As the world continues to figure out how to deal with the explosion of deepfake content online, it appears that not all AI-generated videos will spark controversy. Synthesia, a London-based startup building products around highly realistic AI avatar technology, says it has been hugely popular among businesses, with some 60,000 businesses (1 million users) using the technology to build avatar-based avatars from text documents. Videos, for sales and marketing, for training, etc.

Now venture capitalists want in on the action. Synthesia today confirmed it has closed a $180 million Series D funding round that jumps the company's valuation to $2.1 billion. NEA led the round, with participation from new investors WiL (World Innovation Lab), Atlassian Ventures and PSP Growth, as well as previous backers GV and MMC Ventures. Synthesia has raised $330 million to date.

The startup plans to use the funds for hiring, specifically expanding in the Asia-Pacific region (most of Synthesia's current operations are in Europe and North America) and continuing to improve its product.

“We’re doubling down on everything we’re already doing right,” CEO and co-founder Victor Riparbelli said in an interview. "We want to make our avatars better." He said the company's "long-term roadmap" includes more realistic movements; the ability to port avatars to different environments; avatars that can interact with objects, such as providing physics demo; and avatars that can interact with users. It will also try its own approach by building more "agents" to help customers create avatar-based content more easily.

One area where it doesn't chase activity is mergers and acquisitions. Synthesia has not made any acquisitions so far, and Riparbelli said it prefers to build technology in-house while using APIs for things it doesn't build. For example, it works with Eleven Labs to develop speech and leverages and fine-tunes various third-party large-scale language models instead of building its own models.

Synthesia's funding round has been in the works for at least a few months: The Information reported that it raised $150 million in November 2024. For more fundraising context, it's been about 18 months since Synthesia last disclosed funding: in June 2023, it closed a $90 million round at a $1 billion valuation, with previous backers including Kleiner Perkins Kleiner Perkins and Accel.

During this period, artificial intelligence companies have become hugely attractive to venture capital firms, providing a bright spot in a somewhat bleak funding environment. PitchBook data shows that investment in all global startups will reach US$368.5 billion in 2024, of which artificial intelligence startups account for more than 37%. In the United States, this proportion is even more pronounced, with artificial intelligence startups receiving nearly 50% of the $209 billion in investment last year.

Yes, there are many problems in the field of artificial intelligence. The power consumption required to train and run AI models, major copyright issues regarding how models are trained, AI being weaponized in the case of deepfakes or malicious hacks, AI replacing humans and their jobs, and AI going wrong - All these huge problems need to be solved. has yet to be meaningfully resolved. But there are some important advocates who will push the AI ​​industry to greater and more dramatic heights. Synthesia is one of the companies named by the UK government this week as it launches its massive AI action plan, which aims to provide billions of dollars in deals to AI companies to rebuild public services and the economy.

Synthesia says it now has 60,000 enterprise customers, up from 50,000 in June 2023, and it aims to carve out its own niche in the space and become the go-to platform for businesses looking to build video interactions.

It does so at a time when advanced artificial intelligence video capabilities are becoming more commonplace. Some startups are working on the ability to infer full product videos from basic documentation, while others are working on building avatars and live video assistants that can interact in real time. Some claim to be able to create lifelike avatars of users using just one minute of video. (To get a sense of how crowded the market is here, a simple test is to put Synthesia into Google and see how many companies buy search ads based on its name. There are a lot.)

Synthesia isn't immune to product competition. It's been building a "2.0" version of its platform for some time now, and has already released a number of related features, including the ability for users to make their own avatars using their laptop cameras or their phones with emotions; a Chrome extension that can create custom avatars based on screen data Build Basic Video; its own version of an AI-powered video assistant that can convert documents into videos; multi-language options; and collaboration features for people to edit videos at the same time.

More importantly, though, Ripal Bailey believes the company has an advantage in focusing on business users, which its investors say is part of the startup's appeal.

"Synthesia is one of the few companies that can take really cutting-edge AI and turn it into something that has practical uses," Vidu Shanmugarajah, a partner at Google Ventures in London, said in an interview. "It's very customer-focused. They're obsessed with driving value in a physical environment. Putting that into a secure and compliant platform is very difficult."

It's also interesting to see Atlassian investing in this round. The company has been infusing AI capabilities into its various applications, and it seems only a matter of time before a platform like Jira might start adding more video tools to the mix, opening the door to collaborations with its portfolio companies.