- NEX creates "controllable and IP-consistent" image-generation models for creative tasks.
- Foundation Capital led the company's $3.5 million seed round.
- CEO Darius Lam strives for NEX to be the "visual media tool for the AI era."
Creating an advertising campaign is no easy feat. "It's a whole operation," said NEX CEO Darius Lam. To pull off a campaign, major brands fly celebrities and athletes to a studio, hire professional photographers, and shoot images with the product they're marketing, whether it's jewelry or a handbag, said Lam. "Gathering a set of images might cost thousands or tens of thousands of dollars."
Lam aims to change that with NEX, which develops AI image tools for entertainment, gaming, and marketing. Founded in late 2023, the company recently raised a $3.5 million seed round led by Foundation Capital, with participation from Point72 Ventures.
NEX creates "controllable and IP-consistent" image-generation models for creative tasks, according to its website. Instead of time-intensive and costly photoshoots, customers can automatically generate images for a fraction of the cost, Lam said.
NEX does this by enabling users to generate images inspired by other images, not just text inputs, in a "real-time canvas." The platform gives users a high degree of control by enabling enhancements to image quality and adjustments to color, pose, or depth in the images.
Additionally, NEX can integrate with customers' data to generate content consistent with their IP. The startup's tech can also tailor the AI models based on a customer's existing data, a process called finetuning, or even develop models from scratch. For example, "if you are a film studio that's been operating for decades and you have access to huge amounts of data," said Lam. "Then we could effectively build the model from scratch and it would be completely IP safe."
NEX initially took a consumer-focused approach, attracting nearly 100,000 users within a few months of launching. But the company has since pivoted to targeting business users, selling to small businesses and enterprise customers. The startup now generates revenue by charging up to $50 a month per user for access to its platform and through fees for model development and maintenance.
While models like OpenAI's Dall-E and consumer-focused Midjourney are popular, many don't offer the level of control and customization that design professionals require, said Lam. Some customers may seek to develop in-house models on top of open-sourced ones but it ends up being very expensive, he added.
Setting up and onboarding machine learning engineers, GPUs, and infrastructure is very difficult and costly. "It's much easier to work with a provider like NEX that not only is building the models but can integrate the models directly with applications that their creative teams can use."
Other visual AI companies include BRIA, which raised a $24 million Series A in March and targets enterprise users, and Cuebric, an early-stage company focused on the film production process.
Lam leads a small team of four employees at the young startup. Before founding the company, he was the head of product for computer vision at Cerebras Systems, which develops AI chips and confidentially filed for an IPO earlier this year. While companies like Adobe and Autodesk thrived in the internet age, Lam expects AI to completely upend the software landscape for creative tasks. This transformation will extend to areas like 3D design for game development, product design for software development, and visual creation for entertainment. Lam's vision for NEX is to be the "visual media tool for the AI era."