28 April 2026 | Interaction | By Editor Robotics Business NEWS <editor@rbnpress.com>
In this insightful interview with Robotics Business News, Haitham Eletrabi outlines a bold vision for the future of tennis training—one driven by intelligent automation and seamless player experience. With innovations like the Partner V2 and Rover, Tennibot is creating a fully autonomous training loop that combines AI-powered drill generation, real-time feedback, and automated ball collection. By integrating advanced data analytics and wearable connectivity, the company aims to democratize elite-level training while enhancing efficiency for players, coaches, and clubs alike.
The long-term vision is a fully autonomous training environment, one where every aspect of practice, from drill design to performance feedback to ball retrieval, happens without friction. The player shows up, the system adapts, and they walk off the court better than when they walked on. The Partner V2 is the first real step toward that.
Tennis training has been fundamentally unchanged for decades. You book a lesson with a coach, hit balls, pick them up, repeat. The technology has been largely passive, a hopper on a stand that fires balls at a pre-programmed rate. What we set out to do with Partner V2 is make the machine an active participant in the training process. It's not just feeding balls; it's using 4K cameras to understand what's happening on the court and responding intelligently.
Think of the implications it has on tennis training. Now, coaches can use their time on the court more efficiently, so instead of just feeding balls on the other side and shouting instructions from the other side. They can be on the same side as the player, giving them real-time feedback on their footwork, angles, and ball contact, and making modifications to their style of play.
Same as the clubs and academies. Now, instead of losing court reservations on court time because there is no player to hit with that customer or that client, now they can rent out the sport machines or simply provide them as a service for their players to have a realistic and enjoyable hitting session on the court without having to worry about having someone who canceled or a coach that's not available.
Finally, players can experience a realistic playing session as if they are hitting with a coach and a teammate on demand, and they can create their own custom drills that target their specific weaknesses. This offers a level of training that was never available to them before.
Another element is democratizing elite-level training. A Division I college program or a top ATP academy has access to high-speed video analysis, custom drill regimens, and coaching staff around the clock. Most players don't. They're working on their game alone, two or three times a week, with limited feedback. AI levels that playing field. The Partner V2's custom drill engine can generate training sequences tailored to a player's skill level and goals. That's a capability that was previously locked behind expensive coaching relationships.
The hardest problem wasn't the AI, it was the translation layer between machine intelligence and human expectation.
Our AI can generate a drill sequence based on your skill profile in seconds. But if a 4.0 club player opens the app and sees a drill described in technical terms with adjustable parameters they've never encountered, we've failed. The sophistication has to be invisible. So we spent enormous engineering and UX time on that interface to make sure the app spoke the language of the player, not the language of the model.
The other design challenge was calibrating the AI's recommendations to feel earned, not arbitrary. Players have strong intuitions about their own game. If the system generates a drill that feels random or too easy or too hard, trust evaporates immediately. We built in mechanisms for players to rate and adjust drills, and those signals feed back into the model. Over time, it tunes to the individual. That feedback loop was technically straightforward, but getting the UX to feel natural, like a conversation, took multiple iterations.
The honest answer is that the most technically sophisticated thing we did was hide the technology well enough that players forget it's there.
Yes, and I want to be precise about what kind of disruption I think is actually happening, because it's nuanced.
We're not replacing coaches. A great coach provides something AI can't fully replicate right now: real-time biomechanical assessment, psychological coaching, the ability to read a player's emotional state during a difficult session, and the accumulated intuition of thousands of hours on court. That human dimension has enduring value.
What we are doing is changing the math of independent practice. Right now, the bottleneck for most players is access: to quality repetition, to structure, and to feedback when a coach isn't present. Apple Watch integration is a small example of a bigger idea: training data should be ambient. You shouldn't have to stop practice to change a drill type; for example, that context flows naturally into the training record without interrupting the flow of the session.
The Rover is where I think the shift is most concrete. The reason players don't train alone as much as they should is because solo practice is tedious. You hit forty balls, spend ten minutes picking them up, hit twenty more. The ratio of hitting to chasing is terrible. The Rover eliminates that. When you pair it with the Partner V2, you get a genuine closed loop: the machine feeds, you hit, the Rover collects, the Partner reloads. You can sustain high-quality, high-volume practice by yourself for hours. That's a structural change in how players can develop.
Where coaching models shift: coaches who embrace these tools will be able to run higher-quality sessions more efficiently. They can assign specific drills for between-session practice and actually know whether those drills were completed. The data creates accountability and continuity that hasn't existed before at this price point.
Directly and specifically. The Gen 1 customer base was vocal, and we listened hard.
The number one piece of feedback was portability. People loved the performance of the original Partner, but they wanted to get it in and out of a car, on and off a court, more easily. A 35-pound machine sounds heavy until you compare it to traditional ball machines in the 50–70 pound range, but we knew we could do better. The V2 came out 13.6% lighter and 14.4% smaller in volume. Those numbers might sound incremental, but players who use both machines notice the difference immediately.
The second major theme was vertical range. Tennis at the club and competitive level involves a lot of high balls — topspin exchanges where the contact point is above shoulder height, defensive positioning, high backhand drills. Our first generation had a more limited elevation range. The V2 increased vertical range by 37%, which dramatically expanded the drill possibilities and made the machine more useful for intermediate and advanced players who want to simulate realistic match conditions.
Three things: depth of AI integration, multi-sport versatility, and the closed training loop.
On AI: most competitors in this space are selling hardware with connectivity features bolted on. They'll call it "smart" because it has an app. What we built is different — 8 TOPS of on-device neural processing, 4K cameras with 3D object tracking, and a custom drill engine that generates and adapts training based on player input. The AI isn't a feature; it's the product architecture. That's a meaningful technical moat that took years to develop and isn't easy to replicate.
On versatility: the Partner V2 handles tennis, pickleball, and padel. The addressable market for racquet sports equipment is enormous and growing, pickleball in particular is seeing compounding growth year over year. We solve training across the racquet sports spectrum.
On the training loop: no one else offers what the Partner V2 and Rover create together. The combination of an AI ball machine and an autonomous ball collector is genuinely novel. It's not just a differentiator on a spec sheet — it's a fundamentally better solo practice experience. Players who use the bundle don't want to go back to manual pickup. That stickiness matters.
And then there's the intangible: we're a team of engineers who love racquet sports. My co-founder Lincoln Wang and I built this because we saw the problem firsthand. That perspective, building for players, not for spec sheets, shows up in every product decision.
This is one of the most important strategic decisions we've made.
Traditional ball machines start at $1,500 for a basic unit and climb past $10,000 for feature-rich machines from established brands. For most club players, that's simply out of reach. We launched the Partner V2 at $2,245 — significantly below what you'd pay for comparable or lesser technology from legacy manufacturers — and the Rover at the same price point. The bundle of both comes in at $3,995, which is less than many single-machine competitors charge.
That pricing reflects a deliberate choice to grow the market rather than extract margin from a narrow premium segment. More players training with better tools is good for the sport, and it's good for us. Happy customers at the club level become advocates. They tell their teammates, their coaches, their club managers. That word-of-mouth has been a significant driver of our growth.
We also back the price with substance: 60-day risk-free trial, three-year warranty, free shipping and free returns. We can make those offers because we're confident in the product. It removes the purchase anxiety for a customer making a considered investment.
The goal is that a serious club player, a collegiate athlete, and a teaching pro all look at this product and feel it was priced for them, not as a stripped-down consumer product, and not as an enterprise-only luxury.
Absolutely. The hardware is the entry point; the platform is where the long-term relationship with the player lives.
The app today gives players full drill control, custom drill creation and Apple Watch connectivity. That's a strong foundation, but it's early. The data layer we're building, session history, drill performance, training frequency, intensity over time, is genuinely valuable in a way that compounds. A player who trains with Tennibot for a year has a record of their development that no coaching notebook could replicate. The platform becomes more useful the longer you use it.
Where we see this evolving: deeper performance analytics, progress benchmarking against similar players, coach-facing dashboards that let instructors assign and monitor drills remotely, and eventually integration with broader sports health and performance ecosystems. Wearable connectivity is a gateway to understanding training load. Combining the physical output of a session with recovery data creates a more complete picture of player development.
We're also building for coaches and clubs, not just individual players. A club that deploys multiple Tennibot units across its courts will have a facility management layer we want to serve — utilization data, drill library management, and member progress tracking. That institutional market is substantial.
The vision is a platform that knows your game better than anyone except your best coach, and that grows smarter alongside you over your entire playing career.
