Last reviewed · 4 July 2026

Where It’s Heading

So far, digital intimacy has mostly lived on screens — the texts, video calls, dating apps, and AI companions described elsewhere on this site. The next chapter is about bodies. Companies are now trying to give digital closeness a physical form: robots that share your home, devices that transmit touch across an ocean, and virtual spaces where presence feels almost real. Some of this is shipping today. Some of it has been “five years away” for twenty years.

Companionship is getting a body — just not the one you’d expect

The obvious mental image is a humanoid robot partner. The reality arriving first is smaller and cuter. Pet-like and tabletop companion robots — Casio’s furry Moflin, the huggable LOVOT, the ElliQ desktop assistant for older adults — are multiplying, and a fresh wave of small robots marketed not as assistants but as friends was a headline trend at CES 2026. New York State’s aging agency reports that older adults given ElliQ interact with it more than 30 times a day and claims a 95% reduction in loneliness measures — though that figure comes from the agency and the vendor, not an independent trial. A small peer-reviewed study of LOVOT with older adults reported comfort, connection, and a sense of purpose. These devices deliberately sidestep the hardest robotics problems by not trying to look or move like a person.

At the frankly adult end of the market, physical AI intimacy is already commercially real. Chinese manufacturer WMDoll embeds conversational AI (built on open-source language models) into its sex dolls for an extra $100–$200, and expected a 30% jump in sales in 2025. The dolls don’t move on their own; the intimacy upgrade is conversation and memory, not robotics.

The humanoid question

True humanoid robots did cross a threshold: the robotics firm 1X says it sold out its first-year run of more than 10,000 NEO home robots in five days at $20,000 each — a company-supplied figure, but if accurate, the first mass-market humanoid meant to live inside homes, marketed for chores and companionship. One caveat drew wide attention at launch: NEO initially relies partly on remote human operators to handle unfamiliar tasks — a real limitation for both capability and privacy. Meanwhile Realbotix, the company descended from sex-doll maker RealDoll, has repositioned itself around “embodied AI” companionship, selling humanoids from $20,000 to $125,000 — and now explicitly states its robots are not built for physical intimacy. Even the industry’s most sexualized lineage sees the near-term value in social, not physical, connection.

Forecasters see a big market but a slow consumer arrival. Goldman Sachs projects $38 billion and 1.4 million humanoid units by 2035; Morgan Stanley projects $5 trillion by 2050 but expects adoption to stay “relatively slow until the mid-2030s”, with factories getting robots years before living rooms do.

The strongest skeptical voice is robotics pioneer Rodney Brooks, co-founder of iRobot, who argues that humanoids achieving human-level dexterity “any time within decades” is “pure fantasy thinking” — a human hand has roughly 17,000 touch receptors that current robots simply lack, and today’s machines can neither touch safely nor reliably avoid falling. Safe, gentle physical contact is exactly what intimate companionship requires, making it the hardest robotics application imaginable. History backs the caution — when Rolling Stone profiled the industry’s flagship “robotic love doll” in 2017, the state of the art was an animatronic talking head attached to a motionless doll body, and the walking, feeling android promised back then has still not shipped. And demand is uncertain too: the share of Americans who would consider sex with a robot rose from 16% in 2017 to 22% in 2021 — growing, but a minority, with men consistently more open than women.

Touch and presence without robots

Two quieter technologies may matter more in the next five years. The first is haptics — hardware that transmits physical sensation over the internet. Remote-touch devices already exist, from app-controlled intimate toys to bracelets that relay a squeeze, and the haptic-device market is projected to roughly triple from $4.8 billion in 2023 to $13.7 billion by 2030. Researchers note mediated touch creates genuine feelings of connection but raises novel consent questions: what does agreeing to a touch mean when it can be transmitted, recorded, or replayed?

The second is virtual presence. Social VR already hosts real relationships — VRChat set a record of more than 66,000 simultaneous users on Steam around New Year 2025, a count that doesn’t include players on standalone headsets — and researchers have documented “phantom touch,” where users physically feel avatar contact with no hardware involved at all, suggesting the brain can supply some of the physicality that machines can’t. Photorealistic telepresence is closer than robots, too: Meta’s Codec Avatars — lifelike digital doubles — can now run on a consumer headset, though creating one still requires a studio rig of more than a hundred cameras, and ordinary video calls are the expected first mainstream use. For the long-distance couples and transnational families described in Digital Intimacy in Everyday Life, feeling “in the room” together may arrive well before any robot does.

The forecast

Intimacy scales in software first. AI companion apps reached about 220 million cumulative downloads and a $120 million annual revenue pace in 2025, while robot volumes are measured in thousands. The most likely 5–15 year path is that existing AI relationships gradually acquire bodies — cheap dolls, desktop pets, haptic gadgets, photoreal avatars — rather than robots creating new demand from scratch.

The canonical long-range bet is AI researcher David Levy’s 2007 prediction that human-robot marriage would be legal by 2050. Nearly twenty years on, the software half of his forecast — emotional bonds with AI — is arriving far faster than the hardware half. Regulation has begun moving alongside the technology, too: in late 2025 California passed the first US law aimed specifically at companion chatbots (see FAQ & Further Reading for the safety debate).

What no one knows: whether people will want humanlike robot partners at scale, whether the hard physical problems yield in ten years or forty, and what living with embodied AI does to human relationships. Digital intimacy is entering the physical world this decade, in modest and mostly non-humanoid forms. Humanlike robotic companions, if they come, belong to the 2035–2050 window. Anyone quoting a confident date is guessing.