Inside BOT: What RoboStrategy is buying in the emerging $10 trillion robot economy

How do you invest in robots?

RoboStrategy is a non-diversified, closed-end management investment company registered under the Investment Company Act of 1940, trading on the Nasdaq as BOT. As of August 31, 2026 it reported $276.7 million in net assets and a NAV of $11.35 per share.

The pitch is access: private, pre-IPO robotics and physical-AI companies that public market investors normally can't reach, wrapped in a listed structure. Between April 30 and July 31, the fund reported NAV up 85% and NAV per share up 54%. Since its IPO it has deployed roughly $124 million across six companies.

I had a chance to spend some time with Kang and ask him what he's investing in.

TLDR of our convo: what RoboStrategy is buying and what Andrew Kang believes about humanoid robots

No time? Here's the too-long-didn't-read version.

Or, too long didn't watch.

• Kang's three-name U.S. conviction list: Tesla Optimus, Figure AI, Apptronik. Nothing else in America clears Kang's bar. (Yet?)
• RoboStrategy owns two of the three. Apptronik is 14.2% of net assets, Figure 13.8%. Tesla is zero ... Kang says so on the record.
• The largest single position isn't a humanoid company. Standard Bots, industrial automation, is 31.4% of a $276.7M book.
• Picks-and-shovels is not automatically a winning trade. Component suppliers commoditize. The exception Kang is underwriting is the actuator and motor layer.
• TAM: "tens of trillions." Bottom-up unit economics: $5,000–$7,000 per year for humanoid labor.
• Timeline: millions of units 2028–2029, humanoids common on streets and in homes by 2030.
• Supply, not intelligence, is the near-term constraint.

Disclosed portfolio: what RoboStrategy holds

Here's what we know publicly about what RoboStrategy holds:

Remaining positions (~8% combined): Path Robotics, Project Prometheus, Eccentric Machines, REK, GMI Cloud, Coco Robotics, Nox Metals, Endiatx, Allonic, Purple Rhombus.

In other words, if you buy BOT as pure humanoid robot exposure, you're making a mistake. What you're actually getting is roughly 28% humanoid pure-plays and about a third industrial arms. 

That's actually smart: robotics is big, robotics is wide, and not every robot working in a factory or a home will be a humanoid robot. But you should be aware of that, if you're looking for just humanoids.

Who will win in humanoid robots?

Kang has a pretty clear opinion here: not everyone.

In fact, he names just 3 key players in the U.S.:

"That is Tesla Optimus, that is Figure AI, that's Apptronik. Those are really the three key companies that we have strong conviction are going to deliver a real product to the market where a humanoid can manipulate things around us, can understand the physical world, and that they will grow to become really massive enterprises."

Why?

Kang sees talent density (and talent shortages) as the key gating factor, not capital.

Speaking as a robot company CEO, he says:

"I need an expert in robot learning, I need an expert in robot behaviors, I need an expert in computer vision, in hand engineering, in battery engineering, and fleet management systems. And then there are only a few hundred people, maybe, in all these categories that are really, really competent. And then putting them together in one company and raising billions of dollars of capital, that's not very easy."

Why this matters:

A talent-constrained market concentrates. It doesn't distribute returns across a basket. If Kang is right, index-style exposure to "humanoid robotics" underperforms badly against the two or three names that clear the bar.

The risk, obviously, is that when you pre-pick winners ... you're not always right.

On Tesla and Optimus: Kang conceded the criticism and kept the position

Andrew Kang is bullish on Tesla, despite the fact that we haven't seen amazing things from Optimus yet. When I pressed him on the gap between Optimus's public showing and Figure's, he didn't defend Tesla's current state:

"Actually, you're exactly right. At this point in time, Figure is ahead of Tesla on the AI side. You can see it."

His attribution is CEO bandwidth: Elon Musk has been dealing with SpaceX's listing, the xAI merger, the Cursor acquisition ... but thinks that when Musk really starts paying attention, Tesla will turn it around.

And, he says, founder track record matters:

"But when he turns his attention, I think he gets a lot of stuff done, and can get as much stuff done as the combination of maybe five other CEOs put together."

His hedge is real, and investors should note it: "Is he going to be able to commercialize humanoid robots at scale? I think yes. Is it going to be number one? I'm not sure."

On Agility and the SPAC signal

He's not bullish on Agility Robotics. Agility is going public via a $2.5 billion SPAC merger. Against Unitree's $66 billion peak, that looks cheap to me. 

But Kang declined the comparison:

"I'm not sure if I would agree with them being cheap as a valuation."

His thinking:

"There's a lot of capital in the private capital markets that is now only just beginning to kind of turn on to robotics. And if your technology is really improving, if you have a good pace of development, if you are really winning core contracts and people believe that you're going to solve general robotics, then there's no lack of funding for it. So if you have to kind of go public via a SPAC, maybe there's a reason."

And the analogy that frames his whole approach:

"You could have looked at OpenAI or Anthropic two years ago and said, hey, look at these other companies that are valued at 10% of the value of OpenAI or Anthropic... Sometimes the winners keep winning, and there are reasons why they are valued highly."

Picks and shovels: the moat isn't where you think it is

It's relatively popular to make a pick-and-shovels bet on an emerging market. The thinking is simple: you can't necessarily pick the winners; but you can pick a broad swathe of suppliers.

RoboStrategy is a little bit pickier. According to Kang:

"It's not always the case that something in the supply chain for a big product is actually a winner. You look at Apple. Most of the component suppliers are heavily commoditized... they're just taking razor-thin margins, and they're not big companies."

The counterexample of course, he adds, is Chinese battery makers: unique, hard to replicate, larger than the automakers they supply in some cases.

Where he's putting money: actuators and motors. 

Kang's case is that motor design has been optimized for continuous single-direction rotation, not the constant reversals humanoids demand, and that the actuator menu — harmonic, planetary, cycloidal — is all trade-offs with no clean answer. Harmonics deliver torque density, gear ratio and precision, but demand tolerances that make manufacturing scale brutal and cost prohibitive at whole-robot scale.

Case in point: RoboStrategy led the pre-seed in Eccentric Machines, developing an AI-native actuator the firm calls Sentor, designed to handle unexpected loads at the joint rather than routing everything back to a central controller. Kang's framing in our conversation: an actuator that "can sense the world around us or any forces that happen to the actuator and be able to respond to it without needing to go back to the brain of the human. It has its own reflexes."

He flagged it as early stage. Treat it accordingly.

Of course, there's a complicating factor for the whole picks-and-shovels thesis ...and it's called vertical integration. 1X builds its own actuators and its own tendon-driven hands. (Figure has publicly called tendon systems a dead end.) 

When the leading buyers build their own components, the supplier TAM is smaller than the robot TAM implies.

The FCC covered list

In August the FCC added foreign-produced advanced robotic devices to its Covered List, restricting imports of new models. 

Kang's expectation is that the list softens over time.

"I think the government will approach it the same way that they did with the drone ban that also came from the FCC, where they started out really restrictive, really broad-reaching, and then they listened to the domestic industry... I think it won't be a huge burden to the US from that sense. I think there will be accommodations that will be made."

The portfolio implication is obvious: 

Restrictions on Chinese research platforms push U.S. labs toward U.S. hardware. Dexmate — 9.2% of the fund, described in RoboStrategy's August shareholder letter as the leading U.S.-based seller of humanoids with hundreds of robots shipped — is the direct beneficiary. 

RoboStrategy led both seed rounds, at $123 million and $216 million post-money.

The developer tier is a separate market

On Nori Robotics' $1,688 A3, assembled in San Francisco, Kang was positive but categorical about segmentation:

"I think the market is going to segment, where you're going to have companies that develop for the research market. I would think about that as like maybe the Raspberry Pi-type companies, which is a market, but it's very different from the companies that build computing for the AI world for everyday use."

Worth noting: his correction to the "China is cheap" assumption: Unitree's EDU model, the one researchers actually buy, runs into the tens of thousands of dollars, and serious Chinese industrial units run six figures.

His credit to Unitree is the more interesting investor point:

"Without them, actually, a lot of the US companies wouldn't be as ahead as they are ... they enabled a lot of research to be done by universities, by people that are publishing their results and innovations publicly."

China watchlist

Kang is tracking Galbot (which he singled out as the most frontier team on robot learning, citing its viral tennis demo), Agibot, Engine AI, LimX, plus automakers and phone makers entering the field, including BYD and Honor. 

His own count of the Chinese field: "100-plus."

(Yeah, we see a few more than that ... 194 so far.)

For Europe, RobotStrategy hasn't place a humanoid bet yet. He's watching Uma (out of Hugging Face) and Genesis AI, which he likes for a simulation-first approach paired with co-designed hardware and software.

Market sizing and timeline

"It's clearly going to be tens of trillions of dollars. And I don't say that as a joke."

The bottom-up:

• One humanoid, amortized over seven years, running three shifts ≈ three humans' output
• $5,000–$7,000 per year for humanoid labor
• At a $50,000 unit price, one million units = $50 billion in revenue
• At 10–100 million units per year: "multiple trillions, tens of trillions of dollars of revenue, which would mean tens of trillions of dollars of market cap"

The timeline: millions of units as of 2028–2029, accelerating from 2030. Humanoids are common on streets and in homes by 2030, with earlier concentration in China, San Francisco, and among affluent early buyers.

The adoption shape: Kang expects a threshold effect, not a ramp, "like GPUs and other applications that, before AI really reached that threshold of, this is really useful, the demand was small. And then when it just crossed that threshold, it just jumps up instantly. It gaps up."

The constraint:

"I can snap my fingers and you can generate a million instances of an AI chatbot, but for robots, we've got to build out the infrastructure, which is happening right now. And then you've got to actually build the robots, and that'll take a little bit of time."

Things to watch for in RoboStrategy

1. Whether the three-name list holds. The fastest way to check Kang's thesis is a fourth U.S. company shipping at Figure's reliability.
2. Unitree's post-IPO trade. Down ~53% from its first-day high as of early September, still over 3x the IPO price. It's the category's public price discovery mechanism.
3. FCC accommodations. If the FCC restrictions don't soften, domestic component costs and research access both get worse.
4. Actuator unit economics. Eccentric Machines and its peers are the test of whether the picks-and-shovels layer produces a CATL or an Apple supplier.
5. BOT's premium or discount to NAV. Closed-end funds frequently trade at a discount. A structural discount would tell you what the market thinks of the marks on Level 3 private positions.