NEWSLETTER

By clicking submit, you agree to share your email address with TFN to receive marketing, updates, and other emails from the site owner. Use the unsubscribe link in the emails to opt out at any time.

From consumer app to enterprise platform: How Meshy is scaling AI-powered 3D creation

Meshy
Image credits: Meshy

The conversations around generative AI have been about text, images, video, and coding in the past few years. But in the background, a new conversation has been picking up momentum: AI-generated 3D content. Devices for immersive computing, e-commerce, robotics, additive manufacturing, gaming, and product visualisation are proliferating, and the demand for 3D assets is quickly outstripping our ability to produce them. Current modeling workflows haven’t changed much in 30 years, continue to require specialised software, rigid processes, extensive training, and create long production cycles. The result is a direct increase in cost and time-to-market, whether we’re building a startup or a major corporation.

That gap has fuelled one of the more intriguing subsectors of creative AI. Rather than automating the workflows of existing designers, a new breed of companies is betting on making 3D creation mainstream for everyone.

One of the startups getting onlookers excited: Meshy AI, the Silicon Valley startup that has gone from a consumer-friendly text-to-3D tool into a full suite for developers, designers, educators, manufacturers, and enterprise.

Why AI-generated 3D is becoming an investment category

The next generation of AI platforms is characterised by the production of multimodal content. Images do not have an independent life any longer. They serve in product configurators, virtual shops, game levels, fabrication flows, and digital twins.

However, traditionally, these assets have had to be created and annotated by hand: the manual creation of a single high-quality textured model, and, depending upon complexity, could take hours or even days. For enterprises at scale, needing to build thousands or millions of assets, it is clear that the maths do not work.

AI changes that.

Instead of supplanting veteran 3D artists, generative systems now tend to mostly offload monotonous design support duties, granting designers the ability to iterate faster and affording those without modeling experience less friction in the process.

Opening up the opportunity to a much larger audience. Independent developers, marketing teams, educators, product designers, architects, and manufacturers, among others, can all create or edit assets without having to hire dedicated 3D experts.

For those watching the market for creative AI, investors will recognise the pattern: the more generative tech transforms the creative process, the more valuable it becomes for businesses to own media for it rather than the tools themselves.

Meshy’s rapid growth reflects wider market demand

Founded in 2023 by CEO Ethan Hu, a Ph.D. graduate of MIT and the inventor of the Taichi GPU programming language, Meshy has scaled very rapidly.

The company points to more than 10 million users on the service who have made 100 million 3D models; a sign of how much AI-powered 3D design is being adopted by a new wave of non-specialist, niche technical users.

A lot of that scale has been about reducing friction.

Instead of having to be a Blender expert or know how to use other creation tools, users can just describe an object, upload a reference image, or explain what they want to see. The platform then generates the geometry, creates the texture, optimises it, and spits it out in standard file formats that can be dropped into tools like Unity and Unreal Engine, Maya, and Cinema 4D.

And for startup executives, this flexibility of the latest generation of creators could shrink the time it takes to get a new game to market dramatically.

Moving beyond text-to-3D

Early AI creativity tools were typically designed to do one thing. Make an image. Write a bit of code. Generate a model.

The bigger opportunity is to manage entire creative pipelines.

Meshy’s AI 3D generation powers include an AI 3D model generator that doesn’t require any modeling knowledge. You type in a simple text string, and get back fully-textured 3D models in seconds to a minute. Game props, collectibles, prototypes, product concepts, or assets ready to be detailed.

The platform also supports image-to-3D generation, AI texture generation, auto-rigging, animation, remeshing, and retopology, streamlining what are typically multiple discrete production processes into one process.

That’s currently reflected in the types of formats that can be exported: GLB, FBX, OBJ, STL, USDZ, 3MF, and Blender, making it possible to incorporate it into professional pipelines, rather than having it locked inside a proprietary ecosystem.

It’s an emerging trend in the field in general. As AI products have matured, developers have become less interested in products that are essentially creative tools and more interested in products that are completely integrated production environments.

Introducing conversational 3D creation

One of the most exciting innovations on the Meshy product roadmap is the release of the Meshy 3D Agent. It marks a divergence from text prompting to creating prompts as a conversation.

Instead of demanding that users write longer and more specific lines of prose, you can submit an image, a sketch, a half-baked idea, or even simple free-form spoken input. The Meshy system is prompted to pitch many detailed design concepts, to refine them iteratively with dialogue, and then to output a download pack of 3D geometry, all happening in the same conversation.

That’s reflective of what’s happening with AI software more broadly today, where conversation interfaces are increasingly taking the place of software menus.

This is good for creators, who don’t have to keep so much specialist language in their heads all of the time. And it’s good for companies, which can unlock the product development and content generation process for their teams without needing them to code up a trained model.

It also fits better with how teams tend to think and work about things today within their companies, by talking through cycles of iteration, and not through rigid software structures.

Enterprise expansion through Workspace 3.0

The most straightforward way to see the shift from consumer product to enterprise platform that Meshy has undergone is to look at Workspace 3.0. Unlike its predecessor, which provided capabilities exclusively for generating assets, this new iteration incorporates project-based workflows specifically designed for collaborative teams engaged in professional element production.

Image generation, modeling, animation, and printing are seamlessly organised into project spaces with shared asset libraries and API endpoints optimal for professional-grade production. This shift in emphasis is important because model fidelity is no longer the primary criterion for effective enterprise AI collaboration; instead, it is about workflow integration and compatibility.

Businesses don’t want one more standalone creative application. They want AI capabilities that plug into existing production tools. By delivering APIs with its collaborative workspace, Meshy is able to brand itself as a new kind of creative application: infrastructure.

The approach may echo best practices from enterprise software in multiple AI verticals.

Closing the loop between AI and manufacturing

Another praiseworthy aspect of Meshy’s approach is its hardware partnerships. Instead of contenting itself with creating digital assets, it offers a means to connect AI renders directly to the physical printing process.

In its FlashForge implementation, designs can be made directly within Flash Studio and then sent on over to compatible printers, along with auto-generation of texture-to-filament colours.

Meanwhile, the partnership with Bambu Lab makes the model output process even simpler by introducing one-click print functionality, while also maintaining a high compatibility rating with slicers for character and figure models. This is based on robust testing conducted in-house.

Probably more important is the Formlabs partnership. The integration connects AI-created models to professional SLA and SLS printing through a service bureau to deliver a workflow that brings a model from a text model to a professionally manufactured object with minimum human intervention.

To investors thinking through platform quality, these partners are more than just technical integration partners. They are a signal of growing enterprise legitimacy and a growing surface area for the company to tap into an expanding manufacturing and industrial ecosystem.

xTool and Snapmaker are building their own bespoke creative experiences on top of Meshy’s API, solidifying the company’s already strong product-market fit and pushing the reach of the platform further outside of its own interface.

Real-world adoption across industries

Tech products become more compelling once their utility is made tangible to a wider base of customers. The broad swath of customer types currently employing Meshy, ranging from game devs and educators to those in cultural heritage and tabletop design, shows the relevance of AI-generated 3D outside the initial scope of experimental intent.

A single dev, Jlemarchand, created the character art for Navy Island Confidential in Meshy, while a horror game creator, EvilKris, relied on Meshy 3D agents to iterate quickly on monster creation for Famished. The creator cyber_fox created assets for an English learning app in Meshy and reported that the workflow reduced a process that in Blender took around 300 hours to create the same single character silhouette to about 5 minutes per model.

In education, Teacher Noel found his way into VITA School’s 8th-grade game design class, devising classroom projects where total 3D novices created and imported Roblox-ready assets within a lesson. 85% described the platform as more intuitive than they had expected.

Organizations are using AI-made 3D in production pipelines as well. Annie, “Colorful Weaving Threads” exhibition R&D director, turned multi-day modeling projects into 5-minute asset creations, while tabletop creator Chad Hunter made highly-detailed, 10-inch 3D print-ready dioramas in under a day, without navigating the steep learning curve of traditional 3D modeling software.

The next phase of AI creative infrastructure

The broader AI sector has been making the transition from shiny new object to core operational tool. For the most part, people increasingly expect AI models to work with existing processes, to collaborate with teams, to integrate with systems, and to generate things that are directly usable in a commercial context. In this frenzied environment, the future of AI-designed 3D may owe less to big live demos and more to the development of full-stack workflows.

For the companies that can do it, from ideation to generation to optimisation to collaboration to fabrication to deployment in a few clicks within a connectivity-rich ecosystem, the enterprise rush is where the upside is and where the action is. Meshy is essentially going full stack and increasingly joining the ranks of the broader enterprise design stack.

Instead of focusing exclusively on generating models faster, the company is building infrastructure that supports creators throughout the entire lifecycle of digital and physical production.

For startups building products, manufacturers exploring AI-assisted design, and investors tracking the next generation of creative technology platforms, that evolution is likely to be just as significant as the underlying AI itself.

Total
0
Shares
Related Posts
Total
0
Share
tfn-logo-2-220x220-removebg-preview

Get daily funding news briefings in the tech world delivered right to your inbox.

Enter Your Email
join our newsletter. thank you
TFN Banner