Production | Huxiao Commercial Consumption Group
Author | Zhao Xi, the author sent by Hu Xiu to Singapore
Edit | Miao Zhengqing
Head picture | The Fantastic Drifting of the Young School
Tiger sniffing note: This is the tenth article of the Singapore Business Watch series "Snowball", focusing on the field of AI venture capital in Singapore.
The Snowball series presents the current situation of the subdivided commercial track, reflects the changes of Singapore's business environment, and provides samples and cases for players of all sizes who go to Singapore. If you are an entrepreneur rooted in Singapore, a sea going enterprise, or an investor who has long followed Singapore, please keep an eye on this series, and you are welcome to share industry clues.
A Starbucky in the core business district of Singapore, Fiona, who has experienced wheat skin in the tropics and painted eyeliner, introduced her AI products to me lively. The young CEO, who graduated from Zhejiang University and was born in 1999, founded GENPULSE in 2021. He went all the way south and directly entered the venture capital market of Singapore in northern latitudes.
GENPULSE aims to develop the best hair AI in the world. On April 22, 2024, GENPULSE released Lushair Hair Follicle Intelligent Precision Measuring Mirror, which uses image recognition technology and segmentation models in the feature field to intelligently analyze scalp images under the hair follicle mirror.
Although the main body of the company is registered in Singapore, the Singapore market has never been the main battlefield of GENPULSE. GENPULSE China is dominated by Hangzhou Siyue Technology, and 90% of its products are sold to China. After the financing in the second quarter of this year, Fiona will begin to expand its market in Australia and North America in the third quarter.
One of the reasons for choosing Singapore is that GENPULSE has been incubated by the accelerator of the National University of Singapore, and will receive hundreds of thousands of Singapore dollars (more than one million yuan) of investment upon expiration. The recognition from Singapore has led to "learning from the famous" and "natural flow" in the Southeast Asian market.
This fledgling AI company is a small epitome of Singapore AI start-ups: they enjoy all kinds of resources brought by this country, while undertaking the commercialization problems caused by the narrow market, and almost without exception need to expand outside. A large number of founders are from other countries, and there are few local entrepreneurs.
Singapore has strong academic capacity and government support. And because of its unique strategic positioning, Singapore can undertake AI business that some countries cannot, and become a global overseas data center. A VC investor from China working in Singapore told Tiger Smell that more AI companies would land in Singapore in the future.
But at present, Singapore's name is easy to be ignored in the global AI competition led by China and the United States. The managing director of a capital in Singapore directly said to Hu Xiang: "AI technology in China and the United States is relatively strong, and the opportunities in Southeast Asia are not particularly great."
In short, Singapore seems to have the right time and place in the AI field, but it is facing fierce international competition and market positioning challenges.
So, what's the problem?
Find location
Singapore is trying to find its own position in the conflict of interests between China and the United States, demonstrating its cooperation advantages in the AI field.
For example, in January 2019, S Iswaran proposed the world's first AI ethics framework at the World Economic Forum in Davos.
At the same time, Professor Wen Yonggang (an academician of the Singapore Academy of Engineering and a professor of the School of Computer Science and Engineering of Nanyang Technological University, whose research mainly focuses on the research and development of large-scale distributed computer systems and the innovation of related core technologies) told Hu Xiang that Singapore is short of energy, so he focuses on energy conservation research, and is comparable to Google in terms of energy conservation in data centers using AI.
He also said that Singapore is in a leading position in the academic field of artificial intelligence. Take the data technology and artificial intelligence major of Nanyang Technological University as an example. In the QS University ranking in 2024, this major ranks eighth in the global universities.
"Singapore's AI talent density is the highest in the world." He recalled that at several famous AI conferences, Singapore, as a country with a small population, generally ranked among the top four in the total number of papers published.
But talent mobility is too high. Many young scholars who received education and did academic research in Singapore were poached by other countries when they were about to become famous or just became famous. For example, Professor Wen's colleagues sometimes joke that "Singapore has a large talent deficit in Australia's AI field."
Academic achievements cannot be translated into the business world. An investor told Hu Xiang that he has been dabbling in the AI field, but has never invested in the AI project in Singapore. Last year, he talked about a project, the paper is very excellent, the results were learned by the academic and industrial circles. The reason for not investing is that we are still exploring a feasible path in commercialization.
The Singapore government has given great support to AI, which can be called "stepping down in person".
On February 16, 2024, Singapore announced its financial budget for the new year, and the authorities said that the investment in the next five years would exceed 1 billion Singapore dollars (about 5.35 billion yuan) for AI computing, talent training and industrial development.
One of the policies in the plan is to subsidize Singaporeans aged 40 and above in the middle of their career, who can obtain 4000 Singapore dollars (about 21400 yuan) for continuous education from May. Congressman Chen Youming said that the next 10 years will be the world of AI, so it is necessary to help Singaporeans learn how to use AI to improve Singapore's competitiveness.
At the same time, Singapore is committed to attracting the world's top AI creators to work in Singapore, cultivating AI professionals, and fully developing the country's AI potential. According to Singapore's National AI Strategy 2.0 (NAIS 2.0 for short), the government will expand the number of AI practitioners in Singapore to 15000, create a landmark base for AI creators and practitioners to work together, and cultivate the AI community in Singapore.
Nevertheless, in business, Singapore is far from the dream of becoming one of the leaders of AI.
Business Dilemma
"More and more capital is flowing into Singapore, but the development speed and commercialization degree of AI industry are still much lower than the speed of capital inflow. China is just the opposite, and Chinese companies are relatively pragmatic," said one entrepreneur. In particular, entrepreneurs from large Chinese factories have strong code ability and very flexible commercial thinking.
This triggered the grand occasion of China's 100 model war, and also created the chaos of some people shaking half a bottle of water and daring to sell classes online without anything. Such a situation of prosperity and chaos has not occurred in Singapore.
"Singapore has a good environment for entrepreneurship and innovation, and government policies are in place, but limited by its own market, its development potential lies in exports to ASEAN," said Professor Wen Yonggang. There are many international enterprises stationed in Singapore, but they have no business in Singapore, so they seldom purchase ToB application equipment. Most enterprises aim at financial security, or rely on Singapore to enter the Southeast Asian market.
"The infrastructure growth of AI in Southeast Asia is one of the fastest in the world. Due to the influence of local politics, Chinese and American companies will go to Southeast Asia to build their own data centers. Therefore, the training of AI is gathered here, and the data centers are growing very fast. But its market is not here."
However, the Internet in Southeast Asia has not formed a complete market. Unlike China and the United States, which are single markets, Southeast Asia has ASEAN, but its market is very fragmented, which is not conducive to commercial expansion.
According to the public information queried by Huxiao, a Singapore AI voice star startup company has raised a sum of 30 million dollars in 2022, and has not announced its financing since then.
Some investors commented: "The current AI companies are still looking for ways to improve their retail scale and income, and exploring the cash flow model, which is relatively far from real profitability."
Mao Hua, founder of Singapore based startup PlatoX ai, who focuses on personal assistant of artificial intelligence, said: "Singapore enterprises have no chance to encounter the application with such a large number of users in China, so their experience is not enough."
Local enterprises are not sensitive to this, and "aggressive" is obviously lower than Chinese start-ups. Fiona mentioned that she has cooperated with a local brand in Singapore, which has only 30000 users since its establishment in 2017, but is not ready to expand. "There is not much room for imagination."
There are also two challenges for Singapore AI entrepreneurship: capital and talent.
Professor Wen Yonggang said that Singapore and other capital markets have abundant medium - and long-term funds, but the early funds are few.
Mao Hua also said: "The wave of AI investment seems to be lively and full of funds, but everyone is cautious, and few can really make up their minds to invest."
"Cross river dragons" from other countries are also grabbing investors' attention. Professor Wen Yonggang mentioned that more and more Chinese AI companies came to Singapore to go to sea. From 2018 to 2022, AI research departments of Alibaba, Tencent, Shangtang, iFLYTEK and other enterprises will come to Singapore; In recent years, a number of medium-sized enterprises and early stage enterprises that are close to listing have also begun to flock to Singapore.
Capital has observed that such enterprises have mature technology and complete products, and can roll up directly only by capital injection. It is more inclined to invest in such enterprises than in Singapore's local start-ups.
At the talent level, although Singapore AI has strong academic strength, its talent size is obviously insufficient. On the one hand, there is an outflow of academic talents; On the other hand, local talents have low entrepreneurial willingness.
Mao Hua's intuitive feeling is that the weak local entrepreneurial atmosphere stems from the fact that the Singapore government has taken care of citizens too well. In other words, most Singaporeans can enjoy a happy and stable life without having to struggle in the business world.
A professor of humanities at Nanyang Technological University in Singapore told me that the stability of life also made Singaporeans dare not take risks, because the risk of accidentally taking a life fork was unacceptable to many citizens.
Therefore, many of the resources provided by Singapore's universities and government for entrepreneurs fall into the enterprises of foreign entrepreneurs.
An employee of an AI startup founded by the founder of Pakistan told me that their company, in cooperation with the National University of Singapore, is located in the office provided by the government, and there is only one company on the whole floor.
The inertia and foam of Web3
"Some so-called projects combining Web3 with AI have no users. Although there are no users, the valuation is already very high." The founder and CEO of a Singapore based Web3 infrastructure company told Hu Xiu. The CEO, who did not want to be named, is also an angel investor, and his investment field is involved in AI.
As the rich soil of Web3, Singapore has continued its entanglement with Web3 in the AI wave. Many Web3 entrepreneurs, smelling the breath of AI being able to finance, rushed to the market and formed a new bubble.
Daniel Zou, the founder and CEO of Blueshirt Technology, told Hu Xiu: "The wave of AI has attracted a large number of technology industry practitioners. Everyone wants to use AI to describe a story, and then issue a coin to the chain."
Generally, the market will recover in the first half of the year when Bitcoin is halved. The supply of Bitcoin is reduced, but the market demand is still there, and the price of Bitcoin will rise in response.
The current bull market of Bitcoin originated last October. The number of bitcoins is halved, the ETF is approved, and the price of bitcoins rises sharply from 30000 dollars to 70000 dollars. "After this wave, everyone is looking at what concept is worth hyping." The CEO of the aforementioned Web3 infrastructure company said, "After the popularity of OpenAI, investors are looking at what projects are well integrated between AI and Web3."
For example, there is a problem of data leakage in big models such as OpenAI, and entrepreneurs have a starting point of how to solve existing data privacy problems through decentralization.
Chris, co-founder and COO of Chainbase, told Hu Xiu that he believed that there were two kinds of directions in which Crypto and AI could combine. (Chainbase focuses on the combination of on chain data and AI, realizes the interoperability of on chain data through its first full chain data network, and trains Theia, the first industry basic model specially built for Crypto, based on this data network. Singapore has the main body, and has completed a round A financing of 15 million dollars.)
The first is Crypto for AI, which is used as a tool to solve some problems in the existing AI workflow. For example, in order to cope with the amazing growth of computing power demand in the AI field, the industry has generated decentralized AI to efficiently solve the problem of computing power matching, and lowered the threshold for the use of high-quality computing power.
For example, a Tsinghua team that started an overseas business founded NetMind AI， Pry a large number of idle computing power around the world, open the computing power network, and let customers pay for the call.
Kai, the founder and CEO of NetMind.AI, told Hu Xiu that AI and blockchain technology are a very natural combination: "We have been committed to the research of training, fine-tuning and reasoning in multi node asynchronous mode. This is a distributed and decentralized thing in itself. At the same time, idle computing power is better utilized and deployed through algorithms to form a network for our users. Many of the early computing power comes from mining projects such as Web3's Bitcoin Ethereum. "
At the same time, NetMind.AI has built its own netmind layer 1 chain. In the future, it hopes to use the blockchain for bookkeeping, including price, task allocation and other information, which will be more transparent and further eliminate the price difference caused by information difference.
The other is that AI for Crypto uses AI to make the Crypto industry widely popularized and develop faster.
Chris mentioned that Chainbase, like Ritual and Bittensor, which are well-known in the industry, belong to this kind of combination mode: "Bittensor is an open source protocol, which promotes the development of AI through the incentive structure driven by the blockchain, putting the reasoning part on the chain and computing it off the chain. Chainbase also hopes to create the first crypto native foundation model by making use of the characteristics of transparent and causal integrity of data development on the chain, transforming static data on the chain into dynamic model assets, so that the entire crypto has intelligent capabilities and further drives the large-scale popularization of the industry. "
Kai said that there are few successful business models combining Web3 and AI. Many people are hyping this concept, but companies that really create real value have not yet emerged.
Singapore, as the absolute center of Web3, is still in the starting vacuum period of the industry regarding the integration with AI.
Write at the end
One day after I returned to Beijing, Fiona sent a good news: "This week, our Leads came from Malaysia, Singapore, Brazil, the United States, Australia, Hong Kong, Vietnam..."
Starting from Singapore, a world port, young AI start-ups can link to resources around the world. In this global AI competition, Singapore, like these young companies, is looking for and consolidating its position.
However, with these huge potentials and resources, Singapore still faces many challenges in commercialization and market expansion.
