The team behind Spotify's recommendation engine just raised $10 million to do the same thing for online shopping

Key points

• Malachyte raised $10 million in seed funding, co-led by Bessemer Venture Partners and Gradient, announced Thursday.
• The three co-founders previously built Vector AI, the system behind roughly 90% of Spotify's recommendations to its 800 million users.
• Malachyte's platform has been available to Shopify merchants since June 2026 through a native integration.
• The startup worked with more than 20 enterprise customers across travel, grocery and retail before focusing on e-commerce.

Three former Spotify engineers built the recommendation system that decides what song plays next for 800 million people. Now they want to do the same thing when you are shopping for work boots at midnight.

Sidd Motwani, Ian Anderson and Shivaditya Sinha spent years developing what Spotify calls Vector AI, a system that predicts what a listener wants to hear next rather than simply repeating what they already played. Motwani says it powers around nine in ten of every recommendation Spotify makes.

This week, the trio announced their new startup, Malachyte, had closed a $10 million seed round, co-led by venture firms Bessemer Venture Partners and Gradient, with Harpoon Ventures also participating. The money goes toward scaling the product and hiring commercial and product staff.

What problem is Malachyte actually solving?

Most online shops treat every visitor roughly the same. A first-time browser sees the same generic homepage as a loyal customer, and even repeat shoppers get suggestions based on old purchases rather than what they need right now.

Malachyte's fix is what the company calls "two-headed Vector AI," a personalisation engine, meaning software that tailors a shopping experience to each individual, that builds a real-time picture of two things at once: your general tastes and your immediate intent in this session.

Crucially, it starts working before you click anything.

"Our system starts forming before the first click, using the context available the moment the page loads," Motwani told TechCrunch AI. He gave a concrete example: search for "heavy-duty boot," then click two steel-toed options, and the system quietly moves work pants and gloves up the page while pushing dress shoes down. No account. No login. No history needed.

Every hover, scroll, search edit and add-to-cart sharpens that picture as you browse. The system even reads device and time-of-day signals: a phone visitor arriving at 11 p.m. from an email link is in a different frame of mind than the same person on a laptop at 10 a.m., and Malachyte treats them differently.

Who is paying for this, and who can use it?

The platform first went live in autumn 2025 with novelty retailer Fun.com. Since June 2026 it has been open to merchants selling through Shopify, the e-commerce platform used by millions of independent and mid-sized online stores. Larger retailers can connect through an API, a technical bridge that lets external software talk to Malachyte's system directly.

The company tested with more than 20 enterprise clients in travel, grocery and retail before settling on e-commerce as its focus.

Should online shoppers care?

If it works as described, yes: you would spend less time scrolling past irrelevant products. But remember, this data comes from following your every move on a site, including hovers and pauses. That is worth knowing, even if the experience feels smoother.

Stories like this often spotlight the founders who raised the money without asking how many of the 20-plus early test customers actually saw measurable sales gains. Motwani's Spotify credentials are real, but an 800-million-user music app and a mid-sized Shopify store are very different environments.

One honest takeaway: if you run a Shopify store, the native integration means you can trial this without a big technical project. Test it on real traffic before committing to a paid tier.