Regular updates on the latest VC-backed AI startups. Follow along to stay informed!
Gan.ai, an AI-powered video creation platform, raised $5.2M in Seed Round funding. The deal was led by Surge, Sequoia Capital’s rapid scale-up program, with participation from Emergent Venture and other angel investors.
Problem to be Solved
Video content is an effective way to reach customers, but personalized video ads are too time-consuming to scale. Gan.ai allows users to record a video once and then automatically customize it for every customer.
How They Use AI
GAN stands for generative adversarial network, the AI model architecture that is responsible for many of the so-called “deepfakes” that have been trending on Twitter. A generative adversarial network (GAN) has two parts: (1) A generator, which learns to generate plausible data, and (2) a discriminator, which learns to distinguish the generator's fake data from real data. Both the generator and the discriminator are neural networks. The generator output is connected directly to the discriminator input. Through backpropagation, the discriminator's classification provides a signal that the generator uses to update its weights.
Business Model
Usage-based. The company generates revenue by charging a fee based on the number of videos users generate. Early customers include Samsung and Zomato.
Sapient, an AI-powered test coder, raised $5M in Seed Round Funding. 8VC, Correlation Ventures, and GTMfund participated in the round.
Problem to be Solved
Writing test code manually is time-consuming, prone to human error, and a major roadblock to building better software faster.
How They Use AI
Given that Sapient just raised their seed round, it is unlikely that they are using any self-trained foundation models (even just the electricity costs are in the millions). This means they’ve probably built their tech on top of a third-party model, like ChatGPT. Their differentiators are their prompts, IDE (like MS Word but for coding) plugin, and platform. At a high level, they are using Gen AI to create unit tests according to software development principles.
Business Model
TBD. For now, software developers can download SapientAI Test Coder as a free plugin.
Together, a developer of open-source generative AI models, raised $20M in Seed Round funding. The round was led by Lux Capital with participation from Factory, SV Angel, First Round Capital, Long Journey Ventures, Robot Ventures, Definition Capital, Susa Ventures, Cadenza Ventures, and SCB 10x.
Problem to be Solved
Training, fine-tuning, or productizing open-source generative AI models is challenging because current solutions require that users have expertise in AI and are able to manage the large-scale infrastructure needed. Together is building a cloud platform for running, training, and fine-tuning open source models that the co-founders claim will offer scalable compute at “dramatically lower” prices than the dominant vendors (e.g., Google Cloud, AWS, Azure).
How They Use AI
Together’s platform is not built with AI as much as it is built for AI. They’ve made some notable contributions to the open-source model space already with their collaboration on the RedPajama dataset, which is considered on par with state-of-the-art proprietary datasets used to train models like GPT-4. High-quality, open-source datasets will pave the way for increasingly better open-source models (like Meta’s LLaMA models).
Business Model
TBD. The platform has yet to launch in GA and has made no revenue. Monetizing the open-source models may prove difficult, as the number of alternatives from community groups and large labs grows by the day. On the AI hardware infrastructure front, Together will likely follow the same business model as the big 3 cloud service providers but at a lower price (similar to CoreWeave).