## Is Humanoid Robotics the Next Major Asset Class for Investors?

Morgan Stanley predicts humanoid robotics will become a $5 trillion global industry by 2050, with Barclays projecting a nearer-term $200 billion market by 2035. Those numbers are now driving real capital allocation decisions — most visibly in the launch of KOID, KraneShares' first US-listed humanoid robotics ETF. Stephanie Link, chief investment strategist at Hightower, which manages $353 billion in assets, told Business Insider she has recently begun investing in the sector and is actively advising clients to take a position. The investment thesis rests on three converging forces: an aging global workforce creating structural labor shortages, generative AI dramatically accelerating robot autonomy, and hardware cost compression — Barclays cites a roughly 30-fold decline in humanoid robot prices over the last decade — that is finally making commercial deployment economically viable.

For engineers and founders tracking this space, the financial world's attention matters. Capital inflows at this scale fund the hardware iteration cycles, the data collection infrastructure, and the foundation model training runs that actually move the technology forward.

---

## The KOID Thesis: Labor Scarcity Meets Hardware Cost Collapse

KraneShares CIO Brendan Ahern framed the core argument bluntly: generative AI has been the "nitrous oxide" that accelerated a hardware trend already in motion. The mechanical side — actuators, battery density, structural components — has been improving steadily, with Barclays' cited cost decline of nearly 30 times over the past decade being the single most important commercial unlock. The AI layer on top of that cheaper hardware is what makes [Physical AI](https://humanoidintel.ai/glossary/physical-ai) plausible at scale rather than just in controlled lab environments.

The labor replacement arithmetic is straightforward, per Barclays' analysis cited in the piece: even a humanoid operating at half the efficiency of a human worker — accounting for the shorter operational window of current battery systems, which max out between four and six hours per charge — would still deliver roughly 25% more daily output. That calculation changes significantly as battery technology improves and as [whole-body control](https://humanoidintel.ai/glossary/whole-body-control) capabilities mature to handle more complex, variable tasks.

Ahern draws an explicit parallel to KraneShares' 2017 launch of a global electric vehicle ETF, timed to the early expansion of the Chinese EV industry. His firm is now applying the same thesis to humanoids: China is the current global leader, a robot recently set a global half-marathon record in Beijing, and the country's deployment velocity offers a preview of where the rest of the world is heading.

[Unitree Robotics](https://humanoidintel.ai/companies/unitree-robotics) featured in the article's opening scene — two units performing a synchronized dance in KraneShares' Midtown Manhattan office — serves as a representative data point for how far consumer-facing robot capability has advanced, even as Joseph Dube, KraneShares' head of marketing, acknowledged they don't look particularly realistic in person either. That honesty is actually useful: it calibrates where we are on the capability curve.

---

## What Lit the Match for Mainstream Investor Interest

Link's entry point into the sector is instructive for understanding how institutional attention shifts. Amazon CEO Andy Jassy's 2025 shareholder letter, which highlighted the company's deployment of robots across half of its distribution facilities and referenced a fleet of one million robots, was her stated catalyst. "If one of the biggest companies in the world is seeing massive efficiencies," she told Business Insider, "then the number of robots is going to go much higher over the long haul."

That framing — a Fortune 10 CEO explicitly endorsing the ROI case in a shareholder letter — carries more weight with institutional allocators than any analyst note. It signals that the technology has cleared the threshold from experimental to operationally validated, at least in structured logistics environments.

Link's baseball metaphor for the sector's maturity is worth noting: if generative AI is in the third inning, and cybersecurity is in the second, humanoid robots "haven't even stepped into the batter's box." That positioning, from a strategist overseeing $353 billion in client assets, suggests institutional allocation to the space is still in its earliest stages.

---

## The Skeptical Read: What the Bull Case Leaves Out

The investment framing in this piece is almost entirely bullish, and several meaningful risks are underweighted.

**Deployment complexity is non-trivial.** The Barclays cost-per-unit decline is real, but unit economics only tell part of the story. Integration costs, safety certification, retraining pipelines, and the human infrastructure required to supervise robot fleets in unstructured environments add materially to total cost of ownership. [Dexterous manipulation](https://humanoidintel.ai/glossary/dexterous-manipulation) in variable, real-world conditions remains genuinely hard — the gap between a robot completing a task in a demo and completing it reliably across thousands of SKUs in a warehouse is not closed by cheaper actuators alone.

**The 2050 horizon is a long time.** Morgan Stanley's $5 trillion figure is real enough to cite, but a 24-year projection in a sector defined by hardware-software co-evolution and geopolitical supply chain risk carries substantial uncertainty. The EV parallel that Ahern invokes is instructive both ways: the Chinese EV industry did indeed scale dramatically, but it also triggered tariff walls, stranded assets, and incumbent collapses that weren't priced into early ETF positions.

**ETF structure concentrates exposure in public-market proxies.** The actual frontier of humanoid development — [Figure AI](https://humanoidintel.ai/companies/figure-ai), [Agility Robotics](https://humanoidintel.ai/companies/agility-robotics), [Apptronik](https://humanoidintel.ai/companies/apptronik), and most of the technically interesting Chinese players — is private. Public-market humanoid ETFs will largely hold component suppliers, conglomerates with robotics divisions, and publicly listed manufacturers, not the pure-play startups where the highest-risk, highest-return outcomes will be determined.

**The "diversify across the breadth of change" advice** that closes the article is sound as far as it goes, but it's also what ETF product marketing always says. Investors who can access private rounds will have very different exposure than those buying KOID.

---

## Industry Trajectory Implications

The financialization of humanoid robotics — ETF launches, mainstream wealth management coverage, shareholder letter validation — has a direct feedback effect on the engineering and startup ecosystem. It increases the pool of capital available for hardware iteration, which in turn accelerates the [sim-to-real transfer](https://humanoidintel.ai/glossary/sim-to-real-transfer) cycles that actually improve deployed performance. More capital also means more data collection infrastructure, which feeds the foundation models that companies like [Physical Intelligence (π)](https://humanoidintel.ai/companies/physical-intelligence) and [Skild AI](https://humanoidintel.ai/companies/skild-ai) are building.

The China dimension cannot be ignored from a competitive standpoint. Ahern's observation — that his Asian travels inspired the ETF launch, just as they did for the EV fund — implies that the manufacturing scale, government subsidy structure, and deployment velocity on the Chinese side are far enough ahead to be the reference case rather than just a competitor. That has direct implications for which companies and which hardware architectures end up setting the de facto standards.

For founders raising now: mainstream financial media attention of this type historically precedes, by 12 to 24 months, a significant expansion in the LP base willing to write robotics checks. The pipeline is filling.

---

## Key Takeaways

- **Morgan Stanley projects a $5 trillion humanoid robotics market by 2050**; Barclays puts the 2035 addressable market at $200 billion — both figures from the source article.
- **KraneShares launched KOID**, the first US-listed humanoid ETF, with the same strategic thesis the firm applied to Chinese EVs in 2017.
- **Hardware cost compression** — Barclays cites roughly a 30-fold price decline over the last decade — is the foundational commercial unlock; generative AI is the accelerant on top.
- **Current battery limitations** (4–6 hours per charge for most humanoids) remain a real operational constraint, though the labor economics can still work at half human efficiency.
- **Institutional capital is early**: a $353 billion wealth manager only recently entering the space signals that mainstream allocation has barely begun.
- **Public ETFs provide indirect exposure**; the most technically significant humanoid companies remain private.
- **China is the current deployment leader**, and the competitive dynamics there will shape global hardware and software standards.

---

## Frequently Asked Questions

**What is the KOID ETF and what does it invest in?**
KOID is KraneShares' US-listed humanoid robotics ETF, designed to provide diversified exposure to the humanoid robotics industry rather than concentration in a single company. Because most pure-play humanoid startups are private, the fund's holdings will primarily include publicly traded companies with significant humanoid robotics exposure.

**What is the projected market size for humanoid robots?**
Morgan Stanley projects the humanoid robotics market will reach $5 trillion by 2050, with more than a billion robots deployed globally. Barclays offers a nearer-term figure of $200 billion by 2035. Both projections are cited in the KraneShares investment thesis.

**Why are humanoid robot prices falling so fast?**
Barclays cites a roughly 30-fold decline in humanoid robot prices over the last decade, driven by improvements in actuator manufacturing, battery technology, and structural components — the same hardware cost curves that previously played out in consumer electronics and electric vehicles.

**How long can humanoid robots operate on a single charge?**
Most current humanoid robots max out at between four and six hours per charge, per figures cited in the Business Insider source article. Despite this limitation, Barclays' analysis suggests a robot operating at half human efficiency could still deliver approximately 25% more daily output.

**Why is China considered the global leader in humanoid robotics?**
China's combination of government industrial policy, manufacturing scale, and aggressive domestic deployment has put it ahead on metrics like production volume and public demonstrations — including a robot completing a half-marathon in Beijing this year. KraneShares' Ahern explicitly cited observations from Asian travel as a key input to launching the humanoid ETF, drawing a direct parallel to the firm's early positioning in Chinese EVs.