Granite Vehicle Ventures LLC v. Tesla Inc

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Granite Vehicle Ventures LLC (Texas)
  • Defendant: Tesla, Inc. (Texas)
  • Plaintiff’s Counsel: Heim, Payne & Chorush LLP; Miller, Fair, Henry PLLC
• Case Identification: 2:24-cv-01007, E.D. Tex., 02/19/2025
• Venue Allegations: Plaintiff alleges venue is proper in the Eastern District of Texas because Defendant is a Texas corporation, thereby residing in the district, and because Defendant maintains regular and established places of business within the district where acts of infringement have allegedly occurred.
• Core Dispute: Plaintiff alleges that Defendant’s vehicles equipped with its Full Self-Driving (FSD) (Supervised) software infringe three patents related to systems for controlling the driving modes of self-driving vehicles.
• Technical Context: The technology at issue concerns methods for dynamically determining whether a semi-autonomous vehicle should operate in autonomous or manual mode, based on assessments of the vehicle's processor competence, the human driver's competence, and external conditions.
• Key Procedural History: The Asserted Patents claim priority to a patent application filed on September 25, 2015. The rights to this application were originally assigned to International Business Machines Corporation (IBM) and subsequently transferred through multiple entities before being assigned to Plaintiff Granite Vehicle Ventures LLC.

Case Timeline

Date	Event
2015-09-25	Priority Date for Asserted Patents
2016-10-01	Approximate start of Accused Product hardware availability (all vehicles made after Oct. 2016)
2023-03-07	U.S. Patent No. 11,597,402 Issues
2023-08-29	U.S. Patent No. 11,738,765 Issues
2024-07-16	U.S. Patent No. 12,037,004 Issues
2025-02-19	Complaint Filing Date

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 11,597,402 - "Controlling Driving Modes of Self-Driving Vehicles"

The Invention Explained

• Problem Addressed: The patent describes a technical environment where self-driving vehicles (SDVs) can operate in either a manual mode controlled by a human or an autonomous mode controlled by an on-board processor (Compl. ¶45; ’402 Patent, col. 6:41-54). The implicit technical problem is managing which mode is safer or more effective under various circumstances, particularly when an "operational anomaly" occurs (’402 Patent, col. 9:60-63).
• The Patented Solution: The invention proposes a system that determines a “competence level” for both the vehicle’s control processor and the human driver in the context of a given operational anomaly or roadway condition (’402 Patent, abstract). Based on a comparison of these two competence levels, the system selectively assigns control of the vehicle to whichever is determined to be relatively more competent (’402 Patent, abstract; ’402 Patent, col. 2:41-54).
• Technical Importance: This approach provides a framework for dynamic, context-aware risk management in semi-autonomous vehicles, moving beyond simple on/off system engagement to a comparative assessment for control allocation.

Key Claims at a Glance

• The complaint asserts independent claim 4 (Compl. ¶50).
• The essential elements of independent claim 4 include:
  • A self-driving vehicle (SDV) comprising a sensor system, vehicle controls, and a computer system.
  • The computer system is capable of receiving sensor readings, operating vehicle controls, and determining the SDV’s operational state.
  • The computer system is capable of determining a vehicle fault.
  • The computer system is capable of determining a competence level of the processor.
  • The computer system is capable of determining a competence level of a human driver.
  • The computer system is capable of determining a corrective action based on a comparison of the processor and human driver competence levels.
  • The computer system is capable of implementing the corrective action and issuing an alert indicating it.
• The complaint reserves the right to assert additional claims (Compl. ¶52).

U.S. Patent No. 11,738,765 - "Controlling Driving Modes of Self-Driving Vehicles"

The Invention Explained

• Problem Addressed: As with the related ’402 Patent, the invention addresses the challenge of deciding when an SDV should transition between autonomous and manual control, particularly when a fault occurs (’765 Patent, col. 1:21-30).
• The Patented Solution: The patent claims a computer program product that executes a method for controlling the SDV’s driving mode. The method involves determining if a "fault" (defined as a "current weather condition") has occurred that exceeds a "threshold for danger" (determined by assessing a "control processor competence level") (’765 Patent, claim 1). If the threshold is exceeded, the system determines a corrective action using a "fault-remediation table" and implements it by transferring control to the driver and issuing an alert (’765 Patent, claim 1; ’765 Patent, col. 8:29-45).
• Technical Importance: This patent provides a more structured decision-making logic by explicitly linking the fault to weather conditions and introducing a "fault-remediation table" as the mechanism for selecting a pre-determined corrective action.

Key Claims at a Glance

• The complaint asserts independent claim 1 (Compl. ¶76).
• The essential elements of independent claim 1 include:
  • A computer program product for performing a method comprising:
  • Receiving sensor readings, including from a GPS sensor.
  • Determining if a fault has occurred, where the fault is a "current weather condition."
  • Determining if the fault exceeds a danger threshold by "determining a control processor competence level."
  • Determining a corrective action using a "fault-remediation table."
  • Implementing the corrective action, which comprises transferring control to manual and alerting the driver to take over.
• The complaint reserves the right to assert additional claims (Compl. ¶78).

U.S. Patent No. 12,037,004 - "Controlling Driving Modes of Self-Driving Vehicles"

• Technology Synopsis: The ’004 Patent claims a computer program product that uses specified sensors (steering wheel, speedometer, GPS, camera) to determine the competence levels of both a human driver and the SDV (Compl. ¶109). The system compares the human driver's competence level against multiple, different thresholds to determine if a "first fault" or a more severe "second fault" has occurred, and then uses a "fault remediation table" to implement distinct corrective actions for each fault level, such as a warning versus a system disengagement (Compl. ¶109).
• Asserted Claims: The complaint asserts independent claim 1 (Compl. ¶108).
• Accused Features: The complaint alleges infringement by Tesla's FSD system, which monitors driver inattentiveness (e.g., looking at a phone, hands off the wheel for a period of time) against different thresholds and implements tiered responses, such as an initial warning ("first corrective action") followed by a system disengagement ("second corrective action") for continued inattention (Compl. ¶¶140-150; Compl. ¶157; Compl. ¶159).

III. The Accused Instrumentality

Product Identification

The Accused Products are Tesla’s Full Self-Driving (FSD) (Supervised) software program and all Tesla vehicles compatible with it, including the Model 3, Model S, Model X, Model Y, and Cybertruck models manufactured since 2016 (Compl. ¶46).

Functionality and Market Context

The complaint identifies FSD (Supervised) as an SAE Level 2 driver-assist system, which requires constant driver supervision (Compl. ¶37). The system relies on a suite of cameras and a specialized "FSD Computer" to process data through neural networks and execute driving maneuvers such as steering, stopping, and navigating intersections (Compl. ¶53; Compl. ¶57; Compl. ¶59). A diagram in the complaint shows the placement of cameras around a Tesla Model Y, including on the door pillars, front fenders, rear license plate, and windshield (Compl. p. 14).

A key feature alleged to be relevant is the system's driver monitoring functionality, which uses a steering wheel sensor to detect turning force and an interior cabin camera to monitor driver attentiveness (Compl. ¶55; Compl. ¶66; Compl. ¶67). The system is alleged to issue warnings or disengage if it determines the driver is inattentive or if external conditions, such as poor weather or sun glare, degrade its own performance (Compl. ¶62; Compl. ¶70). The complaint notes that as of April 2024, approximately 400,000 Tesla owners use the FSD (Supervised) service (Compl. ¶32).

IV. Analysis of Infringement Allegations

’402 Patent Infringement Allegations

Claim Element (from Independent Claim 4)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
a sensor system comprising a plurality of sensors	The accused vehicles include multiple cameras, ultrasonic sensors, and steering wheel sensors to monitor the environment and driver.	¶54	col. 6:39-40
the computer system is capable of determining the operational state of the self-driving vehicle (SDV)	The system uses a steering wheel icon on the touchscreen to display whether FSD is available but not engaged (gray icon) or actively engaged (blue icon).	¶60	col. 12:47-49
the computer system is capable of determining a vehicle fault	The system disengages FSD when it cannot operate safely, such as when sensors are blinded by the sun or in poor weather, which the complaint frames as determining a fault.	¶62	col. 7:62-8:3
the computer system is capable of determining a competence level of the processor	The system determines its own competence is insufficient in non-ideal conditions (e.g., snow), triggering alerts that FSD may be degraded.	¶64	col. 10:10-14
the computer system is capable of determining competence level of a human driver	The system uses a cabin camera and steering wheel sensors to monitor driver attentiveness, which the complaint equates to determining the driver's competence level.	¶66	col. 10:55-58
the computer system is capable of determining a corrective action...	The system determines to issue a warning if the driver is distracted, or to disengage FSD if the processor's competence is low due to poor weather.	¶70	col. 11:25-30
the computer system is capable of implementing the corrective action	If a driver is inattentive, the system suspends FSD for the remainder of the drive, flashes the screen, and slows the vehicle to a stop. A screenshot shows the message "Autosteer unavailable for the rest of this drive" (Compl. p. 22).	¶74	col. 11:10-13
the computer system is capable of issuing an alert indicating the corrective action	If the driver's competence is deemed insufficient (e.g., inattentive), the system issues a visual and audible warning to pay attention. A screenshot shows the "Please pay attention to the road" alert (Compl. p. 23).	¶75	col. 11:13-17

• Identified Points of Contention:
  • Scope Questions: A central question may be whether Tesla's monitoring of driver "attentiveness" through sensors constitutes "determining a competence level of a human driver" as claimed. The defense may argue that "competence" implies an assessment of skill or capability (e.g., in specific weather), whereas Tesla's system measures only the state of paying attention.
  • Technical Questions: It may be disputed whether the FSD system's response to poor environmental conditions (e.g., disengaging due to snow or sun glare) is equivalent to "determining a competence level of the processor." The defense could frame this as a simple operational limit or sensor failure, rather than an assessment of processor "competence" as described in the patent.

’765 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
determining... whether a fault has occurred; ... the fault comprises a current weather condition of the roadway	The complaint alleges that FSD may be degraded or disengaged due to poor weather conditions, such as snow or low visibility from direct sun, which it equates to determining a weather-based fault.	¶89; ¶100; ¶101	col. 8:29-32
determining whether the fault exceeds a threshold for danger comprises determining a control processor competence level	The complaint alleges that when FSD determines its performance is degraded by poor weather, this is a determination that the processor's competence level is low and the fault exceeds a danger threshold.	¶90; ¶102; ¶104	col. 10:10-14
determining a corrective action associated with the fault using a fault-remediation table	Plaintiff alleges that FSD uses a neural network for decision-making, and that this process is "akin to cross-referencing a table" where specific faults correspond to specific corrective actions instilled during training.	¶92; ¶95; ¶96	col. 8:29-34
the corrective action comprises transferring driver controls to manual control and alerting a human driver to take over immediately	When conditions are sufficiently poor, the system displays a "Take over Immediately" instruction, sounds a chime, and transfers control to the driver. A screenshot shows this exact alert (Compl. p. 30).	¶97; ¶105; ¶106	col. 11:13-17

• Identified Points of Contention:
  • Scope Questions: The case may turn on the construction of "fault-remediation table." The complaint's theory that a neural network is "akin to" a table suggests this is a known point of dispute. A court will have to decide whether a complex, probabilistic system like a neural network can be construed as the structured, deterministic "table" described in the patent's specification.
  • Technical Questions: The infringement theory hinges on equating FSD's performance degradation in bad weather with the patent's sequence of "determining a fault," assessing a "processor competence level," and then determining if a "danger threshold" is met. The defense may argue that the accused system follows a different, more integrated logic that does not map onto these discrete, sequential steps as claimed.

V. Key Claim Terms for Construction

• For the ’402 Patent:

  • The Term: "competence level"
  • Context and Importance: This term is the central metric for the patent's core function of comparing the processor and the driver to decide who should control the vehicle. Its definition is critical because the infringement case depends on mapping Tesla's driver attentiveness monitoring and system performance limits to this specific claimed concept.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The specification describes the term functionally as the "competence level... in controlling the SDV while the SDV experiences the current operational anomaly" (’402 Patent, col. 2:45-48). This language could support a broad definition covering any factor that affects control, including general attentiveness.
    • Evidence for a Narrower Interpretation: The abstract links the comparison to "a current roadway condition which is a result of current weather conditions," and the specification provides specific examples like a driver's poor night vision or an SDV's handling of rain (’402 Patent, abstract; ’402 Patent, col. 11:21-25). This may support a narrower construction where "competence" relates to specific, assessed skills or capabilities under particular conditions, not just a binary state of being "attentive."

• For the ’765 Patent:

  • The Term: "fault-remediation table"
  • Context and Importance: This term defines a key structural element of the claimed method. Infringement hinges on whether Tesla's use of a neural network for decision-making can be considered a "table." Practitioners may focus on this term because the complaint's "akin to" language suggests a potential mismatch between the claim language and the accused technology.
  • Intrinsic Evidence for Interpretation:
    • Evidence for a Broader Interpretation: The patent does not explicitly define "table." An argument could be made that any system that functionally maps a set of inputs (faults) to a set of outputs (corrective actions) performs the function of a table, regardless of its underlying implementation.
    • Evidence for a Narrower Interpretation: The common specification describes a table where "each row refers to a fault condition, a first column refers to a condition that gets manifested by that fault, and a second column that refers to the mode in which the vehicle should be driven" (’402 Patent, col. 8:29-34). This description of rows and columns may strongly support a construction limited to a structured data lookup format, as opposed to a trained neural network.

VI. Other Allegations

• Indirect Infringement: The complaint alleges that Tesla induces infringement by its customers (end-users) (Compl. ¶168; Compl. ¶181; Compl. ¶194). It alleges Tesla does so with knowledge and intent by manufacturing and selling the Accused Products and providing advertisements, user manuals, and over-the-air software updates that instruct and encourage users to operate the systems in an infringing manner (Compl. ¶170; Compl. ¶183; Compl. ¶196).
• Willful Infringement: The complaint alleges that Tesla’s infringement is willful at least from the filing date of the complaint, thereby establishing post-suit knowledge as a basis for enhanced damages (Compl. ¶171; Compl. ¶184; Compl. ¶197).

VII. Analyst’s Conclusion: Key Questions for the Case

This case will likely involve significant disputes over both claim construction and the technical operation of the accused system. The key questions for the court appear to be:

• A core issue will be one of definitional scope: can the patent's conceptual term "competence level," which is used for a comparative analysis, be construed to read on Tesla's system for monitoring driver "attentiveness" and its own operational safety limits? Similarly, can a complex, trained neural network be construed to meet the limitation of a "fault-remediation table"?
• A key evidentiary question will be one of technical implementation: does the accused FSD system perform the discrete, sequential steps recited in the claims—such as determining a fault, then determining a competence level, then comparing against a threshold, then consulting a table—or does its integrated neural network architecture operate in a fundamentally different way that does not map onto the claimed methods?