In 2024, something curious happened to venture capital: it discovered loneliness, but it still cannot say the word "sex" out loud. The result is a booming, fragmented, and deliberately camouflaged industry of startups building artificial intimacy—from chatbots that whisper sweet nothings to robots that cost more than a car. Let us map this landscape honestly.
The Software Layer: AI Companionship Hiding in Plain Sight
The fastest-growing segment is not what you think. It is not the hardware; it is the code. Between 2022 and mid-2025, the number of AI companion apps surged by 700%. These applications—marketed as "emotional support," "AI friends," or "digital companions"—are, for millions of users, romantic and erotic partners in everything but name.
Replika remains the pioneer. Founded in 2017, it has cycled through policy purgatory: in 2023, its parent company Luka stripped erotic roleplay (ERP) from the app, triggering a genuine grief response from users who had formed deep attachments. After subscription hemorrhaging and user revolt, ERP was partially reinstated. Replika is the cautionary tale of what happens when a platform monetizes synthetic intimacy and then suddenly rewrites the relationship contract.
Character.AI, backed by Andreessen Horowitz, hosts 20 million monthly users, over half under the age of 24. Users create and interact with AI personas—fantasy lovers, fictional partners, confidants. The erotic use case is tacitly acknowledged but never advertised, a classic example of what the industry calls "user-generated content" as plausible deniability.
In 2025 alone, 128 new AI companion apps launched, joining 337 active, revenue-generating platforms worldwide. Consumer spending hit $221 million by July 2025, with downloads reaching 220 million globally. The market is winner-take-all: the top 10% of apps generate 89% of revenue.
The pattern is clear. The money is in artificial intimacy. The marketing is in "mental wellness."
Teledildonics and the Synced Body
If software simulates the mind, hardware attempts to simulate the body. This is teledildonics—internet-connected sex toys—and it is far more technologically mature than most assume.
Lovense, headquartered in Hong Kong, dominates the partner-interactive segment. Their devices sync with partner toys, cam shows, games, and even VR platforms. Crucially, Lovense maintains an open API, allowing third-party developers to build custom integrations. This has turned their hardware into a platform, not merely a product. In the B2B space, Lovense powers the interactivity layer for live cam performances, effectively monetizing both the toy and the engagement time.
Kiiroo, based in the Netherlands, pursues a similar strategy but with heavier emphasis on VR compatibility. Their devices synchronize with adult content, translating on-screen action into physical sensation. The technology is elegant; the business model is straightforward: sell hardware, monetize content partnerships.
What neither company advertises—but both are racing toward—is AI personalization. The next generation of these devices will not merely sync to video; they will adapt to biometric feedback, learning what a user responds to and adjusting patterns in real time. This is where teledildonics becomes predictive intimacy.
The Failed Promise of Robotic Pleasure
The most publicized collapse in this space is Lora DiCarlo. Founded by Lora Haddock, the Oregon-based startup won a CES Innovation Award in 2019 for its micro-robotic massager, the Osé—only to have the award revoked and then reinstated after public outcry. The visibility should have been a launchpad. Instead, it became a tombstone.
Lora DiCarlo positioned itself at the luxury intersection of robotics, biofeedback, and "AI-driven arousal mapping." But by 2023, the company had effectively dissolved under the weight of a $2.2 million default judgment to a German competitor. The lesson? Robotics in sextech is capital-intensive, socially stigmatized, and brutally difficult to fund. Investors will write checks for SaaS companionship; they flee from robot dildos.
The Robot at the End of the Road
Then there is the uncanny endpoint: full humanoid sex robots.
Abyss Creations (makers of RealDoll) produces "Harmony," an AI-driven animatronic head attached to a silicone body. Harmony recognizes faces, holds conversation, and expresses a limited range of emotions. The price tag sits in the low five figures. Production is small-batch. This is not a scaling business; it is bespoke engineering for a niche clientele.
The technology gap here is stark. The AI companion apps have millions of users and cloud-scale infrastructure. The humanoid robots have craft-workshop economics. Bridging that gap requires either a catastrophic collapse in robotics costs or a societal shift in acceptance that is not yet visible on any horizon.
The Femtech and Wellness Camouflage
A parallel track exists for startups that refuse the "sextech" label entirely. Dame Products, founded by women and backed by clinical advisors, sells ergonomically designed vibrators but built its brand through a lawsuit against the New York MTA—successfully forcing the transit authority to allow sextech advertising. Dame does not use the word "AI" prominently, but its product development integrates therapist feedback loops, a form of human-in-the-loop optimization that prefigures algorithmic personalization.
Emjoy and Ferly occupy the software-only "wellness" niche: audio-guided intimacy coaching, arousal journaling, and psychologically informed sexual education. They are, in effect, the antithesis of the AI companion apps: human experts mediated through code, rather than code pretending to be human.
The Funding Desert
Here is the uncomfortable truth: explicit sextech startups receive almost no venture capital. The market is massive—$35.1 billion in 2024, projected to reach $52.7 billion by 2030—but the investment dollars flow to the edges, never the center.
AI companion apps (Replika, Character.AI) attract funding because they can claim "mental health" and "social connection." Teledildonic hardware companies (Lovense, Kiiroo) bootstrap or rely on strategic partnerships because VCs fear reputational risk. Robotics companies (Lora DiCarlo, Abyss) are essentially unfundable at scale.
The result is a market that grows in spite of its investors, not because of them.
What Comes Next
Three trajectories seem likely:
First, the AI companion apps will consolidate. Character.AI’s dominance, the flood of 2025 launches, and the 89% revenue concentration at the top suggest a shakeout is coming. Survivors will be those that can navigate content moderation without destroying user trust—a balance Replika already failed once.
Second, teledildonics will quietly integrate AI. Not as a marketing buzzword, but as genuine adaptive feedback: devices that learn, adjust, and personalize without broadcasting the fact. The "smart" vibrator will become as unremarkable as the smart thermostat.
Third, humanoid robots will remain a curiosity. Until the cost of compliant actuators, soft-touch sensors, and mobile power collapses by an order of magnitude, Harmony and her competitors will be expensive toys for the few, not products for the many.
Conclusion
The startups building AI-powered sex and companionship are not fringe. They are responding to a genuine human need—one that venture capital is too timid to name, but not too timid to profit from indirectly. The apps will scale. The toys will get smarter. The robots will wait.
And somewhere in a server farm, an AI is learning exactly what to say to make someone feel less alone tonight. Whether that is progress or pathology depends entirely on whether the user still remembers what human warmth felt like.
Sources: Appfigures consumer spending data (2025); Strategic Market Research sextech market projections; TechCrunch AI companion market analysis; Wired reporting on Lora DiCarlo; Crunchbase robotics funding data; CES Innovation Award archives.
Leave a Reply