The AI venture capital circle is booming: only one year after its establishment, the super dark horse raised 1.3 billion US dollars, but the big hit only raised 25 million?

Text: Juny Editor: VickyXiao

The AI investment boom continues. Today, the new financing of several popular companies in the AIGC field is once again officially announced!

But if you take a closer look, the difference between the company and the company is really a bit big.

Among them, Inflection AI, which has just been established for more than a year and has only 35 employees, has become a "superstar" with an absolute dark horse attitude. It has received an investment of up to 1.3 billion U.S. dollars, and its valuation has exceeded 4 billion U.S. dollars in one fell swoop. **This round of financing is not just as simple as having more money, looking at its list of investors is enough to make your jaw drop.

In addition to the collective support of the two AI overlords, Microsoft and Nvidia, individual investors include: Microsoft founder Bill Gates, former Google CEO Eric Schmidt, LinkedIn co-founder Reed Hoffman …

But on the other hand, Stability AI, the super darling of the capital market, which was previously tied with Midjounery for two big mountains in the field of AI-generated images, seems to be in financing difficulties. **Recently, there have been frequent scandals and high-level executives leaving Stability AI. This time, it can only raise 25 million US dollars through the sale of convertible notes. **

However, Runway, a popular unicorn that used to be "incompatible" with Stability AI and focused on AI video generation, continued to chase after the victory and made great strides, and received a $1.4 investment led by Google and Nvidia.

Inflection AI: Although the company is small, giants and bosses are chasing after them

Let's take a look at the biggest winner this time, Inflection AI.

Inflection AI is a very young company. It was just established in Palo Alto, Silicon Valley in March last year. Currently, the total number of employees does not exceed 35. Although the company is small in size, its team strength is very strong.

Inflection AI has three co-founders, namely: Mustafa Suleyman, co-founder of Google DeepMind, Reid Hoffman, co-founder of LinkedIn, and Karen Simonyan, a former senior researcher at DeepMind.

The main business of Inflection AI was established around the training and development of large language models. Only two months after its establishment, it received US$225 million in Series A financing. **Inflection AI launched its first product - Pi in May this year, which is a chat robot similar to ChatGPT, but unlike ChatGPT's general artificial intelligence positioning, Pi is positioned as an artificial intelligence for personal services. ** Mainly around the needs of personal life, such as helping you organize your schedule, prepare for meetings and learn new skills, remember conversations with you, etc.

(About the entrepreneurial story of Inflection AI and more evaluations of Pi, Silicon Stars will publish a special report, please continue to pay attention!)

Pi is developed based on Inflection's internal large language model. The funds raised by Inflection AI will also be used to cooperate with Nvidia and computing infrastructure company CoreWeave to build a supercomputer and continue the training of large-scale artificial intelligence models.

This round of financing is also the first time Nvidia has appeared on the investor list of Inflection AI. For example, Microsoft, Bill Gates, and Eric Schmidt have been among the investors since the inception of Inflection AI.

It is worth noting that the single financing amount of US$1.3 billion is very impressive in the entire AI field. This also allowed Inflection AI to directly squeeze into the top five companies with the highest single financing in the entire AI field in the world (not just AIGC). The companies with a larger single financing amount are currently only OpenAI, Argo AI, Cruise .

**Inflection AI’s round of financing is valued at as high as US$4 billion, ranking third in the AIGC field after OpenAI and Anthropic. **

Image via Forbes

**The "scandal-ridden" Stability AI scares away investors? **

Compared with the infinite scenery of Inflection AI, Stability AI has recently seemed a lot lonely.

To say that the first fire of AIGC should be "fired" by AI Wensheng Tu at the beginning, and then ChatGPT turned out to push the craze to its peak. The two hottest and most promising start-up companies in the field of AI textual graphs are Midjourney and Stability AI.

Different from Midjourney taking the user route, the Stable Diffusion open source technology used by Stability AI has created a prosperous plug-in ecology, and its strong customization output capabilities make it the choice of more professionals. Relying on the success of Stable Diffusion, in October last year, Stability AI received a US$100 million seed round investment led by Lightspeed, with a valuation exceeding US$1 billion. **

Logically speaking, as a leading star unicorn and riding this wave of AIGC boom, Stability AI should have great potential. Referring to the current financing situation in the industry, it should not be a problem to raise more than 100 million yuan. However, due to the recent frequent scandals of Stability AI, investors are discouraged.

Earlier this month, Forbes published a feature article about exposing the problems behind Stability AI, including but not limited to: founder Emad Mostaque's academic qualifications, fabricating that he has a master's degree in computer science from Oxford University; bragging about Stability AI's business, such as in many Amazon, the United Nations and other institutions were listed as "strategic partners" on the road show, but they were not actually authorized; employees' wages were owed many times, and the company's internal management was chaotic, etc.

**And most importantly, even Stable Diffusion, its most famous product, is suspected of plagiarism. **

According to reports, the predecessor of Stable Diffusion was actually Latent Diffusion developed by the start-up company Runway and a group of researchers from German universities. At that time, due to limited funds and computing power, Stability AI actively approached them and offered to help them with model training. But after significantly improving the performance of Latent Diffusion, Stability AI also quietly took Latent Diffusion as its own and began to show it to investors in the name of Stable Diffusion.

In fact, this matter has been traced long before this report. Last year, after Stability AI announced its $100 million financing, Runway and Stability AI had a public "mutual tear" and turned their faces directly.

Picture taken from Forbes

Although Emad Mostaque recently stood up to respond to the doubts in the article one by one and bombarded Forbes' report, but the response also indirectly admitted the facts such as false academic qualifications and exaggeration. But it also once again challenged Stability AI's plan to seek financing at a $4 billion valuation.

Recently, **multiple channels showed that Stability AI finally raised $25 million through the sale of convertible notes. But for the crazy money-burning model training, this money is undoubtedly a drop in the bucket. **

At present, Stability AI is still saying that the company is doing well, not short of money, and is in the process of orderly expansion. But just this week, David Ha, director of research at Stability AI, and Ren Ito, chief operating officer, announced their resignations, and another big change happened at the top. Therefore, whether the current state of Stability AI is good or not, it is estimated that only the founder himself knows.

Runway: continue to focus on the AI-generated video track, and further raise 140 million

Compared with the current situation of Stability AI, Runway, who has been looking for another track after "parting ways", has recently received a lot of good news.

Today, everyone knows that the original basic model of Stable Diffusion was developed by Runway and some German researchers, but after Stability AI joined and participated in the training, Runway was inevitably marginalized.

After that, Runway still insisted on the AIGC field, but the subdivision field changed from picture to video editing and generation. The modes provided by Runway include text to video, **enter text and pictures to turn pictures into videos, and directly turn pictures into videos, etc. At the same time, you can also customize a video to your desired style through text description. **

A large number of special effects in the "Blinking Universe" that swept the Oscars before were completed by Runway, and because of Runway's technical support, the visual effects team of "Blinking Universe" is only composed of 5 people.

As a large model company, Runway's previous financing scale was not too large. At the end of 2021, Runway completed its Series B financing of US$35 million at a valuation of US$200 million. At the end of 2022, Runway completed its Series C financing of US$50 million, with a valuation of US$500 million.

Just a few months after the last round of financing, Runway was revealed to have completed a new round of financing of 100 million US dollars in early May. **Today, there is definite news that this round of financing should total $141 million, and the latest investors to join include Google, Nvidia, and Salesforce. ** This round of financing valued Runway at US$1.5 billion, which means that Runway’s valuation has tripled in less than half a year.

With the news of large-scale financing and acquisitions in the AIGC field coming one after another, this upsurge has not abated in the slightest.

The start-up companies that squeezed into the AI table, how to secure their position may be the next challenge they will face.