Vad | Exylos Robotics

Turning human intent into validated robot-ready manipulation skills for complex high-variability tasks. @exylos_ai

[Vad | Exylos Robotics @vadcrypto]

Excited to showcase EXYLOS product and share our vision, see you in 2 hours

[CyreneAI @CyreneAI]

Robotics is where AI was before data.

- No internet-scale dataset.
- No foundation layer.
- No shortcut.

@exylos_ai is building the missing infrastructure - turning human actions into verified, robot-ready training data for the next generation of intelligent machines.

The future won't be trained on prompts.

It'll be trained on physical experience.

Join us at 12:30 PM UTC as we showcase Exylos AI.

[Vad | Exylos Robotics @vadcrypto]

Next level red light THERAPY. #dontdie

[Bryan Johnson @bryan_johnson]

never waste a good crisis

[Virtuals Protocol @virtuals_io]

https://t.co/J7vKHqIcA8

[Vad | Exylos Robotics @vadcrypto]

EXYLOS approach combines quality signal with exponential scaling through contributor network multiplied in simulation.

[Exylos @exylos_ai]

Billions are flowing into physical AI, but the robots still can't tidy a room. There isn't enough high-quality training data to train them on.

Three approaches dominate for closing this gap, each with a trade-off:
→ Real-world teleoperation: high quality, hard to scale.
→ Pure simulation: cheap and clean, but often breaks in reality.
→ Egocentric video: abundant footage, slow cleanup.

@exylos_ai takes a fourth route: teleop-in-sim powered by a global contributor network.

Trainers play robot tasks inside EXYLOS software. Curators verify, simulation multiplies accepted signal. $EXY coordinates access, rewards, reputation, and quality.

Output is a high-quality Skill Pack robot teams can train on directly.
27k+ Hugging Face downloads on public samples show the demand is real.

[Vad | Exylos Robotics @vadcrypto]

@rohankanojia_in Exactly right. Whatever the architecture of the model - the data is the limiting factor for its performance. And data hunger will only grow bigger as models progress.

[Vad | Exylos Robotics @vadcrypto]

Just shared detailed breakdown on how and why robot ready training signal will be produced inside EXYLOS, and how the value flows over time build stronger and more qualified contributor network.

[Exylos @exylos_ai]

https://t.co/BXc7Ys1C6V

[Vad | Exylos Robotics @vadcrypto]

@shivst3r The signal-to-noise ratio is brutal, but you're completely right. It takes a special kind of resilience to keep shipping real tech when the timeline only cares about the gamblers.

[EllioTrades @elliotrades]

Andrew Kang turned a few early stage robotics investments into a $500 Million publicly traded empire @RoboStrategy

He is widely considered THE leading investor in robotics

@Rewkang believes the ChatGPT moment for Robotics is happening RIGHT NOW and it will be BIGGER THAN BITCOIN

This interview is a masterclass on Robotics investing in 2026

1:18 – Why robotics today feels like crypto in 2015
4:23 – Is robotics a bigger opportunity than Bitcoin?
10:30 – Robotics 101: players, problems & bottlenecks
16:40 – Who's leading the race: Tesla vs Figure
23:10 – China's 100+ robotics companies & the real investment risk
25:28 – US vs China: who's actually ahead?
33:40 – How many human jobs will be replaced?
38:00 – The "ChatGPT moment" for robotics
43:55 – UBI & the new social contract
46:30 – Why Andrew built Robo Strategy
51:20 – Lessons from Bitmine, MicroStrategy & Saylor
1:07:50 – The Boston Dynamics question
1:12:09 – His most contrarian bet: avoid model-only companies
1:19:55 – Tesla or Figure?
1:23:26 – Figure's valuation in 5 years: bear, base & bull
1:29:20 – Final words of wisdom

[NVIDIA Robotics @NVIDIARobotics]

New NVIDIA physical AI agent-ready skills are changing how robotics researchers work. 🤖

With NVIDIA robotics skills, researchers can automate the most common development steps from scene preparation to simulation and learning with Omniverse libraries, Isaac frameworks, and physical AI open datasets.

Specialized skills like Isaac Mobility extend that even further.

Learn more ➡️ https://t.co/WwGv2XYpEi

[Exylos @exylos_ai]

The @FrameworkVC putting their thesis on the table publicly: the data layer for physical AI is the position to own.

Exylos is built on the same thesis with a different structure.
The supply comes from a coordinated network, not a centralized payroll, and the data is open: sample datasets are already live on Hugging Face for anyone to inspect.

The Framework call validates where the value sits. We're shipping the onchain version of it.

[bigwil @bigwil]

Virtuals Weekly Episode 105 Recap!

➥ We were joind by @NodarJ who explained his move to @virtuals_io and his plans for UniClaw. Sadly, shortly after the show, Nodar rugged the token, sold ~$200 of it and spun up another token in its place. Pretty disappointing all things considered.

➥ We dug deeper into the @reppo move to @BaselineMarkets and scratched the surface of what it could mean for the Virtuals ecosystem more broadly. Those paying attention might have had a small glimpse into the future here!

➥ We continue to push the probable importance of having runs on the board with a Virtuals agent. @base MCP is also on our radar and being early is the approach here.

➥ It looks like the Degen Arena is dying out. We want to see these new initiatives succeed and are hopeful the team can get pushing it forward.

➥ We previewed the latest A tier launch in @exylos_ai - It has been delayed one week but definitely one to take a look at. The launchpad remains quiet but it's good to have a solid launch to focus on.

↛. You can listen back to the recording for episode 105, and set your reminders for episode 106, in the comments.

[Exylos @exylos_ai]

Something is close, and it changes how you'll see all of this.

Better you hear it before the launch, not after. A net win for the community

So Exylos TGE is now June 11.

The next seven days are going to be worth following.

[Exylos @exylos_ai]

The newly announced NVIDIA https://t.co/wnQONXTKcG makes this point sharper: it does not make robot data less important. It makes raw video less defensible.

If world/action models can connect pixels, future states, and trajectories, the bottleneck moves to verified signal: task intent, action compatibility, failure boundaries, recovery behavior, quality settlement, and evaluation.

Cosmos 3 pushes raw video-to-action imagination into commodity infrastructure.

The scarce asset is no longer just more footage.

The value moves to deciding what deserves to become trusted robot training data.

[Exylos @exylos_ai]

Robotics companies are paying real money for training data right now — not in five years, not when humanoids ship at scale. The question is who's positioned to supply it.
Recap of our X Space on the data layer behind physical AI.
With @umeirzz from @virtuals_io
@rgvrmdya from @reppo
and our CEO @vadcrypto

[Vad | Exylos Robotics @vadcrypto]

@exylos_ai Gm guys, hope you enjoyed the space and we managed to answer your questions!

[Vad | Exylos Robotics @vadcrypto]

This Monday 1PM UTC we’re hosting X Space With @umeirzz @rgvrmdya and @exylos_ai:
🔥 Can token networks help produce, validate and massively scale the exact data robots need to learn?

Would be awesome to see the Virgen community there! Bring your questions💪

[Vad | Exylos Robotics @vadcrypto]

👋 Hey Virgen community! First post here - excited to join this awesome community Big thanks to @100xDarren
and everyone building in the Virtuals ecosystem! This is Vad from @exylos_ai
- building a Physical AI project inside the Virtuals ecosystem.

[Exylos @exylos_ai]

DePIN networks coordinate bandwidth, storage and compute.
Now the question is whether the same market logic can coordinate something harder: robot-ready training signal for Physical AI.

Can token networks help produce, validate and scale the data robots actually need to learn?

Join us for an X Space on June 1 at 1PM UTC to explore the topic with @umeirzz @rgvrmdya @vadcrypto

https://t.co/ZH0AQVQ5dn

[Vad | Exylos Robotics @vadcrypto]

Mark the date!

[Exylos @exylos_ai]

$EXY launches June 4 on @virtuals_io
@exylos_ai is a Skill Factory for Physical AI.
Robots keep getting smarter, but they still don't have enough data teaching them how to interact with the real world.
That's the data we make, and we ship it as ready-to-train datasets for robotics teams.

[Exylos @exylos_ai]

The robotics data space is 90% beautiful landing pages and 10% actual datasets.

Meanwhile, we just passed 26,000 downloads on @huggingface

The #PhysicalAI bottleneck isn't going to be solved by more basic sim data or raw egocentric video.
The real demand is for structured, enriched, ready-to-train datasets.
That is the only way to build reliable manipulation models.

Links to our datasets are in the replies below. 👇

@LeRobotHF #PhysicalAI