From Generic AI Tools to Owning the User Relationship
AI company acquisitions are increasingly focused on buying non-AI products with loyal users and deep domain expertise so that artificial intelligence can be embedded directly into specific consumer experiences instead of remaining a generic, stand‑alone tool. This shift from horizontal platforms to vertical, AI-infused services signals a new phase in generative AI expansion where product integration and real-world engagement matter more than model benchmarks. The strategic consolidation of AI around existing communities, workflows, and emotional touchpoints is reshaping how these technologies reach ordinary people, often without them even realizing that an AI layer is in the loop at all.
Seen in this light, the headline-grabbing deals around horoscopes and live shopping are not random; they are roadmaps. Midjourney buying astrology app Co-Star and live commerce platform Whatnot acquiring recommendation startup Shaped describe the same ambition: AI vendors want to control not only the underlying models but also the high-frequency, emotionally charged moments where those models can quietly become indispensable. This is strategic consolidation in AI, and it is happening faster than most users notice.
Midjourney + Co-Star: Generative AI Expansion Into Feelings, Not Prompts
Midjourney, known as one of the world’s leading generative AI companies with tools licensed by platforms like Meta, has acquired Co-Star, one of the most popular astrology apps. On paper, a horoscope app and a text-to-image lab look worlds apart. In practice, this is a move into personalization, ritual, and daily engagement—areas pure image generation has struggled to own. Co-Star brings a sticky user base that opens the app for intimate guidance, not for playing with prompts. That kind of habit is almost impossible to replicate from scratch.
Co-Star’s founder Banu Guler announced the acquisition and emphasized Midjourney’s focus on “beauty, imagination, and supporting human thriving with technology,” in part to counter public fear of AI after a Reuters/Ipsos poll found that 71% of Americans worry their job will be replaced by AI. Crucially for users uneasy about AI, Midjourney says Guler will retain “complete control” of Co-Star. At the same time, she becomes Midjourney’s Chief Design Officer, so the seamless design touches that made Co-Star a standout are expected to influence the Midjourney ecosystem. That is AI product integration via design leadership, not APIs.
A Mind–Body–Spirit Platform Play, Not a Quirky One-Off
Midjourney’s generative AI expansion has already been drifting into the physical and emotional realm. The company has explored health technology, including a full-body ultrasonic scanner, and floated “Midjourney Medical,” a concept for spas that combine saunas, hot tubs, cold plunges, and scanning tech into a single experience. Co-Star fits this arc: it covers the “spirit” part of a mind–body–spirit platform. Far from a novelty purchase, the astrology app is a ready-made emotional operating system that can sit on top of Midjourney’s visual and health experiments.
Strategically, Midjourney is not buying infrastructure—it already owns world-class generative models. It is buying a language for meaning and identity that resonates with young users who, as Co-Star’s base, are no strangers to tech anxiety. If Midjourney successfully weaves Co-Star-style personalization into its tools, the company stops being a background AI utility and starts becoming an everyday companion. That should worry competitors who still think in terms of horizontal APIs rather than vertical, emotionally anchored products.
Whatnot + Shaped: AI Product Integration for Real-Time Commerce
On the other side of the spectrum, Whatnot—an already AI-aware livestream shopping platform—has acquired Shaped, a machine learning company that specializes in real-time recommendation and search technology. Here, the logic is less mystical but more unforgiving: live commerce moves too fast for slow or generic AI. Whatnot processes more than 500,000 hours of live video and millions of interactions each week to refine recommendations. In that environment, mediocre personalization is not a UX flaw; it is lost sales.
Emmanuel Fuentes, Whatnot’s VP of Data and AI, says the company has spent six years cutting recommendation latency from about a day to minutes and expects the Shaped integration to push that closer to real time. Shaped has already built systems that mix customer data with large language models for personalized search and discovery. Rather than rebuilding such expertise internally, Whatnot is folding in Shaped’s founder Tullie Murrell and nearly a dozen engineers and researchers, who will form a new Applied AI Research group. This is strategic consolidation in AI: acquire the team that already solved the domain-specific problems and plug them directly into a high-growth engine that has surpassed 1 billion orders and raised USD 225 million (approx. RM1,035 million) in Series F funding at a valuation above USD 11 billion (approx. RM50,600 million).
The New AI Playbook: Buy Verticals, Not Just Models
Put Midjourney–Co-Star and Whatnot–Shaped side by side, and a pattern appears. Leading AI players are no longer content to sell generic capabilities; they are snapping up specialized domain expertise and ready-made user bases instead of building them from scratch. In one case, the target is an astrology app with cultural cachet. In the other, it is a focused ML shop that has already serviced commerce clients like Outdoorsy and QVC. The common thread is AI product integration into vertical experiences where latency, emotion, and habit matter far more than raw model size.
For users, the impact will be subtle but significant. Co-Star fans keep their familiar app and its founder in control, while Midjourney quietly gains new channels to experiment with personalized, design-led AI. Shoppers on Whatnot see more responsive, context-aware recommendations as auctions unfold in minutes or hours. The deeper message for the market is blunt: the era of standalone AI tools is giving way to embedded intelligence inside niche, high-value workflows. The next wave of strategic consolidation in AI will not be about who has the biggest model; it will be about who owns the most meaningful moments where those models are felt but rarely seen.
