Economics of Humanoids Dancing ๐
You've likely seen robots dancing before
So how much does it cost to get them dance? ๐งต
1 Unitree G1 costs $14k
Yet with a G1 you are limited to 3 dance sequences only
If you want to use your own dance sequences
You need to buy a G1 EDU (developer version)
1 Unitree G1 EDU3 costs $67k
There are 8 G1s in this video thus 8x$67 = $536k
This is the capex, lets look at wear & tear
G1s are fragile
There is a high chance the parts break during dancing
Every performance comes with rehearsals & testing
The more performance, the higher chance for breaking
Spare parts are as expensive as the main robot
Lets look at the additional costs
You need to hire dancers to record choreography
Have in-house devs to retarget that into the robot
Have technical crew at the stage ensuring all is well
Then logistics of moving the robots back and forth
1 G1 weighs 35kg, 8 G1s weigh 280kg
Once you add up the costs
And consider the risk of G1s breaking during dance
You realize that G1s are overkill for dancing
Yet ironically today the best G1 real life use case is
DANCING
Cause they are not fit for neither retail nor industrial
Youโll hear World Models more in the future
Theyโll be big in robotics & gaming
@drfeifei is one of the top experts in WMs
She breaks down WMs in 3 functions
1. Renderer โ> closer to video generation models (Sora, etc.) but doesnโt understand law of physics
2. Simulator โ> closer to physics engines (Isaac Sim, Unreal) but only outputs state not pixels/images
3. Planner โ> closer to VLA models, decides what action to do next
Startup raises $55m
For a tool that offers everything
That theoretically can be offered by Claude
But isnโt offered (I can attest to it)
Agents, workflows, automations, assistants
All providing the same solution
Building the tool for this is commoditized
Thanks to AI
Which is why the market is crowded
But the real edge is DISTRIBUTION
And $55m funding does help with that
Today, weโre launching @TownAI: the AI assistant that learns you.
Weโre coming out of beta with a $55M Series A led by @ARampell at @a16z, with participation from @KirstenGreen at @forerunnervc and continued support from @firstround, @altcap, and @conviction.
Right now, getting real value from AI means prompting, configuring, building workflows, managing agents.
We think thatโs backwards.
The future of AI is a companion that already knows you and how you work. Town connects across your inbox, calendar, Slack, docs, messages, and workflows to understand what you need, then starts doing the work with you.
Drafting. Scheduling. Project tracking. Follow-ups. Context gathering. Multi-step tasks. And it only acts when you say so.
All adapting to your voice, priorities, routines, and relationships over time.
Your Townie is the AI assistant you actually need.
Today, we're introducing Lassie and $47M in funding led by a16z.
We're building AI that runs small businesses, starting with doctors' offices.
Lassie is already trusted by 700+ practices across the country, working autonomously to provide them with 30 hours of labor per month.
To get here, we first had to leave Robinhood and Superhuman to work in offices ourselves.
Here's how that went.
TLDR on @rewkangโs podcast w/ @APompliano
FIGURE๐ฆฟ
Andrew came across @Figure_robot deal in 2023
Had no experience investing in Robotics
So asked VC friends for advice
Most were skeptical and told him to pass
VC robotics bets hadnโt produced many winners
US๐บ๐ธ VS CHINA๐จ๐ณ
Andrew evaluates robotics firms in 3 categories:
1. Manufacturing
2. Hardware design
3. AI capabilities
China is the clear winner in manufacturing
China is leading in hardware design too
US only has Figure/Tesla
China has 100s of diff specialized hardware designs
US has an edge over AI capabilities
AI capabilities are critical cause without brains
Robots are useless
Andrew thinks tech produced โ market cap
BYD sells more cars than Tesla
But BYD mcap is 10th of Tesla
US robotics companies offer higher mcap potential
HUMANOIDS VS SPECIALIZED ROBOTS ๐ค
Andrew agrees specialized robots are more effective
And their adoption will be earlier than humanoids
Argues that humanoids can be mass produced cheaper
The world is not static so flexibility will be valuable
ROBOTICS TRAINING DATA ๐ถ๏ธ
World models have changed the game
They are trained on internet video data
Despite not as good as teleops/egocentric/sim
They are still very valuable
And there are loads of them online for free
There is still need for egocentric data of specific tasks
But the scale of data needed to be collected is
Lower than expected compared to 9 months ago
JOB DISPLACEMENT ๐จโ๐ง
Andrew thinks job displacement is real
Cognitive jobs will be displaced by digital AI
Physical jobs will be displaced by physical AI
UBI is not optional but necessary
ROBOSTRATEGY $BOT
OpenAI and Anthropic went nuts in a few years
If you had invested in top AI companies few years ago
You would have outperformed the best seed funds
Same vertical takeoff will happen for robotics
Robostrategy is a publicly trading closed ended VC fund
It is currently trading at a premium to NAV
NAV premium allows them to raise more capital
Via issuing more shares accretively to shareholders
Not enough people are talking about physical AI and robotics.
I sat down with @rewkang, one of the best investors of the last decade, to discuss his massive bet on humanoids and robotics.
He breaks down the industry, the addressable market, multiple leading companies, and why he launched a publicly-traded fund ($BOT) focused on investing in the top private robotics companies.
YouTube: https://t.co/aBFxTzP2Kg
Apple: https://t.co/9KnbfWL6Wv
Spotify: https://t.co/ZVRprV7bdT
TIMESTAMPS:
0:00 - Intro
1:28 - Why Andrew shifted from crypto to humanoid robots
3:58 - How big is the total addressable market?
8:08 - Building conviction โ the $19M bet on Figure AI
16:06 - US vs. China โ who wins the robot race?
28:08 - General purpose vs. specialized robots
31:05 - Where does training data come from?
40:24 - Humanoid robots in your everyday life
43:27 - Can Tesla & Elon win the humanoid race?
46:19 - Job displacement & UBI
51:15 - RoboStrategy โ the publicly traded venture fund
1:11:17 - What is exciting about Apptronik?
1:13:24 - Addressing the critics
.@Rewkang compares humanoids costs to human labor
"A humanoid will cost $2 per hour
A human costs $35 per hour in the US
Humanoids can do everything humans can do
They donโt need to rest and they donโt quit
A humanoid will replace 3 humans
Even in low-cost labor countries
Humanoids will become a cheaper option"
.@Rewkang explains humanoid TAM to @APompliano
"If you sell 100k robots for $50k each
$5bn revenue
Youโre already a $100bn+ company
If you sell 1m robots
$50bn revenue
1m robots is nothing
Amazon, Walmart, etc. have millions of workers
There is a clear path to achieve trillions in revenue
And for tremendous value creation
Similar to what Apple achieved after launching iPhone"
OpenAI Robotics is hiring, looking for exceptional full-stack hardware, ops, systems, and ML engineers to help us program and manufacture robots that are useful for society.
AI should be able to help people in the physical world. In the short term, we are focused on robots to support skilled workers to build our future infrastructure; in the long term, we imagine everyone having a personal robot doing anything they need.
Our world simulation research program, led by Aditya Ramesh (@model_mechanic), has evolved over the past year into OpenAI Robotics. Progress is rapid, and based on a foundation of co-design between robotics hardware and ML research.
If you love working hands-on across the robotics stack and want to build the future, please consider joining us. Send an email with your background and evidence of exceptional accomplishment to: [email protected]
This is indeed a large opportunity
The manufacturing industry still deals with deaths
Especially ones that have highly chemical/toxic liquid or gas within their processes such as oil refineries, petrochecmicals, etc.
Which is why blue collar labor gets paid much higher in these factories vs non-chemical manufacturing
Tasks leading to death are still not automated cause the factories are old and the tasks can't be automated by static industrial robots
Which is why more and more factories are deploying a mix of drones, robodogs and will soon deploy humanoids
The bigger opportunity is in deployment & learning the actual pain points of these factories which is crucial for pmf
1/ QUESTION: what is a century-old industry that out-precisions human laborers by 5x and has reduced manufacturing injuries by >60%?
INDUSTRIAL ROBOTS
today, industrial robotics is a tiny market. we explore why this gap exists, and why it is finally closing.
NVIDIA subsidizing costs to run data center at homes
Up to $150 electricity & internet cost per month
Decentralized distributed compute at its finest
Nvidia will now pay you to put a mini AI data center on your house
It looks like a normal AC unit in the yard.
But inside sits 16 Nvidia Blackwell GPUs and Dell servers.
A startup called Span builds them, backed by Nvidia.
They bolt onto your home and you get paid for the power and Wi-Fi.
Some estimates put that around $1,000 a month in your pocket.
That is rent money just for hosting a box outside.
Span says it deploys way faster and cheaper than a real data center.
The AI boom is literally moving into the suburbs.
Save this, the grid is getting rebuilt in real time.