Whatnot is buying Shaped to sharpen one of the hardest parts of live commerce: helping buyers find the right item while the marketplace is changing in front of them.
The livestream shopping app announced Wednesday that it has acquired Shaped, a machine learning company that specializes in real-time recommendation and search systems. Whatnot says the move will strengthen discovery and personalization as the platform expands into more product categories and serves millions of buyers.
Why live shopping needs different recommendations
Recommendations are simpler when an online store has a stable catalog. A product page can be indexed, ranked, and revisited. Whatnot’s marketplace works differently.
On Whatnot, inventory is in motion. Auctions may end within minutes or continue for hours. Shows begin and end continuously. Buyer demand can shift during a stream, and buyer intent can change while a show is still underway.
That creates a recommendation problem that has to account for time, attention, inventory, and behavior all at once. If the system reacts too slowly, it may recommend an item, a show, or an opportunity after the moment has passed.
Emmanuel Fuentes, VP of Data and AI at Whatnot, described the challenge directly to TechCrunch: "That speed matters because live commerce is a uniquely hard recommendation problem. Inventory changes by the second, shows start and end continuously, and buyer intent shifts throughout a show."
What Shaped brings to Whatnot
Shaped was founded to help businesses build AI-powered recommendation systems. Its technology combines existing customer data with large language models and machine learning to create personalized search and discovery experiences.
Before the acquisition, Shaped’s customer roster included Outdoorsy and QVC. For Whatnot, the appeal is clear: Shaped’s work sits directly in the area Whatnot says it needs to improve as live commerce becomes more complex.
Fuentes told TechCrunch that combining Shaped’s technology with Whatnot’s existing systems can make recommendations faster, more responsive, and more personalized. The goal is not simply to show buyers more products. It is to improve the timing and relevance of what buyers see while live shopping activity is still unfolding.
Whatnot has already spent the last six years reducing recommendation latency from roughly a day to just minutes. Integrating Shaped’s technology is expected to move those recommendations even closer to real time.
The team joining the company
The acquisition also brings Shaped’s people into Whatnot. Shaped founder and CEO Tullie Murrell will join the company, along with nearly a dozen engineers and AI researchers.
Murrell will lead Whatnot’s newly formed Applied AI Research group. The source also notes that Murrell worked at Meta before launching Shaped.
That matters because Whatnot is not treating the acquisition as only a technology purchase. It is also adding a dedicated group focused on applied AI research, which points to recommendation, search, and discovery as ongoing priorities for the business.
Why the timing matters
The acquisition comes during a period of significant growth for Whatnot. Launched in 2019, the company recently revealed that sellers have surpassed 1 billion orders.
Last year, Whatnot raised $225 million in Series F funding, giving the company a valuation of more than $11 billion after adding 20 million buyers over the past year.
The marketplace has also broadened beyond its earlier focus. Whatnot launched more than 35 new categories last year, including art, golf, and vinyl. During the first half of 2026, it added more than 45 additional categories, with new subcategories continuing to roll out each month.
More categories can create more opportunity for sellers and buyers, but they also make discovery harder. A buyer who once came for a narrow set of products may now face a much wider marketplace. Recommendation systems become more important when the platform has to connect users with relevant shows, auctions, and products across an expanding set of interests.
A broader AI race in resale
Whatnot’s move also fits into a larger competitive push. The source notes that resale giants such as eBay and Poshmark are racing to integrate AI throughout their platforms.
For Whatnot, AI is being tied to the specific demands of live commerce. The company says its systems process more than 500,000 hours of live video and millions of real-time interactions every week, using that data to continuously improve recommendations.
That scale helps explain why the Shaped acquisition is focused on speed as much as personalization. In live shopping, relevance can expire quickly. The stronger the system is at interpreting what is happening now, the more useful it can be for buyers navigating a marketplace that changes by the second.