ChinAI #287: Horizon Robotics secures largest Hong Kong IPO of 2024
The Chinese AI chip firm goes public
Greetings from a world where…
four ounces can conquer five hundred kilograms [四两拨千斤]
…As always, the searchable archive of all past issues is here. Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors).
Feature Translation: Horizon Robotics Goes Public
Context: Last week, Horizon Robotics listed on the Hong Stock Exchange, becoming the exchange’s largest tech IPO of 2024. This week’s translation (link to original Chinese) features insights on Horizon Robotics’s journey from from Zhang Peng, head of an influential start-up community and a friend to Yu Kai, Horizon’s founder. ChinAI has covered Horizon since July 2018 (ChinAI #20). Earlier this year, I analyzed the competition between Horizon and Black Sesame [黑芝麻智能] in system-on-chips that power advanced driving assistance systems and autonomous driving (ChinAI #265). One detail from that issue, which might speak to why Horizon’s IPO has been more successful than Black Sesame’s debut on the Hong Kong stock exchange:
Software matters even when the vertical is hardware. Horizon and Black Sesame have two main revenue streams: the SoC-based solutions and algorithm-based solutions, which refers to the licensing of software and development tool chains to customers. Horizon’s revenue from licensing increased from 43% in 2021 to 62% in 2023; by comparison, Black Sesame’s proportion of revenue from software licensing declined from 54 to 26.5 percent over this same period.
My favorite past issue about Horizon Robotics, however, is a two-part history (ChinAI #221) of the Silicon Valley lab of NEC (a Japanese multinational information technology and electronics corporation), which was a crucial wellspring for China’s AI development. The article concludes: “Nowadays, in China’s AI terrain, the Chinese people who came out of NEC Lab in Silicon Valley have laid down roots everywhere, growing with wild abandon.” That statement certainly applies to Yu Kai, who went from director of the NEC lab to lead Baidu’s institute of deep learning and then eventually to start Horizon Robotics.
Key Takeaways: Horizon Robotics comes to a difficult crossroads in 2019.
The company is focused on two main paths: 1) autonomous driving; 2) “Artificial Intelligence of Things [AIoT]” (e.g., smart homes, air conditioners, and sweeping robots). However, despite the fact that their revenues in AIoT had reached nearly 200 million RMB in 2019, Horizon decides to cut off this business line and focus solely on intelligent driving technology.
This decision becomes even more puzzling when one considers that 2019 was a horrible year for the new-energy vehicle industry. From the article: “At that time, Wei (NIO), Xiao (Xpeng Motors), and Li (Li Auto) were nicknamed ‘the three idiots of the new forces’….Li Bin (founder of NIO) was also deemed “the most miserable person of 2019.”
Horizon’s choice reveals much about the difficulty of providing general-purpose services across a variety of verticals.
Even when Horizon invested the resources to cultivate particular AIoT applications, penetration was slow: “It seems that in order to make each vertical field bigger and better, it is necessary to divide troops and cultivate land, which is extremely inefficient. For example, in some fields, Horizon has tried its best, but even if the market demand has been fully verified, it discovered that the market was already being viciously competed over such that victory was as bad as defeat.”
In making this difficult call — which resulted in laying off 500 employees (out of a total of 1,200) — Yu Kai drew on lessons from the development of Microsoft and Android: “Microsoft's success did not come from building a platform from the beginning, but by deepening a vertical scenario (such as white-collar office software) to promote the rise of the Windows platform. Microsoft revolves around white-collar productivity tools, and Office is its core product, which in turn supports the success of Windows. So, for Horizon, if you choose a vertical scenario to support the future platform, it can only be smart driving.”
Horizon has a long way to go, and there are some very strong companies that are their direct competitors, but here’s where the company stands now: “Horizon has reached mass production cooperation with the top ten companies (conglomerates) in China's automobile sales. In the first half of 2024, the cumulative shipments of the Journey series computing solutions exceeded 5 million sets, making Horizon the automotive computing platform supplier with a market share second only to NVIDIA.”
The full translation has some additional notes from Horizon’s 701-pg. prospectus. SEE FULL TRANSLATION: Horizon Robotics Goes Public: With 3 billion in hand that year, he chose to sever a profitable business line
ChinAI Links (Eight to Exchange)
I was able to catch up on some reading this week, so this week’s ChinAI links are a double-stuffed version.
Must-read: Chinese AI Safety Institute Counterparts
By Karson Elmgren and Oliver Guest, this Institute for AI Policy and Strategy report supplies an essential analysis off five Chinese institutions that are doing similar work to the US and UK AI safety institutes: CAICT; Shanghai AI Lab; TC 260; Institute for AI international Governance (at Singhua); and Beijing Academy of Artificial Intelligence.
Should-apply: US-China AI Governance PhD Fellowships
The Future of Life Institute has launched a PhD fellowship (tuition and 40k annual stipend for five years) to support students that want to study risk reduction in US-China relations in AI. Deadline to apply is November 20; more details and application portal here.
Should-apply: Miles Brundage — Why I’m Leaving OpenAI and What I’m Doing Next
I got my start in this field sitting by Miles in the GovAI offices. His reflections on leaving his post as the head of OpenAI’s artificial general intelligence readiness team are well worth the read, and I’m excited to see what he works on next.
Should-read: Moving Data, Moving Target
Samm Sacks, Krystal Chen Zeng, and Graham Webster recently published a fascinating DigiChina essay on China’s 2024 rules on cross-border data flows, including details on free trade pilot zones at the subnational level that could lower regulatory burdens on data flows.
Should-listen: The Maven Smart System and the Future of Military AI
The Modern War Institute hosted Emmy Probasco and Igor Mikolic-Torreira, from the Center for Security and Emerging Technology. They dive deep into the Maven Smart System, an AI-enabled software by the US Army’s XVIII Airborne Corps.
Should-attend: Two book talks at Oxford
I’ll be hopping across the pond later this month to do two book talks. Come say hi if you’re around:
November 26, Tuesday noon (UK time) at the Oxford Martin School (co-sponsored by the Centre for the Governance of AI and the Oxford Martin AI Governance Initiative)
Should-read: A Pregnant Teenager Died After Trying to Get Care in Three Visits to Texas Emergency Rooms
Everyone should take a few minutes to read this ProPublica investigative piece on how Texas’s abortion ban caused the death of an 18-year-old woman.
Thank you for reading and engaging.
These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is an Assistant Professor of Political Science at George Washington University.
Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).
Also! Listen to narrations of the ChinAI Newsletter in podcast format here.
Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99