AI as a Full Marketplace Participant – Not Just a Backend Tool

01 /The Concept

Arcade offers people a way to "turn their ideas into real things" – with AI doing the heavy lifting. The idea is that anyone can visit the platform and use AI to design a physical object, even without any background in design or fabrication.

In principle the approach could apply to almost anything, but the platform has launched with a tight focus on jewelry. The core use case: someone who wants to create a unique piece for themselves or as a gift.

The key differentiator is that a design created with AI can actually become a real object – made by hand by vetted jewelers on the platform.

When a user is ready to order, they tap a button and the platform matches them with the craftsperson best suited to make that particular piece. Buyer and maker connect in an in-platform chat, discuss price and timeline, the payment goes through, and the maker gets to work.

Arcade earns a commission on completed sales. Its rules prohibit manufacturing any platform-designed piece outside the platform – only authorized makers working through Arcade can produce them, and only within the platform's transaction flow.

Makers, in turn, cannot reproduce a user's design for other customers without explicit permission from the original designer. They also cannot advertise or sell platform-commissioned work outside the platform, or accept direct orders that bypass the platform's transaction system.

Users can also design for others, not just for themselves. A designer who grants permission for their piece to be listed opens a royalty stream – up to 15% of each subsequent sale, depending on the creator's status on the platform.

Product cards then credit two parties: the user who conceived the design and the maker who brings it to life.

Earning that royalty status takes some effort. A user's designs need to generate at least three purchases totaling $500 or more, and they need at least 100 platform followers before payouts begin.

The easiest path to meeting those thresholds is inviting friends to sign up, follow the account, and buy a piece – a tidy built-in referral mechanism the platform has cleverly embedded into its monetization logic.

Arcade opened its beta publicly on September 20th, simultaneously announcing that it had closed several rounds of funding totaling $22 million from three private investors: LinkedIn co-founder Reid Hoffman, actor and investor Ashton Kutcher, and entrepreneur Brit Morin, founder of the women's lifestyle brand Brit + Co.

02 /Why It Matters

A [recent review](/review/dazhe-chast-ot-2-trillionov-jeto-ogromnye-dengi) covered Off/Script, a startup conceived in spring of the same year that raised $7 million in its first round in November. Its platform similarly uses AI to let users design clothing, footwear, bags, and accessories – and those designs can also become real products. The mechanism, however, is different.

A user publishes a design and works to collect 100 likes from other platform users. Once the threshold is reached, the platform runs a pre-order campaign, Kickstarter-style. If the minimum order volume is met, Off/Script sources a manufacturer, produces the run, and ships to backers. The platform takes a 30% commission.

Vice Versa, another startup that raised $1 million around the same time, is pursuing AI-generated fashion design but hasn't yet attached any manufacturing capability or business model to its platform.

Looking at these startups together, three trends emerge worth naming.

AI is expanding the creator pool. Historically, making things required craft skills – you had to know how to draw to design a garment, how to use design software to create jewelry specs, how to write code to build an app. Creative vision and technical execution lived in the same person – or not at all. AI changes that equation. It can now translate a vision into a design, a design into a technical brief, and a brief into a file ready for production. The pool of people who can meaningfully create things is about to get much, much larger.

AI could also dramatically expand the custom goods market. Custom physical goods already exist – tailored clothing, bespoke furniture – but bespoke typically requires the buyer to have a clear enough vision to communicate it, and enough execution skill to produce something worth making. AI lowers both thresholds. More people will want unique things made to their own spec, which could significantly grow the total market for custom production.

And then there's the most structurally interesting shift: AI as a genuine third party in marketplace transactions. Conventional marketplaces connect two sides: buyer and seller, or client and freelancer. In certain new marketplaces, AI isn't just a recommendation engine – it's a participant with its own role in the transaction.

Consider a hypothetical jewelry marketplace with AI designers trained on the styles of individual human designers. A buyer engages different AI designers to generate concepts, picks the one they want, and routes the order to a maker. The human designer whose style trained the AI earns a royalty. The AI designer retains the IP unless the buyer pays extra for exclusivity.

This three-party model would be impossible with human designers playing the AI role – the bottleneck is immediate. Replace the human designer with AI and the constraint disappears. New revenue streams become possible, new creator economics emerge, and the marketplace can scale in ways it couldn't before.

03 /Opportunities

AI will show up in marketplaces – that much is certain. The open question is in what form and under what business model.

For anyone who wants to get ahead of a trend that's clearly going to happen: the category to watch is AI-native marketplaces where AI plays a substantial, load-bearing role – not just a convenience layer.

For example: dramatically expanding the potential buyer pool, as Arcade does with jewelry, by making custom design genuinely accessible to people who've never held a design tool.

Or introducing AI as the third party in a transaction, generating new value streams for creators and new variety for buyers – as in the AI-designer scenario above.

This is a genuinely new territory. Not every approach has been tried, which means there's still room to invent something original rather than iterate on what already exists.

Alternatively – there's no shame in taking the Arcade or Off/Script model as a starting point, deploying it in a new category, and then watching closely to see where the market goes.

Any ideas on what's still unexplored? If so – share them. This is new enough territory that two perspectives genuinely beat one.