Wall Street research firm Bernstein says Washington just started a formal split between the US and Chinese robotics industries. That split runs through something more specific than nationality — a 65% American-content test on every new robot's parts list.

Bernstein's note is new. The rule it's describing is not.

On July 28, the Federal Communications Commission (FCC) placed foreign-produced advanced robots on its Covered List — a roster of tech the agency has flagged as a security risk. Gear on that list generally can't get the approval it needs to be imported or sold in the US, unless the maker wins a Conditional Approval, a case-by-case waiver. Robots already authorized before July 28 can generally keep selling. Fixed industrial arms, some medical devices, vehicles, and drones are excluded entirely.

Here's the part that changes the investment case: "foreign-produced" doesn't just mean Chinese-owned. A covered robot only counts as domestic if it's built in the US, with domestic components clearing 65% of total component cost. That threshold runs through 2028, then rises to 75% in 2029.

That's the Joint Tax. Every actuator, sensor, control board, connector, and wiring harness can now affect whether a robot clears the US market-access test.

The robotics race isn't just about who builds the best machine anymore. It's about who can document an American-enough bill of materials.

A humanoid robot isn't one big computer with legs. It's 28 to 70 motorized joints, each with its own motor-drive system: one chip to control power, another to sense current, a third — a magnetic encoder — to track position. Add a microcontroller, battery-management circuitry, and safety wiring, and one joint alone can carry a dozen-plus components.

Across a full robot, that adds up fast. A humanoid can hold hundreds of semiconductor sockets, and potentially thousands of individual parts, depending on what you count.

The AI processor that gets the headlines is often cited at just 1%-3% of that count. But part count isn't economic value — one AI processor can cost more than hundreds of small analog parts combined.

The AI brain wins the headlines. The distributed control system creates the larger number of qualified sockets.

Consulting firm McKinsey & Company breaks down where the cost actually sits. Actuators — the motors, gearboxes, and controls that move each joint — run 40%-60% of a humanoid's total bill of materials. Sensing and perception add another 10%-20%. Compute and control, the flashy AI-processor layer, is only 10%-15%.

The FCC's threshold is written in dollars of component cost, not number of chips. A robot stuffed with cheap capacitors doesn't clear that test any faster than one built around a single expensive processor.

Total chip content per humanoid still runs an estimated $1,400 to $4,000, well above the $500-$700 in an average car. Translate that into a market size, though, and the range gets wide fast.

Goldman Sachs, a Wall Street investment bank, models 1.4 million humanoid units by 2035. It puts the total robotics market at $38 billion that year. Bank of America, another Wall Street bank, sees a steeper curve: 90,000 units this year, 1.2 million by 2030, 10 million by 2035.

Run that same $1,400-$4,000 range against Bank of America's 10-million-unit case. You get roughly $14 billion to $40 billion a year in chip demand alone. Goldman's smaller case implies just $2 billion to $6 billion.

That's not forecast precision. That's venture-style dispersion. The opportunity could be enormous. Nobody's settled which curve is real.

Chips are the story everyone's chasing. Since actuators can run 40%-60% of a humanoid's bill of materials, the gearboxes and motors inside each joint may matter more to the FCC's math than any single processor.

That puts Schaeffler AG (Frankfurt: SHA0) and Harmonic Drive Systems, Inc. (Tokyo: 6324) in this theme. Schaeffler makes the strain-wave gearboxes used inside humanoid joints, calls them a scaling bottleneck, and targets mass production in 2027. Bernstein named both as potential beneficiaries of the split. Risk: both are foreign suppliers, and their FCC-related benefit still depends on where their specific parts get made.

Washington controls who gets access to the US customer. It doesn't control what's inside the joint.

Bernstein flagged rare-earth magnets as China's strongest remaining lever. These magnets go into the high-performance motors that power most robot joints, and China still dominates that supply chain.

A US-content rule can protect domestic robot makers from finished Chinese robots. A Chinese magnet restriction can raise the cost of every domestic joint those same makers try to build. Decoupling doesn't remove dependence. It just moves the bottleneck.

Texas Instruments Incorporated (Nasdaq: TXN). TI is integrating radar, motor control, and power tech into Nvidia's robotics platform. That's a technology collaboration, not a disclosed supply contract. Its broad US footprint could matter as makers redesign around the content test. Risk: no humanoid revenue or volume is disclosed yet.

Analog Devices, Inc. (Nasdaq: ADI). CEO Vincent Roche says humanoids are "steadily increasing" the company's opportunity pipeline. Its core industrial business is already half of revenue, up 56% year-over-year. Risk: most of that growth predates robotics.

Allegro MicroSystems, Inc. (Nasdaq: ALGM). Allegro is one of the few Western analog names willing to put a number on this. It expects robotics to reach roughly 3%-4% of fiscal 2027 sales. Risk: that figure covers more than just humanoids, and a delayed customer ramp could shrink it fast.

Monolithic Power Systems, Inc. (Nasdaq: MPWR). MPS sells an integrated power-and-motion package spanning motor drivers, encoders, and sensing, and uses its own tech internally. What's missing: a disclosed humanoid customer, a named production win, or any robotics revenue figure.

Two more worth a line each. TE Connectivity Ltd. (NYSE: TEL) and Amphenol Corporation (NYSE: APH) sell the connectors wiring every joint together — a lower-margin, volume business, not a pricing-power one. Infineon Technologies AG (Frankfurt: IFX), Melexis NV (Euronext Brussels: MELE), Murata Manufacturing Co., Ltd. (Tokyo: 6981), and TDK Corporation (Tokyo: 6762) are strong global robotics suppliers, but none wins automatically just for being non-Chinese — it depends on where their parts are made and on any Conditional Approval.

Unitree, AgiBot, and UBTech Robotics Corp. Ltd. (Hong Kong: 9880). These are China's leaders in unit volume and cost. Unitree priced its Shanghai IPO this month at roughly a $9 billion valuation, and retail demand topped 8,000 times the shares on offer. That proves investors want the stock — not what it trades at once it opens on August 19. And it doesn't erase Unitree's existing US business; the pressure falls on future models needing new authorization.

Nvidia Corp. (Nasdaq: NVDA), Qualcomm Incorporated (Nasdaq: QCOM), and Advanced Micro Devices, Inc. (Nasdaq: AMD). Nvidia has the strongest robotics ecosystem, with new partnerships spanning Infineon, NXP, and STMicroelectronics; Qualcomm and AMD are building rival platforms. None has enough humanoid volume yet to matter next to its far larger core business.

Tesla, Inc. (Nasdaq: TSLA). Musk calls scaling Optimus "the hardest product to scale manufacturing that we've ever made." Tesla hasn't disclosed cumulative production — as of its July update, first-generation lines were still being installed. The next real signal is sustained output and cost per working hour, not another demo video.

We're not buying the 2035 unit forecast. Goldman's 1.4 million and Bank of America's 10 million are too far apart to treat either one as an investable fact.

Instead, we're screening established industrial businesses where humanoid robotics is optionality, not the only reason the stock works. Two questions: does the supplier own a hard-to-replace, safety-relevant socket inside the joint, sensing system, or power architecture? And does its actual manufacturing location — not its headquarters — help a robot maker clear the FCC's US-content test?

On Conditional Approvals. More constructive if regulators grant these narrowly; more cautious if broad exemptions let makers keep foreign supply chains with just superficial US assembly.

On individual suppliers. More constructive the moment a name discloses a real humanoid production win and where that content gets made — a growing catalog with no disclosed robot revenue doesn't count.

On the magnet risk. More cautious on the whole buildout if China restricts rare-earth magnets faster than Western suppliers can respond.

On the demand curve. More confident in the higher-end forecasts once a major maker reports sustained production and repeat orders, not just installed capacity.

The FCC's July 28 action covers many new foreign-produced mobile robots broadly. It isn't limited to Chinese-owned brands.

The threshold that matters is cost, not chip count. Domestic components need to clear 65% of total component cost through 2028. That rises to 75% in 2029.

The best supplier picks combine a hard-to-replace joint, sensing, or power part with a US plant that can count toward that threshold.

The FCC didn't just restrict a Chinese robot. It changed the math behind every new covered robot trying to reach the US market.

Through 2028, more than 65% of a robot's component cost generally needs to qualify as domestic. Otherwise, the maker needs a Conditional Approval. That makes chip count the wrong scoreboard. The processor matters. So do the motor drivers, encoders, connectors, and sensors. The single largest cost pool may sit in the actuators and gearboxes that turn all of it into movement.

China builds many of the cheapest robots today, and it also controls the rare-earth magnets inside their motors. Washington controls access to the US customer. Neither side controls the whole chain.

The next phase of this race gets fought inside the bill of materials, not the demo video.

The AI brain gets the headlines. The joint may decide whether the robot gets through the door.

The HEAT (Hedge, Edge, Asymmetry and Theme) Formula is designed to empower investors to spot opportunities, think independently, make smarter (often contrarian) moves, and build real wealth.

The views and opinions expressed herein are those of the Chief Executive Officer and Portfolio Manager for Tuttle Capital Management (TCM) and are subject to change without notice. The data and information provided is derived from sources deemed to be reliable but we cannot guarantee its accuracy. Investing in securities is subject to risk including the possible loss of principal. Trade notifications are for informational purposes only. TCM offers fully transparent ETFs and provides trade information for all actively managed ETFs. TCM's statements are not an endorsement of any company or a recommendation to buy, sell or hold any security. Trade notification files are not provided until full trade execution at the end of a trading day. The time stamp of the email is the time of file upload and not necessarily the exact time of the trades. TCM is not a commodity trading advisor and content provided regarding commodity interests is for informational purposes only and should not be construed as a recommendation. Investment recommendations for any securities or product may be made only after a comprehensive suitability review of the investor’s financial situation.© 2026 Tuttle Capital Management, LLC (TCM). TCM is a SEC-Registered Investment Adviser. All rights reserved.