Post

@Rewkang so cool！

@Rewkang Between chatbots and self-driving is a massive grey area — and that's exactly where the real money gets made. Most people don't know how to position for that.

@Rewkang ☝️ Everything shifts from here

@Rewkang This is why Physical AI is such an interesting space. Software can scale instantly, but the real world requires data, hardware and distribution.

@Rewkang The future is now.

@Rewkang The shift from intelligence to supply is the real story Scaling hardware takes years, not compute

@Rewkang intelligence being the bottleneck for so long made this feel like a software problem but if the constraint is physical robots next year, that changes how fast any of this actually scales

@Rewkang Well explained

@Rewkang Hardware is the new software — by next year the real edge goes to whoever can actually ship millions of robots, not just smarter models

@Rewkang A robot needs a supply chain, not just an API, and hardware lead times are the one thing no model upgrade is going to fix next year.

@Rewkang Is there an ETF you recommend?

@Rewkang Intelligence remain bottleneck. Robotics needs massive amounts of high-quality, real-world, labeled data for training and continuous improvement. That’s exactly why should we bullish on the companies which are collecting this data. Like @useKled @mercor

@Rewkang 👏🏼

@Rewkang Great, Andrew, keep building!

@Rewkang Congratulations Andrew! What a wonderful next step.

@Rewkang the honest answer is that most investors still don't grasp the difference between ai-assisted decision-making and true autonomy in robotics.

@Rewkang the supply chain reality check we all needed

@Rewkang adoption will look more like enterprise software rollouts, 5-10 year slog. but yeah the "moment" thing is cope, robotics doesn't have a consumer hook like chat did. margins kill the hype cycle early 🐸

Great letter Andrew. You mentioned 2026 is largely about scaling data collection and processing for robotics. Do you see any genuine role for blockchain here, particularly around provenance, ownership and incentivisation of robot-generated/training data or basically no need at all.

@Rewkang cool！🙃🙃

@Rewkang The 100 meter sprint race held for robots in China a couple of days ago clocked in at 8.85 seconds， I’m truly amazed， the robotics industry is so promising， In the future, every industry will be inseparable from it it’s destined to be the trend of the future

@Rewkang Robots become the bottleneck next