The Digital Heartbeat: Rolex, AI, and Counterfeit Protection

Few brands have matched the impact Rolex has had on the evolution of horology. From the 1926 debut of the “Oyster,”[1] the world’s first waterproof case, to the perfection of the “perpetual self-winding rotor”[2] in 1931, Rolex has consistently defined the standards of reliability and luxury that govern the industry today.

This legacy of innovation is backed by a formidable intellectual property portfolio. As of early 2026, Rolex and its key subsidiaries hold approximately 2,597 global patent filings, representing roughly 976 unique patent families.[3] Of these, approximately 1,455 patents/applications, or an estimated 56% of the total portfolio, remain active.[4]

Rolex is currently leveraging artificial intelligence to bridge the gap between traditional craftsmanship and the future of high-precision timekeeping.

Science of Sound

Traditional acoustic timing machines function as simple peak-to-peak sound analyzers. They focus primarily on the first of five distinct mechanical collisions, or “shocks,” that occur during the escapement cycle:

1. Shock 1: The impulse jewel strikes the pallet fork (unlocking);
2. Shock 2: The escape wheel tooth strikes the pallet impulse plane (impulse);
3. Shock 3: The pallet fork catches the impulse pin;
4. Shock 4: The escape wheel tooth drops onto the pallet resting surface (locking);
5. Shock 5: The pallet fork rod hits the limiting banking pin.[5]

While prior art systems could detect broad groups of noise, they lacked the resolution to distinguish between internal components (e.g., Shocks 2 vs. 3 or Shocks 4 and 5).[6] The introduction of external noise into the testing environment creates a “masking effect” that further degrades the signal, burying these micro-shocks beneath a layer of parasitic interference.[7]  Traditional threshold-based sensors are often unable to “listen through” this noise, resulting in the unstable “snowy” displays common in conventional chronometric testing.

Rolex’s Patent Application Publication: WO/2025/262258

To help solve these technical challenges, Rolex recently published patent WO/2025/262258[8], detailing an innovative “Method for Determining a Characteristic of a Timepiece Movement,” which utilizes acoustic analysis and machine learning to evaluate the mechanical health of a watch. The method replaces traditional Fourier Transforms with Discrete Wavelet Transforms (DWT) using the Daubechies (db5) or Symlets (sym5) families.[9]

A wavelet is a “brief oscillation.”[10] Unlike a traditional wave (like a sine wave) that goes on forever, a wavelet starts at zero, increases in intensity, and then quickly returns to zero.[11] Because they are short and “local,” they can tell you both the frequency and the exact microsecond the sound occurred.[12]

This allows the system to deconstruct the watch’s “tick” into five distinct levels of information:

1. Levels 1 and 2: High-frequency data used for chronometric rate calculations;
2. Level 3: Isolation of mechanical disturbances or “ringing”;
3. Level 4: Damping analysis, revealing energy absorption and the state of lubrication/friction; and
4. Level 5: Intrinsic noise analysis to filter out background interference.[13]

By processing these signals through a neural network trained on a database of “ideal” movement sounds, the method can achieve effective denoising and better distinguish the different shocks during the escapement’s operation.

The “Crown”

The integration of AI in determining characteristics of timepieces marks a shift from repair to preventive maintenance.

Some key advantages are:

1. Non-invasive integrity: This process bypasses the traditional requirement for physical entry, thereby protecting the delicate caliber from external particulates and the risks associated with improper handling. Furthermore, because the case remains sealed, owners can receive detailed diagnostic reports without the risks typically associated with opening a factory-sealed watch.
2. Acoustic fingerprinting and authentication: The AI method creates a unique “acoustic fingerprint” for every timepiece. By decomposing the vibrational signature of the movement, the system can distinguish between genuine factory components and aftermarket parts.
3. Component-level diagnostics: The method can diagnose specific mechanical failures.
4. Adaptive denoising for retail environments: High-precision diagnostic testing can be performed in busy retail environments or service centers without sacrificing the accuracy of the chronometric data.
5. Objective amplitude measurement: The system marks the exact start and end of the “drop” phase to calculate actual amplitude acoustically. This removes the “lift angle” guesswork inherent in traditional timegraphers, providing a measurement that is as reliable as optical laser testing.

Rolex’s ability to innovate and incorporate AI strengthens its position in the industry by improving its ability to develop precise and high-quality timepieces.

Defending the “Crown”

Rolex maintains its reputation as one of the largest names in the watch market through targeted litigation against counterfeit products. Rolex’s approach is often conservative but precise, choosing to defend its rights against third-party vendors who may sell counterfeit versions of its products.

For example, in 2024, Rolex prevailed in a lawsuit against BeckerTime L.L.C., a third-party retailer of watches.[14] Rolex Watch USA, Inc. sued defendant BeckerTime L.L.C. (herein “BeckerTime”), alleging trademark infringement.[15] BeckerTime often purchased used Rolex products, modified and refurbished them to include both Rolex and non-Rolex parts, and resold them to consumers as “Genuine Rolex” products.[16]

BeckerTime attempted to argue that it did not infringe on Rolex’s trademarks because they did no more than repair and restore “second hand” Rolex products.[17] However, as the Fifth Circuit Court of Appeals noted, BeckerTime was doing more than repairing vintage Rolex watches.[18] BeckerTime was producing modified watches with added diamonds, bezels, and additional non-genuine parts.[19] BeckerTime’s changes to the Rolex products amounted to a material alteration of the watches, such that a reasonable customer would be confused about whether the watch is “fully genuine Rolex”.[20] As a result, the 5^th Circuit Court of Appeals affirmed a verdict of trademark infringement in favor of Rolex.[21] Rolex was granted an injunction to prevent BeckerTime from marketing modified Rolex products as “Genuine Rolex” products.[22]

Future Measures to Protect the “Crown”

The results of Rolex Watch USA, Inc. v. BeckerTime, L.L.C. show that Rolex has an interest in ensuring products on the resale market are genuine. Rolex took specific actions in BeckerTime to prevent a resale vendor from selling products that included non-Rolex parts. Notably, BeckerTime’s modifications to the watches were external changes, primarily directed to changes in the external bezels and bands of the watches.

In contrast, other resale vendors may make internal part changes, such as replacing a genuine pallet fork or escape wheel tooth with a non-genuine part, which would be far more difficult to detect. The technology of patent WO/2025/262258 offers a practical solution to the problem. Using this technology, Rolex can quickly determine whether a watch contains genuine internal components through AI. The WO/2025/262258 technology removes any need to take apart a watch, allowing Rolex to easily determine whether a resale product is genuine or not. Rolex may use this technology as a new way to authenticate products and deter the sale of counterfeit Rolex products. Rolex’s use of AI is a good example of how a company can use the technology to not only improve product development and maintenance, but also as a way to strengthen their IP portfolios and guard against infringement.

[1] See Swiss Patent No. 120,851 (issued June 16, 1927).

[2] See Swiss Patent No. 160,803 (issued March 31, 1933).

[3] See Rolex Patent Portfolio – Exemplary Landscape Overview, IIPRD (Sept. 19, 2025), https://www.iiprd.com/rolex-patent-portfolio-exemplary-landscape-overview/.

[4] Id.

[5] See WIPO Patent Application Publication No. 2025/262258 (published Dec. 26, 2025).

[6] Id.

[7] Id.

[8] Id.

[9] Id.

[10]See What Is a Wavelet?, MathWorks, https://www.mathworks.com/help/wavelet/gs/what-is-a-wavelet.html (last visited Mar. 23, 2026).

[11] Id.

[12] Id.

[13] WIPO Patent Application Publication No. 2025/262258, supra note 5, at Fig. 11.  

[14] See Rolex Watch USA, Inc. v. BeckerTime, L.L.C., 96 F.4th 715, 715 (5th Cir. 2024).

[15] Id.

[16] Id. at 719.

[17] IId. at 722, see also Champion Spark Plug Co. v. Sanders, 331 U.S. 125, 126-127.

[18] Id.

[19] Id.

[20] Id.

[21] Id. at 224.

[22] Id. at 225.