The first batch of AI companies have started to go bankrupt.
The pace of development in the AI field is truly rapid. It may only take a few months for an AI venture to go from being highly sought - after to being abandoned by venture capitalists.
While the global AI sector is booming and the huge financing rounds of workflow and Agent companies are dazzling, the UK AI startup Robin AI has been listed on a bankruptcy website due to the failure of its latest round of financing.
Robin AI is no unknown entity. At the beginning of this year, it was regarded as a rising star in the European AI industry. In the past year, Robin AI completed multiple rounds of financing, attracting top - tier global investors. Google, SoftBank, Temasek, PayPal, and the University of Cambridge are all its shareholders. In January 2025, it was included in the Top 10 list of UK technology companies by The Sunday Times.
It only took Robin AI half a year to go from heaven to hell.
Once One of the "Twin Stars" in the Field
Robin AI is a company focusing on legal AI. In this wave of AI, the legal field is considered one of the most promising application scenarios, with numerous startups emerging. Robin AI was one of the early pioneers.
The essence of legal work is the processing of language and rules. Whether it's drafting contracts, analyzing precedents, or writing legal opinions, the core is to conduct reasoning, comparison, and creation within a large, structured text library (laws, regulations, precedents) according to established rules (legal logic).
These are exactly what large - language models excel at. Not surprisingly, the legal industry has become a battleground for AI giants, witnessing a fierce arms race.
In 2022, Robin AI reached a strategic cooperation with Anthropic, integrating the latter's Claude large - language model into its products. Almost simultaneously, another newly - established legal AI startup, Harvey AI, partnered with OpenAI and received investment from the OpenAI seed fund. For a while, Robin AI and Harvey AI were seen as the two luckiest "fast horses" in the legal AI field, with Anthropic and OpenAI, two AI giants, competing indirectly behind them.
There is a reason why Robin AI was chosen by Anthropic.
The founding team of Robin AI is excellent. The two founders are a senior lawyer and an AI researcher respectively, both young and talented. Co - founder and CEO Richard Robinson was the coach of the English national debate team and worked as a lawyer at Clifford Chance, one of the world's largest law firms, for seven years. The other co - founder and Chief Technology Officer, James Clough, is a doctor from Imperial College London and a post - doctoral researcher at King's College London.
As a senior lawyer, Richard Robinson has firsthand experience of the pain points in the legal industry. Even high - paid elite lawyers have to spend a lot of time on "necessary but low - value" repetitive work. Take the PE industry as an example. In a merger and acquisition, lawyers often need weeks to review thousands of contracts, while an AI can finish it in minutes. The goal of Robin AI is to be a "copilot" for lawyers and change the way they work.
In media interviews, Richard Robinson and James Clough described two visions for the company.
One is relatively modest - if Robin AI can free lawyers from one hour of a ten - hour workday, it will have high economic value. This is because (in Europe and the US) lawyers charge by the hour, and the fees are very high.
The other is more ambitious. James Clough predicted in an interview in 2022: "AI will create two types of lawyers: those empowered by AI and those left behind by the times."
13 Fortune Global 500 Clients Alone
Unexpectedly, before lawyers were left behind by the times, Robin AI was abandoned first.
In the past three years, Robin AI has been on a roller - coaster ride. It first reached the peak and then suddenly declined when it was at its height.
Robin AI was founded in 2019. In the early days of its startup, it mainly tried to train proprietary models with legal data and didn't attract much attention. After 2022, with the popularity of large - language models and the cooperation with Anthropic, Robin AI took off.
· In July 2021, Robin AI completed its seed - round financing, with Google and Episode 1 as investors.
· In February 2023, Robin AI completed its Series A financing, mainly led by SoftBank Group and the founder of Wise.
· In January 2024, Robin AI completed its Series B financing, led by Temasek, with follow - on investors including QuantumLight, Plural Ventures, AFG Partners, etc.
· From July to November 2024, Robin AI completed several Series B+ financings, with investors such as PayPal Ventures and the University of Cambridge.
With successive rounds of financing and to compete with Harvey AI, Robin AI expanded its business rapidly.
Robin AI's early products mainly focused on contract review based on self - developed models. After the cooperation with Anthropic, it launched full - text search and intelligent Q&A products based on Claude, supporting natural - language queries of contract terms and having functions such as automatic identification of contract types and key terms. In 2023, Robin officially launched its contract review Copilot product and cooperated with Anthropic to optimize long - text understanding ability. In November 2024, Robin released Robin Reports, a powerful legal AI agent that can summarize thousands of contracts at once and recommend actions based on the analysis results.
While advancing its product line, Robin AI also started global expansion, opening an office in New York in July 2024 and another in Singapore in September.
During the Series B financing, Robin AI claimed that its technology could reduce the time required for contract review by over 80% and costs by 75%. Moreover, Robin AI had won a wide range of customers globally, including 13 Fortune Global 500 companies. Its customer list includes prestigious names such as UBS, General Electric, Pfizer, Bristol - Myers Squibb, PepsiCo, Yum! Brands, Blue Origin, KPMG, and PwC.
In January 2025, Robin AI ranked tenth on the "2025 UK Tech 100" (The Sunday Times 100 Tech list) compiled by The Sunday Times, one of the most authoritative media in the UK.
Sudden Death in Series C Financing
However, from this point on, Robin AI was suddenly abandoned by capital.
At the end of October, the media found that Robin AI was listed on a website called "IP - BID.com", a digital marketplace in the UK for trading insolvent enterprises and non - performing assets. Being listed on this website is not strictly bankruptcy but an attempt to sell at a low price before formal bankruptcy.
Subsequently, more media reported that Robin AI started a new round, Series C financing, at the beginning of 2025. However, the financing still hadn't materialized by October. In fact, Robin AI had been quietly laying off employees due to the failed financing. In February 2025, the total number of Robin AI employees reached a peak of over 200, dropping to about 150 by September, and in October, it announced that it would lay off about 50 people.
By now, failure seems to be the only outcome for Robin AI. So far, no buyers have registered to purchase Robin AI on the "IP - BID.com" website. If the sale fails, Robin AI will soon go bankrupt. Even if the sale succeeds, it will be at a bankruptcy price.
This situation is really sad. How did Robin AI, which was highly sought - after half a year ago, end up like this? It's not that it has no customers or fails to solve industry pain points. Based on various comments, there are two main reasons for Robin AI's downfall.
The first reason is insufficient growth. It is reported that Robin AI's revenue doubled in 2024, and its US business revenue increased six - fold. It sounds good, but investors evaluated it as "far from achieving AI - level growth". An investor told the media that VCs are willing to pay a valuation of 50 - 100 times ARR (Annual Recurring Revenue) for AI companies based on an expected annual revenue growth of 3 - 5 times, and this growth should come from businesses with a gross margin of over 80%.
This shows the cruelty behind the booming AI field. There is abundant capital here, which can quickly create a unicorn. But if you slack off for a moment, you may be out of the game immediately.
The second reason is even more cruel - Robin AI is considered "not AI - enough".
Since its establishment, Robin AI has always emphasized that it is an "AI - native" company and even included AI in its name. However, to achieve product launch, Robin AI chose a heavy - asset model.
As we all know, the legal industry has zero tolerance for errors and attaches great importance to risks. However, even the most advanced large - language models still cannot eliminate "hallucinations" and errors, which has become the biggest obstacle for AI in the legal industry. To solve this problem, Robin AI proposed the so - called "SaaS + services", also known as "lawyer in the loop". The concept is a bit convoluted, but the essence is simple: Robin AI hired dozens of qualified lawyers in - house and set up an outsourcing team in India to conduct final review and quality control of the AI's output.
This model enabled Robin AI to win a group of top - tier financial institutions and Fortune Global 500 customers. However, in hindsight, it was a decision that was correct in the short term but a failure in the long term.
Because the heavy - asset model made Robin AI a "jack - of - all - trades and master of none".
On the one hand, Robin AI did have customers and revenue, but the revenue was not high and far from covering the costs. In 2024, Robin AI's revenue was only about $10 million, while the net loss exceeded $14 million.
On the other hand, although Robin AI is an AI company, it is labor - intensive, and its business expansion depends on labor expansion. In the era of rapid development of AI technology, it seems more and more outdated. Which investor would be interested in a "legal service outsourcing" company?
As a result, Robin AI has no hope of self - sufficiency and cannot obtain financing, so it has to leave the market dejectedly.
Is the AI Bubble Bursting? Not at All
While Robin AI failed in its financing, new legal AI unicorns are still making great progress.
In October 2025, Harvey AI, supported by OpenAI, completed a $150 million financing round, led by A16Z. It is reported that its valuation has reached $8 billion.
It is reported that Harvey AI's Annual Recurring Revenue (subscription revenue of the most recent month multiplied by 12) was $50 million at the end of 2024. The revenue gap between Robin AI and Harvey AI is not huge, but their fates are worlds apart.
Besides Harvey AI, more legal AI startups are emerging like mushrooms after rain.
On October 30, Legora, a Swedish legal AI startup, announced the completion of a $150 million Series C financing, with investors including well - known institutions such as Redpoint Ventures, Benchmark, and Y Combinator. Legora was founded in 2023, and its valuation reached $1.8 billion in just two years.
In fact, with the continuous popularity of AI applications and Agents in 2025, AI companies in various vertical fields such as programming, accounting, healthcare, and law are constantly receiving financing. Statistics show that as of October, the investment in the global legal AI field in 2025 has been twice that of last year, indicating that there is still a lot of capital in the AI field.
In China, statistics show that as of September 2025, about 764 AI companies in China have received investment, with a total amount of up to 83 billion yuan. Among them, Agents are one of the hottest areas for financing, with general - purpose Agents, vertical - type Agents, and edge - side Agents all booming. Xiniu Data said that the Chinese AI agent industry showed a trend of large - scale explosion in 2025. As of July 2025, 23 financing events had occurred, with the disclosed amount exceeding 8 billion yuan, and it is expected to exceed 15 billion yuan for the whole year. Companies such as Shudian, Zhongshu Ruizhi, and Butterfly Effect have all completed financings of hundreds of millions of yuan this year. In addition, giants such as Alibaba, Baidu, ByteDance, and Tencent have also launched Agent products publicly.
However, despite the excitement, whether Agents have barriers is still a hot - debated topic. The speed of AI technology update is much faster than in any previous field. Will a popular concept today become obsolete tomorrow? The fate of Robin AI may be a worthy reference.
Moreover, the current market valuation of AI companies means that even 100% growth is considered a failure.
On the day I was writing this article, the AI sector in the European and American stock markets tumbled across the board. It is currently the third - quarter earnings disclosure season. Many AI giants released what they claimed to be "better - than - expected" results but were mercilessly sold off by investors. Palantir, known as the "first stock in AI applications", reported a 63% year - on - year increase in revenue and a 231% year - on - year increase in net profit, but its stock price dropped by 8%.
This easily reminds people of the investors' verdict on Robin AI as "not achieving AI - level growth".
Not long ago, Goldman Sachs issued a research report warning that the "biggest risk in the current market is that the revenue of technology stocks will start to disappoint investors, and there are doubts about whether the current rate of return can be sustained. Such an event can at least trigger a major correction, although it may not be a full - scale collapse like the bursting of the Internet bubble."
For AI companies, they either achieve exponential growth or may quickly turn from "unicorns" to "cast - offs". There is no middle ground on the AI highway.
This article is from the WeChat official account "China Venture Capital", author: Tao Huidong, published by 36Kr with permission.