Andrew Kang's Shareholder Letter: Robotics Investment Reaches a Turning Point, Private Valuations Still Severely Lagging
Author: Andrew Kang
Compiled by: Deep Tide TechFlow
Deep Tide Overview: Robotics companies are transitioning from concepts to real orders and shipments. Andrew Kang's shareholder letter reveals key developments in portfolio companies such as Standard Bots, Figure AI, and Path Robotics. The letter repeatedly emphasizes the disconnection between private valuations and fair value, which is a signal that investors focusing on physical AI cannot ignore.
August 20, 2026
Shareholder Letter ------ August 2026 {#article-toc-33506-2}
Andrew Kang
To the shareholders of RoboStrategy, Inc.:
It has been just over three months since RoboStrategy went public on May 11. From April 30 to July 31, NAV grew by 85%, and per-share NAV increased by 54% (1). Most of the growth came from value-added financing to institutional investors. I personally invested an additional $10 million into the fund at a market premium of approximately $36.7 per share. Although we have seen a significant increase in investor demand and growth in portfolio companies, NAV pricing is still anchored to the last round of private market valuations. We believe that private financing rounds may not fully reflect the fair value of private companies, and the disconnection between private market valuations and fair value may persist. A recent data point is Unitree's IPO, which opened with a market value of $66 billion (2) — about 39 times its last venture capital valuation of $1.7 billion (3), and 7 times its IPO price (4). This value discrepancy may not represent the overall market situation. Unitree is currently not in RoboStrategy's portfolio, but it aligns perfectly with our investment focus and may be included in the portfolio in the future.
Since our public listing, we have invested $124 million in six companies, increasing our stakes in Standard Bots and Dexmate while also making new investments in Nox Metals, Eccentric Machines, Prometheus, and a robotics company that is about to emerge from stealth mode (5). As we scale, we are more inclined to lead investments in those companies we are most confident in. So far, we have led investments in Dyna Robotics, Standard Bots, Eccentric Machines, REK, and a stealth robotics company (5). As lead investors, we have the privilege of working closely with these truly outstanding founders. For RoboStrategy shareholders, this means the fund may benefit from the terms negotiated as lead investors. These lead terms may include board seats, information rights, pro-rata participation rights, warrants, preemptive rights, and additional investor protections.
The industry has undergone significant changes, such as the U.S. banning foreign mobile robots through the FCC, the rise of open-source models, and the accelerated timeline for developing foundational robotics models — all of which we anticipated. This positions our portfolio favorably, as these changes become tailwinds rather than headwinds. Many of our portfolio companies have moved from pilot phases to growth phases, making substantial progress in real deployments and shipments.
Portfolio Highlights {#article-toc-33506-3}
Standard Bots is the largest AI-native industrial robotic arm manufacturer in the U.S. (6). We led their $200 million Series C financing round, valuing them at $1 billion (7). They expect to deliver 10% of the new industrial robot deployments in the U.S. next year. Their robotic arms are in use at hundreds of companies across nearly every state. Their factory in Glen Cove will expand to 70,000 square feet to meet demand. In June, CEO Evan Beard participated in a White House roundtable on U.S. priority industry policy, hosted by Commerce Secretary Howard Lutnick and White House Trade Advisor Peter Navarro. He was the only founder of a U.S. robotics company in the room. Politico recently reported that the Department of Defense's Strategic Capital Office is underwriting a financing deal for Standard Bots.
Figure AI has begun ramping up production of Figure 03, with humanoid robot output surpassing 1,000 units in July (8). Their existing BotQ factory has an annual capacity of 12,000 robots, and CEO Brett Adcock stated the goal is to reach 1 million units per year within a decade. Figure has expanded from initial pilots at BMW to more complex sorting applications in the logistics of BMW's Spartanburg plant (9). They have signed a commercial agreement with Catalyst Brands to deploy humanoid robots at scale. Catalyst operates iconic brands such as JCPenney, Aéropostale, and Brooks Brothers.
Apptronik has opened Robot Park in Austin, Texas (10), a new 90,000 square foot data collection facility aimed at training Apollo 2 in collaboration with Google DeepMind. Apollo 2 continuously generates training data to train and optimize the Gemini Robotics model, preparing it for real-world deployment in commercial fleets. Apollo 3 is expected to be the first fully commercialized, mass-produced product, targeting a 2027 launch. They have also hired former Waymo product lead Daniel Chu as CPO, along with executives from Boston Dynamics and Amazon.
Dyna Robotics has released DYNA-2, its flagship world-action model, trained on over 1 million hours of human video data, estimated to be two orders of magnitude higher than previous robotic foundational models' human video data. DYNA-2 demonstrates the scaling laws from humans to robots, with the core finding being that robot performance improves smoothly and predictably with more human data, without plateauing. The model breakthroughs bring trainability and physical capabilities of agents, combined with enhanced robotic hardware, positioning them favorably to win broader deployment markets.
Dexmate has become the leading humanoid robot vendor in the U.S., having shipped hundreds of units. They are present in nearly all major physical AI teams, with clients including Nvidia, Amazon, Google, LG, Skild AI, Generalist, and top university labs such as UC Berkeley, Carnegie Mellon, Harvard, Columbia, and NYU. We led their seed round and Seed Plus round, with post-money valuations of $123 million and $216 million, respectively, as we believe a U.S. company has the opportunity to replicate Unitree's business model by providing high-quality humanoid robot platforms for developers and researchers, and they are indeed executing on this.
Path Robotics has signed a performance-based production agreement with HII (11) involving up to $600 million in shipbuilding work, aiming to deploy advanced robotics in U.S. Navy shipbuilding projects. This is one of the largest contracts ever awarded to a physical AI company. Path has also launched Rove, a welding quadruped robot that can move to the workpiece rather than bringing the workpiece to the robot. This is the only method that can achieve welding automation on structures too large to fit into a work cell, making it ideal for shipbuilding. Saronic, which has raised over $2.5 billion to build autonomous vessels, will receive the first Rove units in Q1 2027.
Eccentric Machines is developing Sentor, an AI-native actuator that replaces the standard "single gear motor" architecture with a new motion method designed to respond directly at the joints to unexpected loads, rather than waiting for the robot's central brain to process. Solving this issue brings robots closer to reflex capabilities, embedding the subconscious layer of physical AI into the joints, much like biology does. Sentor is designed with manufacturability in mind and addresses some of the most critical performance, efficiency, and mechanical trade-offs faced by modern actuators.
REK hosted what is reportedly the first six-foot humanoid robot fighting competition in the U.S. this month at its San Francisco venue, attracting a crowd of athletes and creators. Previous events sold out the 2,500-seat venue, and REK is now looking for a venue several times larger than this.
GMI is constructing a $500 million AI factory in Taoyuan, with approximately 7,000 Nvidia GB300s distributed across 96 racks, powered at 16MW, capable of processing nearly 2 million tokens per second. It has also launched a $12 billion sovereign AI infrastructure initiative in Japan and is collaborating with Magna AI to build a global network of sovereign AI factories.
Nox Metals completed a $11.5 million seed round after about seven months of establishment, with participants including Palmer Luckey, Y Combinator, and us. Nox has been cutting custom aluminum for about 100 customers nationwide, from small machine shops to Anduril and SpaceX, and is moving into a 75,000 square foot factory in Detroit to meet demand.
Team Growth {#article-toc-33506-4}
Our goal is not just to create a passive investment tool. We aim to build a high-caliber investment institution that creates excess value for the RoboStrategy, Inc. fund. We must not only deeply understand the companies and industries we invest in but also leverage our expertise in media to help portfolio companies tell their stories to more potential customers, talent, and investors. Since our inception, we have built the core of a marketing engine, covering short video content on X, high-production videos on YouTube, as well as long-form research papers and investment theses. Our own content has garnered over 20 million views on X, and our team's media appearances have exceeded 1.5 million views across various channels. Our reporting has been republished by Bloomberg, Reuters, CNBC, TechCrunch, BBC, and Yahoo Finance. In the coming months, we will roll out across all channels, including in-depth reports, industry shorts, expanding content creator and influencer relationships, growing newsletter subscribers, live streams, and podcasts, covering all content related to robotics and physical AI. We plan to go beyond our core English-speaking audience and build a global presence, initially targeting countries like South Korea, Japan, Taiwan, and China. The goal is to become the primary source of information for all dynamics in the robotics field.
The RoboStrategy Advisors team has grown to 14 members, with plans to expand to 17 by September. Functional departments now cover investment, research, marketing, compliance, finance/operations, and legal/policy. The U.S. is formulating a national robotics strategy. Just this month, the FCC issued a notice banning the import of new foreign mobile robot models. The government can and will take many measures to promote the development of domestic companies, potentially including long-term debt financing programs, direct investments, and tax incentives. We hope to share our insights with policymakers and those of our portfolio companies. To this end, Bill Hughes has joined us as Vice President of Legal and Policy. Bill previously served as Deputy Assistant Attorney General at the U.S. Department of Justice and as Special Assistant to the President at the White House.
We are also forming a team for institutional capital formation and distribution. Robotics technology will attract a growing interest from various investors, and I believe our fund offers a unique value proposition that will appeal to RIAs, private wealth platforms, and various institutions. These capital channels are currently undeveloped for us, but they have always been core growth channels for most large asset management firms. Approximately $200 billion flows annually from private wealth channels into semi-liquid asset funds (12). Expanding the fund size will enable RoboStrategy to dominate more financing rounds, benefiting both the companies we invest in and our shareholders, while also strengthening our distribution engine.
About Research {#article-toc-33506-5}
To underwrite an investment, you need to form a clear judgment on a multitude of interconnected technological trends to accurately assess risks and returns. The growth prospects of a robotics foundational model company depend not only on the strength of its researchers and models but also on the trajectory of open-source model development, the feasibility of cloud-based model deployment across environments, and numerous variables affecting robotic productivity, such as the feasibility of new actuator innovations and the supply elasticity of gear grinders produced by a few small companies. These gear grinders are the bottleneck for harmonic reducers, which are key components in many robotic forms. We plan to publish some of this research to share knowledge with the world and enhance our brand and influence. Our investors know that we adhere to high standards in investment due diligence and extend this standard to all our work.
Industry Outlook {#article-toc-33506-6}
The first half of this year has sounded the alarm for the venture capital industry. People have realized that the scale of venture capital in the pure software domain should be significantly reduced, while investing in companies building the atomic world can yield extreme power-law returns. We hear internal and external pressures from many investment firms' LPs to readjust allocation strategies. Sometimes investors need undeniable evidence, and companies like SpaceX and Anduril, which are entering growth stages, have already provided such proof. In fact, SpaceX itself has created approximately $2 trillion in liquidity events (13) for investors and employees, which may lead to substantial downstream funding flowing to other deep tech companies.
Compared to previous quarters, venture capital in the robotics sector significantly increased in the first quarter, reaching the highest level on the chart. However, venture capital in the robotics sector is still just a small fraction of the venture capital in the AI sector before the emergence of ChatGPT in 2021. This is a turning point for financing in the industry, and we expect the robotics sector to catch up with AI in the coming years.
We believe now is a good time for the fund to invest in private robotics companies, but some companies' financing is getting out of control. Discernment is crucial. Differentiation is key. The rising tide is currently lifting many boats in the sector, but many companies building towards the wrong goals or lacking the necessary talent may run aground even if the tide does not recede. Winner valuations continue to climb, while some hyped participants hit walls; both situations can coexist. We have seen this in the AI sector over the past four years.
The deployment of industrial capabilities is entering the early scaling phase. We believe many key architectural issues of robotics foundational models have been fundamentally resolved, and this year's focus is primarily on scaling data collection and processing. I believe we have been in the GPT-3 era, and the emergence of physical versions of intelligent capabilities is becoming apparent, accelerating towards GPT-5 level performance next year. These emergent capabilities include contextual learning, memory, and natural language instruction following. Researchers have grasped the overall scheme for achieving general physical intelligence. We are on the verge of a significant unlocking, with a substantial increase in sample efficiency in post-training, reducing deployment time for robots in any specific role from weeks to hours or even minutes. For robots, there will not be a ChatGPT moment; it will look more like autonomous vehicles. It will not be suddenly recognized globally, but various shapes and sizes of new autonomous machines will rapidly permeate. By next year, intelligence will no longer be a bottleneck; the real bottleneck will be the robots themselves. Please feel free to reach out to the RoboStrategy team with any questions. Thank you for your support.
Andrew Kang
Chief Executive Officer
-- Price
This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.
You may also like

Bitcoin Spot Demand: The Signal That Hadn't Reappeared Since the October 2025 Record

Vietcombank Warns of AI-Driven Fraud

$1 Billion in 48 Hours: X Prepares Crypto Trading Function on Feed

The forex complex: How retail traders are adapting to volatile 2026 markets

Trump Beats the Fed: How Trading Boom Explodes in the UAE

Crypto investors should favor systematic strategies over Fed predictions, Moon Pursuit Capital says

Crypto: Solana ETFs Record Their Biggest Day of the Year

Xbox and PlayStation Reassess Their Multiplatform Strategy, Emphasizing Exclusivity Again

Interview: Aptos CEO on AI Agents, Stablecoins, and Machine Commerce

BlackRock Lowers Minimum Investment for Bitcoin ETF to $1 Million

Previously Bearish on Crypto Market, Jiang Zhuoer Quickly Turns Bullish: Ethereum is the Engine of the Bull Market

Michael Saylor Targets Bitcoin Orthodoxy: 'Satoshi Is Not a Prophet'

Strive by Vivek Ramaswamy Becomes the 7th Largest Public Holder of Bitcoin

The Best AI Models for Trading: A Comparison of the Top 10 Models

Enflame, the Chinese competitor of Nvidia in AI chips, goes public

AI is in a Credit Expansion Phase: The Stronger AI Becomes, the More the Federal Reserve Needs to Cut Rates

Short-Term U.S. Treasury Holdings Decrease by $29 Billion, Testing Demand for Stablecoins

Analysis of Mining Machine Asset Depreciation and Tax Cost Recovery

Bitcoin Transfer by Metaplanet: Sale or Just Custody on Coinbase?

Opinion: The Bull Market Has Arrived, How to Position in This Cycle?

80% of Gas Stations in Russia Face AI-95 Gasoline Shortage

Bitcoin Bull Run: Arthur Hayes Sees Crypto Market Soaring Higher

Goldman Sachs Optimistic About Crypto Brokerage Trading Platforms: Can the Market Support a New Cycle?

Strive CEO: The Bitcoin Bull Market Has Just Begun, ASST Has Passed Its Bottom

Ravencoin (RVN) Price Forecast for 2026–2050

AI Agent's 'Coming of Age': What Step Is Missing from Simulation Training to Real Trading?

Bitcoin Just Hit $80,000: Last Week WEEX's Prediction Came True — What's Next?
Bitcoin hit $80,000 as WEEX predicted last week. See what drove the breakout, current key levels, and whether $90,000 is next.

Mining Artificial Intelligence: Why Bitcoin Miners Are Changing Their Business Model

CFTC Considers the Viability of Perpetual Futures in GPU Computing












