Why Autonomous Driving IP Is Entering Its Most Consequential Phase

Annual filings in this corpus climbed steadily from 1,239 in 2017, accelerated sharply through 2019 and 2020 to reach 2,801, dipped briefly in 2021 to 2,524, then hit a definitive peak of 2,824 in 2022. Counts for 2023 and 2024 show visible softening, and 2025 and 2026 figures are heavily under-counted due to standard 18-to-24-month publication lag — they do not represent a genuine decline in activity. The overall lifecycle signal points to a field in maturity: the foundational claim space has been substantially staked, competitive differentiation is now shifting toward prosecution strategy, portfolio quality, and licensing leverage rather than raw filing volume.

The technology composition of the corpus reveals where the real engineering battles are being fought. The B60W (hybrid and automated vehicle control) branch is the single largest concentration with 5,619 patents, reflecting the centrality of decision-making and control logic to the entire AV stack. Digital data processing (G06F) and wireless communication networks (H04W) each contribute more than 3,100 patents, underscoring that autonomous driving is as much a connectivity and compute problem as it is a mechanical one. AI-model computing (G06N) at 3,047 patents and control of non-electric variables (G05D) at 2,901 patents confirm that perception-to-action inference pipelines have become a primary IP battleground. Image data processing (G06T) and image/video recognition (G06V) together add nearly 4,900 records, capturing the camera-based perception layer at the center of the FSD-versus-ADS debate visible on X. Radar, sonar, and positioning (G01S) at 1,806 patents rounds out the sensor-fusion layer. The breadth of the IPC map — spanning communications, AI, computer vision, and vehicle dynamics simultaneously — is itself a signal that no single company can own the full stack without a large, deliberately structured portfolio.

The Core IP Battles: Control Logic, Perception, and Connectivity

Three overlapping sub-battles define where autonomous driving patents are concentrated and contested: vehicle control and decision-making, AI-driven perception, and the wireless infrastructure connecting vehicles to infrastructure and cloud.

Who Holds the Autonomous Driving IP: Applicant Rankings and Momentum

LG Electronics leads the applicant ranking with 4,428 patents, followed closely by Intel at 4,315 — a gap of only 113 filings between the top two. Robert Bosch at 2,473, Nvidia at 2,428, and Panasonic Intellectual Property Corp of America at 2,113 round out the top five. Together, these five account for 37 percent of the hundred largest filers in this corpus. What is strategically notable is that the top two positions are held by consumer-electronics and semiconductor firms rather than vehicle OEMs or dedicated AV startups. LG’s mass derives from V2X and wireless communication patents; Intel’s from data-processing and imaging patents that were partly built up through its Mobileye relationship. This distribution means OEM autonomy programs are navigating a patent landscape largely shaped by component and platform suppliers — creating structural licensing dependencies that go well beyond any single sensor or chip.

Filing momentum tells a divergent story within the top tier. LG Electronics grew recent filings by 18 percent, maintaining its pole position with purpose. Nvidia posted a more modest 4 percent increase, consistent with a focused AI-perception strategy rather than broad-spectrum filing. Toyota Motor Corporation recorded the strongest momentum signal in the ranked set at plus 27 percent — a notable acceleration from a traditional OEM that had previously appeared to be ceding the IP race to tech firms. These three are moving against the trend: Intel contracted by 24 percent in recent filings, Robert Bosch by 21 percent, Waymo by 16 percent, Honda by 35 percent, and Volkswagen by a striking 62 percent. The Volkswagen contraction is particularly sharp and may reflect portfolio rationalization or a strategic shift toward licensing rather than origination, though the evidence for any specific strategic rationale is not available in this dataset.

Co-filing data reveals where applicants are pooling technical development at the patent level. The most active co-filing pair is Volkswagen AG and Audi AG with 22 joint filings, consistent with their shared platform engineering within the same corporate group. Intel and Mobileye Vision Technologies recorded 10 co-filings, reflecting their historical technology-development relationship in camera-based perception. Toyota Motor Corporation and Denso Corp logged 9 co-filings, a pattern typical of OEM-supplier collaboration in vehicle systems. Robert Bosch and Cariad (Volkswagen’s software unit) recorded 6 co-filings, as did Honda Motor and ThyssenKrupp Presta. Volkswagen and Cariad show an additional 5 co-filings, and Volkswagen and Porsche another 5, reinforcing the Volkswagen Group’s internally collaborative — if recently contracting — filing posture. LG Electronics and Sogang University Research Foundation logged 4 co-filings, a university-industry collaboration at the smaller end of the tracked set.

Jurisdictional concentration is pronounced. The United States accounts for 14,073 of the 17,762 total patent records — nearly 79 percent of the corpus. Europe via the EPO is a distant second at 3,477 records, followed by WIPO PCT at 1,555, suggesting substantial international protection-seeking activity at the application stage. The United Kingdom at 548, Australia at 149, and Canada at 125 round out the meaningful filing destinations. Singapore, South Africa, the Philippines, and New Zealand appear at minimal volumes. The US-heavy skew reflects both the primary commercial market for autonomous vehicles and the preferred jurisdiction for asserting foundational platform patents. For freedom-to-operate analysis, the USPTO corpus is where the most consequential claim constructions will occur.

Foundational and most-cited patents

The most-cited filings in scope anchor the field; citation counts accrue over time, so this list favours older, broadly-cited work. Open any row to see it in Eureka.

Strategic Implications for IP and R&D Teams

• Component suppliers, not OEMs, set the IP terms LG Electronics at 4,428 patents and Intel at 4,315 patents collectively outnumber every vehicle OEM in this corpus. Any automotive program building autonomous features must map exposure to the V2X communication and data-processing claim sets held by these platform firms before finalizing architecture decisions.
• Toyota’s momentum signals a strategic gear-shift Toyota’s 27 percent recent-filing growth — the highest among tracked applicants — suggests a deliberate move to strengthen its autonomous driving IP position, concentrated in G05D control systems and B60W vehicle logic. R&D teams benchmarking against Toyota should anticipate a materially stronger patent position in the near term.
• White space in AI, connectivity, and control remains exploitable Despite high absolute filing volumes, the G06F (3,205 patents), H04W (3,184), G06N (3,047), G05D (2,901), and G06T (2,472) branches each carry white-space share estimates of six to eight percent, indicating that specific claim niches within these large categories remain underprotected relative to their commercial importance — particularly for teams focused on edge-case inference and adaptive control.
• Contracting portfolios create acquisition and licensing opportunities Volkswagen’s 62 percent contraction in recent filings and Intel’s 24 percent pullback may signal portfolios available for licensing or acquisition. IP strategy teams tracking distressed or rationalizing holders in a maturing field should monitor these positions closely.