin Smart Closure is the official Western name used on Enjing Technology’s logotype but some websites, including Crunchbase, say the company is called Enjing Technology.

Jiamai Co., Ltd. says it uses the brand name of MAXAIR on its products, so MAXAIR seems to be an alternate name for the company and the MAXAIR logotype appears on the company’s headquarters.