Mark Harris | Vox

How failed OS startup Cyanogen pivoted into self-driving construction equipment

2024-12-02T16:59:39-05:00

Feed me. | Photo by Dianne Manson/Getty Images

This post has been updated.

In 2015, when there was still an ounce of hope to compete with Google and Apple in the smartphone operating system world, Cyanogen Inc. was producing an open-source operating system for Android devices. Its software was running on 50 million phones and was on track to hit half a billion handsets by 2020, the company said.

Then it all went wrong. A series of what one of its co-founders called “bad” business deals, and disagreements about the future of its Cyanogen Mod operating system, resulted in the company firing most of its staff and closing its Seattle office at the end of last year.

Now, as Axios first reported yesterday, Cyanogen is back, calling itself Cyngn and promising “innovative solutions for autonomous machines and vehicles.” The company is still in stealth mode, but Recode has pieced together some of the story of its renaissance from documents, websites and interviews.

At its peak, Cyanogen Inc. employed around 150 staff. All but about 15 were let go late in 2016, when co-founders Kirt McMaster and Stefanie Jane also left the company. Cyngn’s CEO is now Lior Tal, who had been COO until McMaster’s departure.

McMaster went on to found his own mobility company, Nere — also in stealth — while Jane is an engineer at Oculus, Facebook’s VR company. Neither would comment on Cyngn’s pivot to autonomy.

That pivot occurred because executives and board members at Cyngn see automated vehicles as being as disruptive in the next five years as smartphones have been over the last decade. Cyngn is now back up to around 30 staff, including autonomous automotive engineers from Mercedes Benz and Udacity, according to LinkedIn, and is still on a hiring spree.

In late September, Cyngn received a permit to test two autonomous vehicles in California. Those vehicles are street-legal electric golf carts. However, unlike many of the other transportation startups testing in California, Cyngn does not intend to build a self-driving passenger vehicle.

A document prepared for a South Korean trade delegation to Silicon Valley in May reveals that Cyngn is planning to convert loaders, excavators and other construction vehicles to become fully autonomous. The presentation said that Cyngn had already developed and integrated a beta version of an autonomous system based on computer vision and deep learning, and that it would be commercializing its technology by the end of August. However, Cyngn now says that this document no longer reflects the current direction of the company.

Job listings on Glassdoor show that Cyngn is working with radar, camera, ultrasonic, and lidar sensors, and using ROS, an open-source operating system for robots. But Cyngn will not be following Cyanogen in offering its software as a free or open-source platform, nor using any of Cyanogen’s intellectual property, according to a source close to the company.

Over its three year existence and funding to Series C, Cyanogen Inc raised a total of $115 million from 15 investors, including Andreessen Horowitz, Benchmark Capital, Qualcomm Ventures, Foxconn and Tencent. The Korean document says that Cyngn now wants to raise a further $200 million in a Series D round, and that the company is also open to being purchased outright.

Update: This post has been updated to note that the document prepared for a Korean trade delegation no longer reflects the company’s current direction.

Editor’s note, December 2, 2024: Stefanie Jane’s name has been updated in this article.

This article originally appeared on Recode.net.

Electric car startup Lucid Motors is short of cash and weighing its options

2017-07-20T10:36:54-04:00

The Lucid Air | Lucid

Build or sell? That’s the choice faced by electric vehicle startup Lucid Motors, which is considering several acquisition offers even as it raises money to build its own assembly plant for its Tesla Model S rival, the Lucid Air.

Contrary to some reports, Ford is still very much interested in acquiring the Silicon Valley-based startup, according to one person very close to the negotiations. Jim Hackett, Ford’s new CEO, visited the company with a number of executives on June 28. Ford subsequently presented Lucid with a draft contract called a term sheet. Even if Lucid agreed to sell, there would be a period of research and due diligence before the acquisition closed.

However, Recode has been told that at least two other unnamed companies are also interested in acquiring the startup, and that the Series D funding round is proceeding in parallel. One source describes the fundraising as “a work in process,” with Lucid sitting on financing offers of more than $200 million while it considers selling the company outright.

Lucid has had a turbulent history. Founded as Atieva in 2007 by an ex-vice president of Tesla, the company initially focused on batteries and drivetrains for electric cars and buses, building up a valuable store of patents. Early in 2014, Beijing Automotive Industry Holding, one of China’s leading state-owned carmakers, and LeEco, a Chinese internet company, invested $100 million so that Atieva could develop its own vehicle.

BAIC was meant to help build and market a sedan to rival the Tesla Model S in China. However, BAIC’s involvement did not last long. “They wanted to be in control of the company,” said a source close to Lucid. “That was very disruptive in terms of day-to-day operations.” In April 2016, BAIC sold its shares, which ended up in the hands of LeEco’s owner, billionaire Jia Yueting, two sources confirmed to Recode.

As it turned out, Jia was just as disruptive as BAIC. In the summer of 2014, he had quietly formed his own EV company, Faraday Future, to develop and build yet another rival high-end electric car. Faraday Future has since suffered a series of high-profile mishaps, including a failed demonstration of autonomous technology at CES and the suspension of a $1 billion factory construction project in Nevada.

In May, Jia resigned from LeEco’s parent company amid concerns over its liquidity and his personal finances. Earlier this month, it was reported that Jia said he had sold his stake in Lucid Motors for several hundred million dollars. Lucid would be delighted to see the back of Jia and any lingering association with the troubled Faraday Future. However, a source close to Lucid says that reports of Jia selling his shares are probably untrue.

Lucid Motors did not immediately respond to requests for comment.

Meanwhile, development of the $60,000 Lucid Air has been proceeding well, and it was recently shown reaching 235 mph on a test track. However, Lucid has had its own setbacks, including abandoning attempts to design its own self-driving technology in favor of using a semi-autonomous system from Mobileye, now part of Intel.

Lucid plans to build an assembly plant in China, which would be financed entirely by local and regional government incentives. The site of its proposed U.S. factory, Casa Grande, Ariz., is being somewhat less generous. Lucid has a $1.5 million job-training grant from Arizona and is negotiating a further $5 million grant, according to a spokesperson for the state’s commerce authority. It could also apply for a tax credit program worth up to another $40 million.

However, the make-or-break loan for the $700 million facility would be a U.S. Department of Energy program to help build advanced-technology car plants. Tesla benefited from a $465 million DOE loan in 2010, while Ford secured one for $5.9 billion. But because the loan reimburses expenses rather than paying them up front, Lucid cannot move forward until it has financing in place.

Lucid began its fourth round of institutional fundraising earlier this year, with the help of investment bank Morgan Stanley. Ford’s acquisition offer came as a surprise, said a source close to Lucid: “The bank approached Ford to see if they wanted to participate in the Series D. They came back and said they wanted to acquire.”

Global carmakers usually shy away from investments where they would be minority shareholders. Buying Lucid outright would let Ford freely use its core technologies in powertrains and batteries, and secure a valuable hoard of patents for the future.

It could also mean the Casa Grande factory never gets built, thinks Joe Paluska, an automotive industry consultant. “My personal point of view is that Lucid is unable to come to market as Lucid, due to the significant financing and operational hurdles that face any Tesla challenger,” he said. “A more likely scenario is that they are acquired by a more traditional carmaker who may need a premium EV play to compete.”

Mark Harris is an investigative technology reporter based in Seattle. In 2014, he was Knight Science Journalism Fellow at MIT, and in 2015 he won the AAS Kavli Science Journalism Gold Award. Reach him @meharris.

This article originally appeared on Recode.net.

Atieva will launch its Tesla competitor by December

2016-10-20T11:28:48-04:00

Atieva will unveil its Tesla rival to select people tomorrow, ahead of a full launch in December, according to two sources close to the company. The vehicle’s look has evolved considerably since early designs of the car, revealed last week by Recode, these sources say.

The soft launch will showcase Atieva’s final design and details of its production plans to hit the company’s ambitious goal of selling cars by the end of 2018. The news comes in a busy week for electric vehicles, which included updates on Tesla’s Model 3 and a new electric vehicle from Chinese tech company LeEco.

The event at Atieva’s Menlo Park office could involve rebranding the company as Lucid Motors, according to one of the sources.

Atieva has also chosen a site in Arizona for its U.S. assembly plant to produce the vehicle, code-named Atvus, these people say. In June, Atieva’s manufacturing director told Reuters the plant would build 20,000 cars a year, eventually rising to 130,000 annually.

One source says that the factory will be built in Casa Grande, which happens to be one of the locations Elon Musk scouted for his own $5 billion Gigafactory to produce battery packs for Tesla vehicles. That facility ended up in Nevada, and is expected to start producing batteries by the end of this year.

“Arizona is not going to want to lose out a second time,” said a senior Atieva engineer who has since left the company. “Apparently, they’re throwing everything including the kitchen sink at Atieva to come to the state.” That likely includes generous tax incentives.

Another motivation for Atieva’s choice might be acting CEO Peter Rawlinson’s competitive streak with Tesla Motors, where he was previously chief engineer for the Model S. Atieva’s high-end sedan is widely known in Silicon Valley as the “vengeance car.”

According to the former engineer, Atieva’s single prototype is currently in Germany for “beautification”; the company will then proceed to make five alpha cars — early development vehicles intended for track and road testing. This is down from an estimated 30 to 35 alpha cars because, one source suggests, the company is struggling for money.

Paying for the assembly plant, costing upward of $300 million, will likely involve raising another round of funding, according to one source. Atieva will apply for the Department of Energy’s Advanced Technology Vehicles Manufacturing program that can cover 80 percent of the cost of building a factory, although this is given as a reimbursement rather than an up-front loan.

One small revenue stream Atieva could garner before its first production car rolls out of the factory could be Formula E, the sister racing series to Formula One that uses only electric vehicles. The series will include Atieva’s technology for the 2017-18 season, according to the sources. Fellow EV startups NextEV and Faraday Future already have teams in the championship.

Questions, however, remain about Atieva’s ownership and future. Its board and management remain conflicted about the direction of the company. Chinese technology company LeEco has a stake in Atieva, but also controls rival EV startup Faraday Future.

LeEco’s founder Jia Yueting announced today that Faraday Future would be unveiling its first production car at the Consumer Electronics Show in January, and that LeEco itself is moving forward with a semi-autonomous electric car called the LeSee Pro.

Neither Atieva, Formula E nor officials from Arizona and Casa Grande replied to requests for a comment on this article.

This article originally appeared on Recode.net.

Tesla rival Atieva’s first electric car looks a lot like a Model S

2016-10-13T12:59:56-04:00

The Atvus should look familiar. | Atieva via public records request

Atieva, the troubled Silicon Valley startup that began by making batteries for Chinese buses, is expected to unveil a rival to Tesla’s Model S in early December, Recode has learned. Using public records requests, Recode has also obtained exclusive images of the car’s design and initial manufacturing facility: An electric bike company in rural China.

The documents, which were filed with the state of California late last year, say that Atieva’s first vehicle will be called Atvus. The design render shows a conventionally streamlined, four-door sedan much like a Model S.

Over the summer, Atieva has been raising its profile with a series of videos showing a van equipped with an electric battery and drivetrain outperforming a variety of high-end cars, including a Ferrari, a Dodge Viper and even a Model S. The van, called Edna, has about the same range and acceleration as a mid-range Tesla, including a 2.74 second 0-60mph dash.

In September, Zhang Xin, Atieva’s deputy general manager in China, told Chinese reporters that the car will have fully independent front and rear electric motors, and will be able to accelerate from zero to 100km/h (about 62mph) in under three seconds.

According to an executive who has since left the company, Atieva’s vehicle will also come with “some elements of autonomous drive baked in.” The firm’s co-founder, Sam Weng, said, “We are creating an experience that is personalized and liberating, erasing [the] constraints of modern transportation.” That suggests a semi-autonomous highway driving system, along the lines of Tesla’s Autopilot.

Atieva has two facilities in Silicon Valley, one just 10 miles from Tesla’s own factory. Both have laboratories and workshops for vehicle development, although the company has yet to announce the location of its full-scale factory. In an interview with Reuters in June, Atieva’s manufacturing director said he wanted a U.S.-based assembly plant that would build 20,000 cars a year at first, eventually rising to 130,000 annually.

Until then, the company seems to be relying on a Chinese electric bike company for manufacturing. When asked by California officials to provide a photo of its manufacturing facility, Atieva submitted a picture of the Dayang Motorcycle company in Luoyang. Dayang makes a range of electric scooters and motorbikes, as well as a tiny two-seater micro-car called Chok with a top speed of 49mph. The Chok is so small it is even allowed in bike lanes in China. Atieva already has some presence in Asia, with a Taiwanese research hub and a facility in Shanghai.

Atieva was formed in 2007 by Weng and two co-founders, Bernard Tse and Sheaupyng Lin. The company initially worked to develop batteries and electric drivetrains, filing more than 100 patents and delivering battery packs for electric buses in China. That all changed in 2014, when two Chinese firms invested around $100 million for a nearly 50 percent stake in the firm.

One was BAIC, a Chinese state-owned carmaker, the other LeEco (then called LeTV), a technology company making televisions, smartphones and electric bikes. Atieva went on a hiring spree and pivoted to developing a complete, autonomous electric car to compete with Elon Musk’s Teslas.

Vehicle development progressed well through 2015, according to former executives, but there was growing friction among investors. Jia Yueting, the billionaire owner of LeEco, had quietly started his own electric vehicle startup, Faraday Future, in 2014, without informing the Atieva board. Then LeEco announced its own plans to develop a luxury, connected EV called the LeSEE.

Meanwhile, BAIC was attempting to get access to Atieva’s battery and drivetrain technology, and wanted to shift the company to focus on Chinese buyers. In November 2015, co-founder Bernard Tse was abruptly ousted as CEO. Then in early 2016, according to former executives, BAIC sold its stake in the company. According to several sources close to Atieva, the buyer of BAIC’s shares was Jia Yueting.

Neither Atieva nor LeEco replied to requests for a comment on this article, although Faraday Future said that it has no direct relationship to Atieva.

While the company still officially lacks a CEO, Atieva’s Chief Technical Officer Peter Rawlinson, a former Tesla VP, is understood to be running the company day to day. Atieva’s current marketing strategy seems designed to keep both its U.S. staff and Chinese investors happy.

“In 2018, Atieva new models will officially go on sale first in the United States, and then extended to the Chinese market,” said Zhang in September. “Although many parts … are still in the prototype stage, we will strive to complete the vehicle off the assembly line in two years’ time.”

It is difficult to overstate just how ambitious that goal is, given that Atieva has never produced a complete vehicle, does not currently have an assembly plant — or even a site for one — and seems to be relying on an obscure e-bike company for manufacturing assistance.

If the Atvus does emerge in 2018, it will certainly have more competition than just the Model S — if only from its sister companies. LeEco is having an official U.S. launch next week that may include details of its LeSEE luxury electric car, and earlier this year Faraday Future broke ground on a billion dollar factory in Nevada that it hopes will also be building cars within two years.

The ownership and organizational relationship of Atieva, Faraday Future and LeEco remains extremely murky. However, if Jia is serious about building an automotive empire, Atieva’s hoard of patents could be invaluable to his other companies. LeEco and Faraday Future’s vehicles will certainly share some core technologies, particularly around connectivity and autonomous drive.

The scale of Jia’s manufacturing vision is breathtaking. Along with Atieva’s 130,000 cars, Faraday Future hopes its Nevada facility will produce 140,000 annually, and a $2 billion LeEco assembly plant in China 400,000 more. Together, that would dwarf even Elon Musk’s promise that his California factory will produce 500,000 mass-market Model 3 electric cars each year, as soon as 2018.

Whether Atieva ends up as a stablemate or a competitor to LeEco and Faraday Future, one thing seems certain: Tesla Motors will be watching December’s launch of its first Chinese rival very closely indeed.

This article originally appeared on Recode.net.

Inside Alphabet’s money-spinning, terrorist-foiling, gigabit Wi-Fi kiosks

2016-07-01T07:00:08-04:00

A LinkNYC kiosk on Third Avenue is part of New York City’s free superfast broadband experiment. | Timothy A. Clary / Getty

The free Wi-Fi kiosks that Alphabet’s urban innovation division Sidewalk Labs is selling — similar to those already on the streets of New York — will come with eyes, ears and a host of environmental, air and digital sensors to give the tech giant an unprecedented snapshot of urban life, according to documents obtained by Recode.

The documents, which formed part of Sidewalk Labs’ pitch to cities participating in the U.S. Department of Transportation’s Smart City Challenge, show that Alphabet — Google’s parent company — wants to monitor pedestrian, bike and car traffic, track passing wireless devices, listen to street noise and use the kiosks’ built-in video cameras to identify abandoned packages. Each kiosk will also generate an estimated $30,000 a year for the company from digital advertising.

“The Kiosk sensor platform will help address complex issues where real-time ground truth is needed: Understanding and measuring traffic congestion, identifying dangerous situations like gas leaks, monitoring air quality, and identifying quality of life issues like idling trucks,” says one promotional flyer. Recode obtained the documents under public records laws.

Sidewalk Labs is led by Daniel Doctoroff, who was New York City’s deputy mayor for economic development and rebuilding under Michael Bloomberg, and was also the former CEO of Bloomberg LP. The firm describes itself as a “new type of company that works with cities to build products addressing big urban problems” and one that’s “building a platform” to “accelerate innovation” in cities.

Around two hundred kiosks, without sensor suites, have already been installed in New York City by LinkNYC. This is a consortium that includes Intersection, a high-tech advertising consultancy owned by yet another consortium, led by Sidewalk Labs.

Each 9.5-foot-high Link kiosk offers a range of free services, including U.S. phone calls, gigabit Wi-Fi access, USB charging ports and transport directions. The kiosks have a keypad, a touchscreen tablet for browsing, a dedicated 911 button for emergencies and are ADA compliant. Sidewalk’s documents quote market research that found 90 percent of New Yorkers believe the kiosks are a “positive initiative for the city.”

But this is only the start of what Sidewalk Labs has planned. Sidewalk offered the Columbus, Ohio, winner of the Smart City Challenge up to 100 kiosks in four of its neighborhoods. “Each Kiosk,” promises Sidewalk, “includes data analytics [that would allow Columbus to] better understand the urban environment via environmental sensors and machine learning algorithms that integrate numerous data sources.”

Sidewalk has been working with the U.S. Department of Energy’s Argonne National Lab to test dozens of different sensors, grouped into four areas. “Environmental” sensors will measure humidity, atmospheric pressure and the temperature of the air, street and sidewalk. “Air Pollutant” devices would track particulates, ozone, carbon monoxide, nitrogen and sulfur dioxides, and more.

The kiosks would also follow “Natural and Manmade Behavior” by measuring vibrations from vehicles, magnetic fields, sound levels, and infrared, visible and ultraviolet light. Finally, “City Activity” monitoring would include collecting anonymized sightings of wireless devices (probably via Wi-Fi) and a video camera watching the kiosk’s surroundings. Each city could choose which sensors it wanted to install in its kiosks.

“We believe making aggregated views of this data public will contribute to the expanding startup community developing innovative products for shared public spaces,” says the document. Sidewalk itself plans to use the wireless device data and video monitoring of passing people and vehicles to calculate average roadway speeds and travel times. Alphabet would then feed this data back into its Google Maps navigation app.

Sidewalk suggests that the video sensors might spot “abandoned packages or objects,” raising the possibility of the technology being used to foil terrorist incidents. More mundanely, the company says the camera could also detect clogged drains and standing water on roadways.

“It is a win-win solution,” says Alexei Pozdnoukhov, director of the Smart Cities Research Center at the University of California at Berkeley. “Cities get environmental sensing to meet public health regulations and figure out the ‘livability’ of streets, and companies get … a playground for building new services.”

Sidewalk’s documents suggest that future developments for the kiosks might include small cells to improve cellphone reception or integrating them into Internet of Things networks.

Sidewalk Labs says that all data would be anonymized, with any personally identifiable information being encrypted and routinely deleted after being aggregated. No raw data would be sold or shared with any third party, including Alphabet and Google. Sidewalk also insists: “We will never collect or use PII gathered through our sensor platform.”

Following a complaint by the New York Civil Liberties Union in March about LinkNYC’s privacy policy, the company has confirmed to Recode that its kiosks’ cameras have never actually been turned on.

Locations of LinkNYC sidewalk kiosks

The documents obtained by Recode also spell out the financial benefits of installing Sidewalk’s kiosks. While the kiosks themselves would be supplied free of charge, cities would be responsible for the cost of installing them ($12,900 each) and connecting the gigabit internet optical fiber ($15,000). Cities would also contribute $5,000 per kiosk to a “warranty and hardware refresh fund,” presumably to allow for repairs and upgrades.

Along with these one-off costs, Sidewalk predicts that each kiosk would run up $1,440 in maintenance fees, $2,400 in power bills and a hefty $8,400 in fiber charges annually. The total first-year expenses for a city installing 100 kiosks would be more than $4.5 million.

If a city chooses to upgrade its kiosks to include advertising, however, the numbers look far more attractive. For a start, Sidewalk would finance the $23,000 cost per unit of installing twin 55-inch screens. It would also sell and place the ads, raising an estimated $60,000 per year from each kiosk. Sidewalk would keep half of this, giving the city annual revenues of $3 million from its fleet. If Sidewalk’s predictions are correct, the kiosks would pay for themselves in less than two years.

Even so, Sidewalk appears to be facing a tough sell. Rory McGuiness of Columbus’s Department of Development says, “We had meetings with Alphabet about the kiosks. There are some intriguing things about them, but there are also other companies that also have some very interesting kiosks. We have not signed any agreements yet.”

Whether it is Sidewalk’s kiosks or those of a rival that end up on the street corners of Columbus, two things seems certain. Urban Wi-Fi is about to get a whole lot faster and easier, and our cities a whole lot more connected.

This article originally appeared on Recode.net.