Back in 2017, we had seven companies in the review. Not to say there were just seven startups on the market, but there were just seven that disclosed enough about their approach. One of the startups decided not to continue with the development, and another two have changed their focus. But the rest four of them are still here and show a good pace! Just check the list of the five most funded automotive radar startups up to date (in USD):
- Vayyar, 188M raised, first mentioned 2017
- Uhnder, 145M raised, first mentioned 2018
- Oculii, 76M raised, first mentioned 2017
- Echodyne, 64M raised, first mentioned 2017
- Arbe Robotics, 55M raised, first mentioned 2017
Now there are 21 companies on the list. And it is three times more than we had four years ago. As ADAS and self-driving topic is on the rise, funding follows. Since the beginning of 2020, more than 156M has been raised by the startups mentioned here. And it happened despite some cooldown caused by pandemic restrictions.
As startups grow, their business develops as well as their technologies. This year, for the first time, we have estimates for the valuation of certain companies for November 2021 (in USD).
- Arbe Robotics, est. valuation 723M, source: IPO prospect
- Oculii, est. valuation 307M, source: M&A prospect
- Gapwaves, est. valuation 138M, source: Public company
- Neteera, est. valuation 40M, source: Valuation in a private financing round
This amount of data, of course, gives just a rough estimate of how much radar startups can be valued at. But as the automotive radar market grows year to year, and four out of five most funded startups are not yet valued, we could name the first automotive radar unicorn in 2022.
With so many companies, what are the approaches being taken by each? There are three major ones:
Proprietary hardware: SoC or separated analog and processing parts. RF part of an imaging radar provides an enormous data stream to the processor, so the startups need to develop a processing part and an analog one.
Hardware-agnostic: It implies that the benefits are gained by some sort of post-processing. And it does not matter which hardware did the signal acquisition and pre-processing. Imagine a radar sending certain frames every second; each contains a point cloud or a list of objects. The issue is that these points or objects will vary from frame to frame. Their range and angular coordinates will also differ. Analyzing a sequence of frames, a processor can approximate the points or objects to increase the accuracy of the measurements. However, there are serious obstacles to this approach. The processing itself takes time and lengthens the entire data processing pipeline. Due to the complexity of post-processing, it also requires more processing power and leads to a higher unit cost.
Complementary components: Such as new antenna configurations and materials. These companies are capable of developing the entire radar but focused on a component in the system.
A couple of years ago, one of the known RF manufacturers, Texas Instruments, released a new family of its AWR transceivers. In early 2021 the lineup was updated. Technically it is just another 77 GHz MIMO solution. But unlike the rest of the industry, Texas Instruments simplified the process of acquiring these RF and SoC components. The comprehensive explanations, including whitepapers, datasheets, CAD files, are available on the website without any restrictions. The development boards are available for just USD 150. The result is that at least five startups mentioned below use TI RF frontends or SoC for the developments. Will the industry follow TI’s approach, thus giving a massive boost to automotive radar startups? We will check it next year.
USA, Laurence KS / China, Shanghai, 2M raised, founded 2015.
Being a startup, Ainstein implements startups’ core value: speed. In the past couple of years, they have changed their hardware and software approaches, which led to the availability of FCC-certified products (not a common thing for a startup). They also established new partnerships with other automotive market players. Ainstein uses existing components from Texas Instruments for typical radar applications: automotive, UAV, and smart city. It seems that Ainstein makes no significant changes to the hardware: the specs and features list resembles other solutions. At the beginning of 2021, Ainstein confirmed that Doosan Bobcat, an equipment manufacturer, became a strategic investor. The company exposed no further details.
Ainstein formed RADAC, a joint venture by Ainstein and ADAC electronics focused on bringing radar-based solutions to the automotive market. It may be an exciting idea for startups, where radar startup carries its technical competencies, and the proven supplier empowers them with its market reputation.
Curiously, there is a Chinese company named Muniu Technology with almost the same automotive, UAV, and smart city lineup and approach to radars as Ainstein.
Israel, Tel Aviv, founded 2015, 55M raised
The biggest news regarding Arbe Robotics is about their financials, not technology. Earlier this year, Arbe Robotics claimed that it would go public via the SPAC merger. This popular way of going public for technological companies gives Arbe a projected 723M dollars valuation.
Arbe Robotics has also told us some numbers. The company’s chipset processes 2304 (48x48) channels — the patented processor processes up to 30 Gbps of raw radar data. The estimated parameters of the radar are 300-meter range with 7.5- to 60-centimeter resolution, 100-degree azimuth with a 1-degree resolution, 30-degree elevation with a 2-degree resolution, and Doppler resolution of 0.1 meters per second. There are undisclosed partnerships already, and the company plans to provide a reference design for Tier-1 and OEM customers for radar system development.
Israel, Holon, founded 2013, 4M raised
The company is developing a 360-degree radar with no moving parts. The primary application is counter-drone defense. No news regarding automotive cases was shared; nothing is disclosed about rounds they raised as well.
Seoul, South Korea, founded 2018, 7,3M raised
As some of the radar companies do, Bitsensing develops a portfolio of products where automotive radar plays a significant role. The company’s approach to high-resolution and imaging cases may imply the use of cascaded transceiver setups by Texas Instruments, Infineon, and/or other RF suppliers. Their claimed goal is to develop an imaging radar with a horizontal and vertical resolution of 1 degree.
Preliminary specs of a sensor named Bitsensing MXR710 can be found on 3rd party websites. They say that radar provides two measurement modes within a single package, short- and long-range up to 60 and 150 meters, respectively. The sensor does look like imaging of any type and is not mentioned on the official website.
USA, Kirkland WA, founded 2014, 64M raised
Having a strong market position in security radar applications, the Bill Gates-backed company is still paving its way to the automotive market. Echodyne’s radars have an outstanding performance compared to conventional automotive radars, but their unit should be costly and massive as well. This is why Echodyne claims it focuses on L4+ cases, where the price is not the limiting factor.
Sweden, Gothenburg, founded 2011, 7.5M raised
One of the critical components of radar is an antenna. Gapwaves is working on its unique waveguide technology that allows boosting the parameters of most patch antennas. Patch antennas are widely used because of their versatility and robustness. Gapwaves claims that their antennas have 3 dB less feeding loss, which leads to doubling the radar range when it comes to energy. Such an approach to antennas is also flexible. There can be a family of antennas for various applications.
The benefits are provided by a more sophisticated construction of an antenna. While a patch antenna is a two-dimensional print on the outer level of a PCB, Gapwaves antenna is built with a 3D structure using metalized plastic.
Automotive is just one of the many applications for this technology with noticeable developmental results; 5G is another good example. Gapwaves has announced its collaboration in this field with Uhnder, another prominent radar startup. Gapwaves technology will be used in the future Uhnder’s high-resolution radar on-chip. Recently, Gapwaves has disclosed that the company strategically partnered with Hella in the field of antenna design. Hella plans to invest about 18M euros in the startup. There is also an order for antenna design and initial series production from another Tier-1 supplier.
Notably, Gapwaves is a public company that has a market cap of 138 million euros. It discloses its financials and reports, which are extremely useful for understanding the business.
USA, Columbus OH, founded 2016, 2.4M raised
Ghostwave approaches the problem of mutual radar interference by using a random sequence of emitting signals. Similar approaches are often called pseudo-random noise radars. The company works on the key applications of its technology, including UAVs, but does not disclose much about it.
GPR, formerly Wavesense
USA, Somerville MA, founded 2017, 23M raised
An MIT spinoff works on GPR, a ground-penetrating radar, which is named GPR™, ground positioning radar™. Such radar scans the ground under the vehicle for a unique imprint of a particular position. These imprints are used for precise mapping and localization of the vehicle. As claimed, it works in garages and street environments and complements other mapping solutions such as GPS.
Wavesense did not mention any pilot projects started, but still, there is good news. The company has raised a 15M follow-on round from Rhapsody venture partners, Cambridge MA venture fund that invests in deep tech scientific projects.
USA, Tucson AZ, founded 2017, 12.5M raised
Lunewave is developing a 360-degree antenna based on the Luneburg lens concept that is claimed to increase the angular resolution of the radar. Lunewave is a successful example of a collaboration between a startup and a corporation. The company has participated in two programs held by BMW: Urban-X startup accelerator and BMW Startup Garage. Lunewave has raised several rounds of funding, including the latest from FM Capital, which focuses on mobility startups of various kinds.
It is said that Lunewave is now evaluating tests of their sample antennas. The initial phase is finished already, next to be performed next year.
USA, Palo Alto CA, founded 2017, 17M raised
As Metawave develops, it explains more about its focus, and this video is a good example of an explanation. First, Metawave creates a special Marconi 77 GHz phase controller, a first-of-a-kind analog beam steering unit that improves angular resolution up to a 0.1-degree level. Second, Metawave works on AWARE, a powerful detection and classification platform that makes radar “5D”, where additional D stands for object classification. Both technologies are used in SPEKTRA radar, which is available as a demo unit.
Except for automotive applications, Metawave works in the communication field, 5G antennas in particular. CorporateVision awarded Metawave Klone&Turbo platform as USA Best Broadband Connectivity Deployment.
The company was awarded a contract from the United States Air Force (USAF) with a subject of radars for electric Vertical Take-Off and Landing (eVTOL) aircraft. Metawave has also raised an undisclosed Round A led by Denso in 2020.
Israel, Jerusalem, founded 2014, 16.7M raised
Having healthcare-related products in its portfolio, Neteera has been working on 120 GHz sensors for the last two years. It led to successful pilot projects in hospitals. But now, the company doesn’t even show automotive applications on its website. The latest automotive news came from early 2020 when Neteera announced a strategic collaboration with Valeo at the Consumer Electronics Show in Las Vegas. Even if Neteera has any news regarding in-cabin applications of its sensors, they are not exposed.
CNBC mentioned Neteera among Disruptor 50 companies back in 2020 with a valuation of USD 40M.
USA, Beavercreek OH, founded 2013, 76.1M raised
This year Oculii has told us more about its approach to imaging radars. It is chip-agnostic software, named Virtual Aperture Imaging Software Technology, that processes acquired data with SLAM-style mechanics. The goal is to compare multiple radar frames to build a more detailed and accurate picture of the environment than any single frame has. Oculii’s software is also using ML/DL techniques to adjust transmitting signal parameters (the chirp) to fit a particular road environment. However, the ability itself is what almost every RFIC on the market has, and the only question is how you find the best parameters for it.
In 2019 Oculii disclosed joint co-development with Infineon of a family of radars starting from single-chip radars for Level 1 ADAS to multi-chip radars for Level 4 Autonomous Driving. Later in 2020, a strategic partnership with Hella was disclosed. Those announcements were followed by a 55M Series B investment round in May 2021. That money makes Oculii one of the Top-3 most funded radars startups — with corresponding expectations, of course. GM’s venture arm was named among Oculii investors.
Finally, it was claimed that Oculii had signed a merger agreement with Ambarella, an AI vision company, with a valuation of 307M.
Israel, Tel Aviv, founded 2017, 3M raised
Radsee is a stealth startup from Israel that is making chip- and DSP-agnostic imaging radars with impressive parameters. 400 m range with a broad 120-degree view and estimation of both angles are claimed. These specs are achieved on off-the-shelf 77 GHz components with Radsee’s patented antenna and unique system architecture.
Radsee has closed a financing round from another radar developer, RATA, which is focused on tactical and avionics radars.
Smart Radar System
South Korea, Bundang-gu, founded 2015, 4.0M raised
A Korean startup shows another example of an imaging radar. Most of the radar setups are based on the new TI AWR lineup: it is directly mentioned for some radars. It may be guessed by analyzing the specs of the others. Single, as well as cascaded setups, are offered. And at least one of the radars, the most powerful one, is claimed to be using 8 AWR transceivers at a time. It is curious how effective such scaling is when it comes to angular resolution. But it should be eight times more expensive than a single transceiver radar for the RF part.
Smart Radar System has disclosed its joint development agreement with known Korean automotive parts supplier Mando in radars this year.
Staal Technologies (Omniradar)
Netherlands, Eindhoven, founded 2017, 2.6M raised
Staal Technologies shared a new video and some specs of their proprietary radar on-chip solution. A 60 GHz RFIC works up to 5 meters with a decent resolution, making it capable of in-cabin applications. No partnerships have been announced yet.
India, Bangalore, founded 2016, 2.1M+ raised
Steradian Semi develops imaging radars with its partner and RF supplier, IDT/Renesas. There are two products in the company’s portfolio: E-band imaging radar and high-level visualization software. As mentioned, radar has 256 virtual MIMO channels for 0.1–0.3 degree resolution, ethernet output, and 15 Hz refresh rate. Two rounds of financing with undisclosed amounts of money happened in 2020.
USA, Austin TX, founded 2015, 145M raised
Being an excellent example of a radar startup, Uhnder systematically disclosed some information about the company’s approach in articles and whitepapers. But the biggest news was not about tech. Series B and C doubled the overall amount of raised money to more than 145M, close to the amount that another company, Vayyar, raised. The company is working on bringing their DCM technology to market what causes huge investments upfront in chip development.
DCM is the technology that is solely developed by Uhnder. It uses phase to modulate the TX signal rather than the frequency. Then a digital code is added to the phase to mark the particular transmission. It also uses MIMO and additional convenient techniques, but the most novelties lie in the DCM technology and the RFIC that implements it. Uhnder did not disclose any test results, but, hopefully, they should be available as soon as this year, since 2021 is named the year when first radars are ready.
Israel, Yehud, founded 2011, 188M raised
Once started from Walabot, a microwave device that finds studs in the walls, the company now offers various solutions based on their proprietary RFIC. Automotive in-cabin sensors for seat belt reminders and child presence detection are among them. Valeo was named as a Tier-1 partner for bringing those applications to market.
In early 2021 the company disclosed that it is launching the new Vayyar XRR — an AEC-Q100 qualified and ASIL-B compliant RFIC covering most ADAS applications. A 48-channel MIMO transceiver helps to reach excellent angular resolution and accuracy in the 0–300 m range.
Thus, it becomes clearer why Vayyar raised an enormous round of $109M led by Koch Disruptive Technologies back in late 2019: chip development and manufacturing have never been cheap.
Canada, Toronto, founded 2016, 3,6M raised
The company is focused on in-cabin radar applications: driver conditions monitoring and occupancy sensors. UWB radars are used for that purpose. The approach allows putting maximum functionality in a single in-cabin sensor. For example, it can help to substitute sensor mats inside the seats, which are now mandatory for any passenger vehicle. The system can even be set up as an aftermarket option.
Xandar Kardian often gets honored at exhibitions like CES and has a vast portfolio of smart city and healthcare products.
USA, San Jose CA, founded 2019, 5,6M raised
Founded by the former radar lead of Lyft’s Level 5 Autonomous Division and his colleagues, Zadar has developed a Software Defined 4D Imaging Radar (SDIR) platform as the basis on which to apply their advanced AI and ML processing algorithms to make radar an intelligent, context-aware sensor.
The company takes a comprehensive approach to radar, meaning they focus on building the entire end-to-end processing pipeline. They use existing RF and silicon technologies, most notably Xilinx has been mentioned as an anchor partner for the processing hardware. The team at Zadar sees a trend in radar hardware commoditization and the need to elevate the function and performance into the software layers. They don’t consider their SDIR platform and the accompanying processing techniques hardware dependent, however, they have developed a family of radars to support the early adoption of their technology.
Zadar develops a family of solutions that complement each other. The sensor platforms are zPROX for short-range, zSIGNAL for mid-range, and zPRIME for ultra-long-range and ultra-high-resolution applications. zVUE is the software suite for data acquisition and processing using machine learning techniques for advanced algorithms including clustering, tracking, and classification. This software suite also allows for dynamic reconfiguration of the radar performance and data output.
It is claimed that these tools may be used together or separately depending on the potential customer’s approach to implementing a radar solution. Zadar most recently was selected to be part of the latest Plug’n Play Mobility Accelerator Batch.
USA, Berkeley CA, founded 2017, 21M raised
Zendar develops a novel radar system that fuses raw information from simple, low-cost radar front-ends distributed around the vehicle to create a lidar-like high-resolution view of the environment. Zendar illustrates a good example of transparency by sharing demos and datasets. As explained in papers and videos, Zendar demonstrates an automotive-grade synthetic aperture radar and semi-supervised machine learning algorithms to improve radar performance. This type of approach is known to work in the area of aerial surveillance with costly radar setups. Zendar may be the first company to put such functionality in street-level hardware within automotive price limitations.
Zendar plans to come out of stealth mode later this year and share more information about their product.