By PYMNTS | July 17, 2026
The evolution of retail has consistently moved toward the reduction of friction. From the advent of the catalog to the ubiquity of one-click online ordering, the goal has always been the same: to collapse the distance between discovery and transaction. Today, that distance is being compressed to a matter of seconds through live commerce—a digital-first, modernized evolution of the television home shopping experience.
However, despite its meteoric rise in the East, the live shopping model has struggled to find a sustainable foothold in the North American market. As of mid-2026, roughly 68% of U.S. consumers report that they have never completed a purchase through social media, highlighting a profound disconnect between the format’s potential and its actual adoption. Industry experts increasingly believe that this gap is not a failure of content, but a failure of discovery.
To bridge this chasm, Whatnot, the $11 billion live-shopping powerhouse, announced on July 15, 2026, that it has acquired Shaped, a cutting-edge machine learning startup. This strategic move aims to revolutionize how platforms match shoppers with the right content in real-time, potentially unlocking the massive, untapped potential of the U.S. live commerce sector.
The Core Challenge: Why Discovery is the "Holy Grail" of Live Shopping
In the traditional eCommerce landscape, search and recommendation algorithms are relatively straightforward. Users browse stable catalogs, and algorithms suggest products based on historical data. Live shopping, however, operates on an entirely different temporal frequency.
Thousands of live streams run simultaneously across a platform like Whatnot. Inventory is not static; it changes by the minute, often driven by high-stakes auctions or limited-time drops. A viewer who arrives even five minutes late to a stream may miss the specific item that would have prompted a purchase, leading to immediate churn.
The fundamental problem is one of "temporal relevance." If a user logs on, they need to be funneled into a stream that matches their current intent, at the exact moment that intent is most likely to be converted into a transaction. Static recommendation systems—the backbone of traditional sites like Amazon or eBay—are simply too slow and too rigid to account for the fluid, chaotic, and lightning-fast nature of live video commerce.
Chronology of a Strategic Acquisition
The acquisition of Shaped by Whatnot represents a significant milestone in the maturation of the creator-led economy.
- 2021: Tullie Murrell, formerly of Meta’s recommendation systems team, founds Shaped. The company focuses on building infrastructure that allows platforms to implement "real-time" machine learning, enabling apps to learn from user behavior as it happens.
- 2025: Whatnot experiences a year of explosive growth, launching over 35 new product categories and reaching a valuation exceeding $11 billion. However, as the platform scales, the "discovery problem" becomes more acute.
- Early 2026: Whatnot aggressively expands its category catalog, adding more than 45 new niches in the first half of the year. With sellers surpassing 1 billion total orders, the sheer volume of live data becomes too complex for legacy algorithms to manage effectively.
- July 15, 2026: Whatnot officially acquires Shaped. Financial terms remain undisclosed, but the deal includes the transfer of Shaped’s team of engineers and researchers to Whatnot’s headquarters to form a new applied AI research group.
- July 17, 2026: Industry analysts and data firms, including NIQ, underscore the stark contrast between the $900 billion Chinese live-commerce market and the nascent, struggling U.S. sector, validating Whatnot’s pivot toward AI-driven discovery as a necessary market correction.
Supporting Data: The Global Gap and the AI Tailwind
The divergence between the Eastern and Western live shopping markets is staggering. According to NIQ, live shopping generated roughly $900 billion in sales in China in 2025—a figure that rivals the entire annual eCommerce market of the United States.
While Western retailers have experimented with live shopping, they have often treated it as a marketing gimmick rather than a core infrastructure play. The lack of personalization has meant that traffic is often sporadic and conversion rates remain low.
However, the tide is turning. Recent data from PYMNTS Intelligence and Visa Acceptance Solutions, titled “The Agentic Commerce Deep Dive: Payment Infrastructure and the Path to Trust for Merchants,” reveals a shift in consumer behavior. Nearly 48% of online shoppers are now using AI to research their most recent purchases. Furthermore, traffic referred to U.S. retail sites via AI agents grew by a staggering 393% year-over-year in the first quarter of 2026.
These metrics suggest that while live shopping in the U.S. has yet to reach critical mass, the underlying infrastructure of "AI-mediated shopping" is already being adopted by the mainstream. By integrating Shaped’s technology, Whatnot is positioning itself to be the primary destination for this AI-driven transition.
Official Responses and Technical Implications
The acquisition is more than just a talent grab; it is an integration of technical stacks. Emmanuel Fuentes, Vice President of Data and AI at Whatnot, emphasized the shift toward responsiveness in a statement following the acquisition.
“By combining Shaped’s technology with Whatnot’s existing systems, we can make recommendations faster, more responsive and more personalized,” Fuentes stated.
Shaped’s proprietary systems specialize in unifying disparate data points—user history, real-time clicks, dwell time, and inventory flux—into a single machine learning model. For a platform like Whatnot, this means the system will no longer just suggest a "category" of interest; it will predict the exact window of time in which a user is most likely to bid on a specific item within a live stream.
Tullie Murrell, the founder of Shaped, brings significant pedigree from his tenure at Meta, where he worked on the recommendation systems that drive engagement on platforms like Instagram and Facebook. Under his leadership, the new applied AI research group at Whatnot is expected to focus on "predictive discovery"—the ability to anticipate what a user wants before they even search for it, and then present it to them in a live, interactive format.
Implications: The Future of Mainstream Live Commerce
The move by Whatnot signals a broader trend: the "Platformization of AI." As retail moves away from static, page-based experiences toward dynamic, video-based experiences, the role of AI shifts from a backend helper to a frontend concierge.
1. The Death of the "Search Bar"
In the future of live commerce, the search bar may become obsolete. If an AI can accurately predict a user’s interest in real-time, the interface of the app will essentially "become" the product the user is looking for. Upon opening the app, a user will be greeted not with a search bar, but with a highly curated live stream that perfectly aligns with their current buying intent.
2. High-Frequency Inventory Management
Whatnot’s reliance on thousands of independent sellers creates a volatile inventory environment. By using AI to track the life cycle of every product being demonstrated—from the initial "show-and-tell" to the final gavel strike of an auction—Whatnot can effectively manage the supply-demand flow, ensuring that popular streams aren’t overwhelmed and niche products find their niche audiences.
3. Trust and Security
As noted in the PYMNTS Intelligence report, the path to trust for merchants is paved with secure, invisible, and AI-optimized payment infrastructure. As the barrier between "content" and "checkout" disappears, the payment experience must become equally seamless. The integration of Shaped’s technology will likely include optimizations that handle high-velocity transactions, ensuring that users can buy with a single tap without the fear of system latency during competitive bidding.
Conclusion: A New Era for Retail
The acquisition of Shaped is a calculated bet that the U.S. consumer is ready for a more immersive, AI-assisted shopping experience. While the cultural preference for traditional eCommerce in North America remains strong, the efficiency and excitement of the live-shopping format—when powered by the right recommendation engine—are undeniable.
If Whatnot succeeds in making live shopping as intuitive and frictionless as a simple search query, it may finally bridge the $900 billion gap between the Chinese and American markets. As AI continues to refine the way we discover products, the live stream is poised to move from the periphery of the retail landscape to its very center. For the U.S. consumer, the future of shopping isn’t just about finding a product; it’s about being in the right place at the right time—and thanks to the rapid evolution of AI, the platform is finally learning how to bring that place to the buyer.
