Current Status and Key Technologies of Autonomous Driving

1. How Silicon Valley VCs Are Reacting — "It Has Reached Escape Velocity"

In early 2026, Silicon Valley venture capital made a clear pivot toward the autonomous driving sector, shifting away from the cautious stance that had prevailed through 2024. The defining symbol of this shift was Waymo's $16 billion (approximately ¥2.4 trillion) round in February 2026.

Sequoia Capital served as lead investor alongside Dragoneer Investment Group and DST Global. Sequoia partner Konstantine Buhler praised Waymo's "operational excellence" and its track record of tripling paid ride counts in a single year. Sequoia has investments in Waymo, Nuro, and Applied Intuition within the autonomous driving space, building billion-dollar positions across both full-stack AV companies and sensor technology.

Andreessen Horowitz (a16z) also participated in Waymo's $16 billion round and is sharpening its focus on AV software, infrastructure, and platforms. While it previously invested in Cruise, following GM's withdrawal it has shifted its focus to Waymo and Applied Intuition.

Kleiner Perkins co-led Applied Intuition's $600 million (approximately ¥90 billion) Series F with BlackRock and also participated in the Waymo round, making clear its concentrated investment posture toward next-generation autonomous driving systems.

SoftBank Vision Fund led UK-based Wayve's $1.05 billion Series C and participated in its $1.2 billion (approximately ¥180 billion) Series D, maintaining its position as one of the largest institutional investors backing autonomous driving globally.

The prevailing consensus in the VC community is that autonomous driving has "crossed the valley of death and reached escape velocity." Three concrete data points have dispelled doubts about technical feasibility: Waymo's more than 500,000 paid rides per week, Tesla's commercialization of its robotaxi service, and Aurora Innovation's commercial operation of driverless trucks. Goldman Sachs projects that the North American robotaxi market will generate approximately $7 billion (roughly ¥1.05 trillion) in annual revenue by 2030, accounting for roughly 8% of the ride-sharing market.

---

2. Current Position of Key Players

2.1 Waymo (Alphabet/Google) — Overwhelming Scale and Safety Record

Waymo is the clear leader in the autonomous driving industry as of early 2026.

• Driving scale: Achieved over 500,000 paid rides per week across 10 U.S. cities. Ten-fold growth from 50,000 rides per week in May 2024. Provided approximately 15 million trips in full-year 2025, triple the prior year. Cumulative ride count exceeds 20 million.
• Vehicles: Operates approximately 3,000 robotaxis. Announced sixth-generation hardware in February 2026 (13 cameras, 6 radars, 4 LiDARs).
• Deployed cities: Currently operating in 10 cities including San Francisco, Los Angeles, Phoenix, Austin, Atlanta, Dallas, and Houston. Plans to expand to 26 cities by end of 2026, including Denver, Miami, Nashville, Seattle, and Washington D.C.
• International expansion: Launched mapping and test drives in London with 24 vehicles. Targeting a pilot in April 2026 and commercial launch in September 2026. Conducting mapping operations in Tokyo in partnership with Nihon Kotsu.
• New vehicle platforms: Geely-made "Ojai" minivan (planned for commercial launch in the second half of 2026) and Hyundai Ioniq 5 (supply agreement for up to 50,000 units under negotiation).
• Valuation: Post-money valuation of $126 billion (approx. ¥18.9 trillion) following a $16 billion fundraise. More than doubled from Series C in October 2024 ($5.6 billion raised, $45 billion valuation).
• Safety: Serious injury accident rate 1/11th that of human drivers. Over 100 million cumulative miles driven at Level 4.

Key investors: Alphabet (approx. $13 billion invested), Sequoia Capital, Dragoneer, DST Global, a16z, Silver Lake, Tiger Global, T. Rowe Price, Fidelity, Kleiner Perkins, Temasek, Mubadala Capital

2.2 Tesla — Mass Production of Cybercab and Rapid Scaling

Tesla launched a limited robotaxi service in Austin in June 2025, sending shockwaves through the industry with its cameras-only (no LiDAR) approach.

• Current status: Approximately 31 Model Y vehicles operating in Austin at $4.20 per ride (approx. ¥630). Safety monitors ride in the passenger seat, though limited fully driverless operation began in January 2026.
• Service area: 243 square miles in Austin (2.7 times Waymo's 90 square miles).
• Cybercab: A dedicated robotaxi with no steering wheel or pedals. The first unit rolled off the line at Gigafactory Texas in February 2026. Mass production began in April 2026.
• FSD v13/v14: Introduced the Temporal-Voxel transformer model, marking an architectural shift from "reactive" to "predictive." FSD v14 is assessed to have reached "near Level 4" thresholds.
• Cumulative miles: Over 7 billion miles (approx. 11.2 billion km) of driving data accumulated across all FSD.
• 2026 rollout plan: Expanding to Dallas, Houston, Phoenix, Miami, Las Vegas, and San Francisco in H1. Public ride-hailing app beta release planned for late Q3–Q4 2026.
• Challenges: Early incidents (wrong-way driving, phantom braking, traffic violations) reported and under NHTSA investigation. Persistent criticism of the limitations of the cameras-only approach.

2.3 Zoox (Amazon) — Differentiation Through a Purpose-Built Vehicle

• Vehicle: A purpose-built, bidirectional robotaxi with no steering wheel or pedals. Approximately 50 units in operation.
• Current status: Offering free public rides on the Las Vegas Strip and in San Francisco.
• 2026 plans: Transitioning to paid service in San Francisco and Las Vegas. Expanding testing to Austin, Miami, Dallas, and Phoenix.
• Uber partnership: Zoox robotaxis to be available on the Uber app in Las Vegas starting summer 2026, expanding to LA by mid-2027.
• FMVSS exemption: Application pending for commercial exemption of 2,500 vehicles.

2.4 Aurora Innovation — Pioneer in Autonomous Trucking

• Historic milestone: Launched the world's first commercial fully driverless Class 8 truck operations on the Dallas–Houston corridor in April/May 2025.
• Track record: Achieved over 250,000 miles of incident-free driverless operation.
• Route expansion: Expanded to 10 routes spanning Texas, New Mexico, and Arizona.
• 2026 goals: Deploy more than 200 driverless trucks by year-end. Planned removal of partner-requested safety monitors in Q2.
• Next-generation hardware: 50% cost reduction; FirstLight LiDAR detection range doubled to 1,000 m (to be introduced mid-2026).

2.5 Cruise (formerly GM Cruise) — Withdrawal and Restructuring

• GM halted funding for robotaxi development in December 2024. Cruise was integrated into GM in February 2025, ending the robotaxi business.
• Annual cost reduction of approximately $2 billion (approx. ¥300 billion).
• Technology repurposed for Super Cruise ADAS (deployed in more than 20 GM vehicle models).
• Microsoft recorded an $800 million (approx. ¥120 billion) impairment on its Cruise investment (Q2 2025).
• Hired Sterling Anderson, former head of Tesla Autopilot, to focus on ADAS enhancement.

2.6 Chinese Players — Baidu, Pony.ai, WeRide

Baidu Apollo Go:

• World's largest robotaxi fleet. Over 1,000 fully driverless vehicles in Wuhan alone.
• Operating in more than 15 Chinese cities. Over 250,000 fully driverless rides per week (as of November 2025).
• Cumulative ride count exceeds 17 million.
• Sixth-generation vehicle cost below $30,000 (approx. ¥4.5 million); seventh-generation targeting below $20,000 (approx. ¥3 million).
• Serious incident (March 31, 2026): Over 100 Apollo Go vehicles in Wuhan simultaneously halted due to a system failure. Passengers were stranded on elevated roads for up to two hours, and collisions occurred. Reports attributed the cause to a cloud service outage. The incident raised fundamental safety concerns about large-scale driverless operation.

Pony.ai:

• NASDAQ IPO in November 2024 ($260 million raised); dual-listed on the Hong Kong Stock Exchange in November 2025 (raising up to HK$7.7 billion, approx. ¥99 billion) — the largest autonomous driving IPO of 2025.
• The only company operating fully driverless robotaxis in all four major Chinese cities (Beijing, Shanghai, Guangzhou, and Shenzhen).
• Achieved unit economics breakeven across entire city coverage with Gen-7 robotaxi.
• Targeting a global fleet of 3,000+ vehicles by end of 2026 (three times current size).
• Market capitalization of approximately $7.07 billion (approx. ¥1.06 trillion).

WeRide:

• NASDAQ listing in October 2024 (valuation: $5.3 billion, approx. ¥795 billion); Hong Kong dual listing in November 2025 (HK$2.39 billion raised).
• Holds operating permits in 6 countries (China, Singapore, Saudi Arabia, UAE, France, and the United States).
• Plans to launch fully driverless public service in Switzerland in H1 2026.
• Partnership with Grab to prepare Level 4 deployment in Southeast Asia.

---

3. The Cutting Edge of Component Technologies

3.1 LiDAR — Chinese Dominance and Price Disruption

The automotive LiDAR market reached $1.26 billion in 2025 and is projected to grow to $25.75 billion by 2035 (CAGR 35.16%). Most notably, four Chinese companies (Hesai, RoboSense, Huawei, and Seyond) control 89% of the global market.

Hesai Technology (禾赛科技):

• World's #1 automotive LiDAR manufacturer for four consecutive years (Yole Group research)
• Achieved the world's first annual production of 1 million units in 2025; cumulative shipments surpassed 2 million units
• ADAS shipments grew 203% year-over-year to 1.38 million units
• Infinity Eye Platform (announced April 2025):

- Infinity Eye A (for L4): 360-degree blind-spot-free coverage via 4× AT1440 ultra-high-definition LiDARs + 4× FTX solid-state LiDARs. The AT1440 features the world's highest 1,440 channels, over 34 million points per second (45× mainstream products), and 300 m detection at 10% reflectivity

- Infinity Eye B (for L3): ETX ultra-long-range LiDAR. 800 channels, 400 m detection at 10% reflectivity, ultra-high resolution of 0.05°×0.025°, 95% reduction in false detections in rain and fog. Mass production planned for 2026

- Infinity Eye C (for L2): ATX compact LiDAR. 256 channels, 200 m detection. Adopted by 11 automakers including BYD, Chery, and Great Wall Motors

• Annual production capacity to double to 4 million units in 2026; new factory under construction in Bangkok (operational early 2027)
• Average selling price in the Chinese market: $450–$500; targeting below $200

RoboSense (速腾聚创):

• World's #1 in shipment volume in 2024 (544,000 units); monthly shipments reached 120,000 units in October 2025
• Achieved first-ever quarterly profit in Q4 2025 (net profit of CNY 103.7 million ≈ $14.3 million)
• EM Platform: EM4 is the only mass-produced LiDAR with 500+ beams (customizable from 520 to 2,160 beams); 300 m @ 10% reflectivity, up to 600 m detection
• Robotics LiDAR revenue surged 1,142% year-over-year (303,000 units)
• Annual production capacity to expand to 4 million units in 2026

Luminar Technologies — Under Pressure:

• Volvo announced it will discontinue use of Luminar Iris by end of 2025 (removed from EX90 and ES90 models from April 2026 onward), citing "failure to meet contractual obligations"
• Q3 2025 revenue was just $18.7 million; full-year guidance revised down to $67–$74 million
• Concerns about potential bankruptcy have been reported — a symbol of the decline of the U.S. LiDAR industry

Innoviz Technologies:

• InnovizTwo: angular resolution of 0.05°×0.05°, 300 m detection range
• Nine InnovizTwo units per vehicle in the VW ID. Buzz AD, providing 360-degree Level 4 coverage
• Supplying InnovizOne for the BMW i7 to enable Level 3 functionality (from March 2024)

3.2 End-to-End AI — A Paradigm Shift

Between 2024 and 2025, autonomous driving AI underwent a fundamental transition from modular, rule-based systems to end-to-end neural network approaches.

Tesla FSD Architecture:

• FSD v12 replaced over 300,000 lines of hand-coded rules with an end-to-end neural network; 48 individual neural networks process input from 8 cameras
• FSD v13 introduced a Temporal-Voxel transformer model, prioritizing long-term memory and transforming the vehicle from "reactive" to "predictive"
• FSD v14 advances further into sophisticated end-to-end AI

Waymo EMMA (End-to-End Multimodal Model for Autonomous Driving):

• Built on Google's Gemini multimodal LLM
• Generates driving outputs (trajectories, recognized objects, road graph elements) directly from raw camera sensor data
• Chain-of-thought reasoning improves end-to-end planning by 6.7%
• State-of-the-art results on the nuScenes planning benchmark; competitive performance on the Waymo Open Motion Dataset
• Direct processing of LiDAR/radar data not yet supported

NVIDIA Alpamayo (announced at CES 2026):

• Alpamayo R1: The first open-source inference VLA (Vision-Language-Action) model for autonomous driving. 10 billion parameters, chain-of-thought reasoning
• Enables AVs to "think like humans," resolving previously unseen edge cases
• Three pillars: open model, AlpaSim simulation framework, open dataset (over 1,700 hours of driving data)
• First deployed in the Mercedes-Benz CLA (NVIDIA DRIVE full-stack platform, driving on U.S. roads in 2026)

VLA (Vision-Language-Action) Paradigm:

• Integrates large language models into driving architectures to build multimodal Vision-Language-Action systems
• Li Auto's MindVLA: "mapless NOA" deployed fleet-wide in China

3.3 Compute Platforms — Exponential Growth in Processing Power

NVIDIA DRIVE:

• Orin (current): 254 TOPS; dominant in the L2/L3 market
• Thor (next-gen): 2,000 TOPS (FP8); features an Inference Transformer Engine that accelerates transformer DNN inference by 9×; integrates cockpit, ADAS, and parking into a single chip. Zeekr is the first customer; mass production begins 2025–2026
• NDAS (NVIDIA DRIVE AV Solution): V1 launched April 2025; dual-Thor L3 highway version in Q1 2027; urban L3 version in the second half of 2027

Qualcomm Snapdragon Ride:

• Snapdragon Ride Flex: Integrates digital cockpit and ADAS into a single SoC, reducing costs by 25–50%
• 20+ ADAS/AD projects from GM, Ford, BMW, Honda, Mercedes-Benz, VW, and Geely launching in 2025–2026

Mobileye EyeQ Ultra:

• 176 TOPS AI accelerator, 12 RISC-V cores; performance equivalent to 10× EyeQ5 in a single package
• Level 4 capable at under $1,000
• Announced acquisition of Mentee Robotics at CES 2026 (expanding autonomous driving perception technology to general physical AI/robotics)

Horizon Robotics Journey 6 (地平线机器人):

• Journey 6P: 560 TOPS, 37 billion transistors, TSMC 7nm
• Cumulative shipments exceeded 10 million chips (as of August 2025)
• Adopted by 20+ automakers including BYD, Li Auto, SAIC, GAC, and VW (co-developed for the Chinese market)

Tesla AI5:

• Approximately 40× faster than AI4, with 8× raw compute, 9× memory capacity, and 5× memory bandwidth; up to 800W
• Manufactured by TSMC; mass production delayed to late 2026–early 2027; first deployment planned for Cybercab

3.4 4D Imaging Radar — Complementing and Replacing LiDAR

4D imaging radar is an emerging sensor technology that adds vertical height detection to conventional radar (range, velocity, azimuth).

• Market size: $2.75 billion in 2025, projected to grow to $5.1 billion by 2030
• Continental ARS540: 4D imaging radar capable of mapping up to 300 m; deployed in the Hyundai IONIQ 5
• ZF: Launched a 4D full-range radar for China's SaicAI (5.3% global market share)
• Rapidly gaining adoption as a replacement or complementary sensor to LiDAR in cost-sensitive applications

3.5 Sensor Fusion — Evolution of Integration

In 2025–2026, sensor fusion evolved "from experimental stage to core infrastructure."

Key Architectures:

1. Unified BEV (Bird's Eye View) Representation: Camera images and LiDAR point clouds are converted into a common BEV coordinate system and fused, integrating the semantic information from cameras with the geometric information from LiDAR

2. Token-Level Cross-Modal Alignment: Transformer-based approach for aligning features across sensor modalities

Fusion Strategies:

• Early Fusion (sensor level): Combines raw data before processing
• Mid-Level Fusion (feature level): Most common in mass production; combines extracted features
• Late Fusion (decision level): Combines independent modality outputs
• Dynamic Fusion (emerging in 2025): FDSNet measures semantic consistency across sensor modalities and adaptively selects the fusion stage

Camera vs. LiDAR Debate: Two camps remain in opposition — Tesla's camera-only approach versus the multi-sensor fusion adopted by Waymo, Chinese OEMs, and European manufacturers. However, as of 2026, virtually all major players except Tesla have adopted multi-sensor fusion.

3.6 The Shift from HD Maps to Mapless

The HD map market stands at $1.09 billion in 2025, but "mapless NOA (Navigate on Autopilot)" is emerging as the mainstream solution, bringing structural transformation to the industry.

Three Approaches to the Mapless Revolution:

1. Segmentation-Based: Generates semantic BEV maps through a decoder and clusters them into structural elements

2. Detection-Based (DETR-style): Directly detects map elements end-to-end from BEV features

3. Tracking-Based: Uses a memory buffer that aggregates past BEV features to ensure temporal consistency

Major Chinese OEMs including Li Auto, XPeng, Huawei AITO, and GAC Aion have adopted the mapless approach. Mobileye maintains differentiation through its crowdsourced REM (Road Experience Management) system.

3.7 V2X (Vehicle-to-Everything) Communication

With the rollout of 5G, C-V2X is becoming established as the global standard.

Regional Deployment Status:

• U.S.: Atlanta became the nation's first "Day One Deployment District" in September 2025 (Georgia DOT + 5GAA)
• China: Dedicated C-V2X frequency band (5905–5925 MHz) established; "vehicle-road-cloud integration" pilots in 20 cities (2024–2026); plans for C-V2X integration in half of all new vehicles by 2025
• Europe: Basic V2X use cases from 2026; advanced use cases from 2029
• 3GPP Release 18: Sidelink enhancements for V2X; LTE-V2X and NR-V2X coexistence
• Vision for 6G integration is also under development

3.8 Simulation Platforms

Simulation is the "hidden killer app" of autonomous driving development and is attracting significant VC attention.

Applied Intuition:

• Valuation of $15 billion; raised $600 million in June 2025 co-led by BlackRock and Kleiner Perkins (Series F); valuation increased 2.5× in one year
• Serves 18 of the world's top 20 automakers
• Executed over 50 million simulations covering billions of miles of driving in 2025
• Expanding into the defense sector

NVIDIA DRIVE Sim + Omniverse:

• Cloud-based, high-fidelity simulation; Omniverse Sensor RTX generates physically accurate camera, LiDAR, and radar datasets
• AlpaSim: Open-source simulation framework for the Alpamayo ecosystem

The overall driving simulation software market is projected to reach $13.6 billion in 2025 and $26.3 billion in 2030.

---

4. Coverage and Analysis by Each Media Outlet and Institution

4.1 U.S. Media

CNBC has positioned 2026 as "the year of the robotaxi," providing detailed coverage of the intensifying competition centered on four companies — Waymo, Tesla, Zoox, and Baidu. It notably juxtaposed Waymo's $16 billion funding round against Baidu's Wuhan system outage, highlighting the "trade-off between scale and safety."

TechCrunch praised Wayve's $1.2 billion Series D as "a victory for the AI-first approach to autonomous driving," while spotlighting Applied Intuition's rapid growth as the "shovel business of autonomous driving" (a reference to selling shovels during the Gold Rush).

The Information analyzed Waymo's internal economics, reporting that its estimated annual revenue run rate has exceeded $350–400 million (approximately ¥52.5–60 billion).

4.2 European Media & Institutions

DIGITALEUROPE published "Embracing the Future of Mobility: A Strategy for Autonomous Driving in the EU" in May 2025, recommending the development of a regulatory framework for Europe.

Allianz analyzed safety data from an insurance industry perspective in its report "HANDS OFF — The Safety Promise of Autonomous Driving," reporting that between 2021 and 2025, there were zero recorded human fatalities caused by ADS (Automated Driving System) vehicles (only one on record, representing 0.1% of all accidents).

The EU legalized Level 3 highway driving at speeds up to 130 km/h effective January 1, 2026. Automated lane changes via ALKS (Automated Lane Keeping Systems) are also permitted, subject to the condition of 360-degree sensor redundancy.

4.3 Japanese Media & Institutions

The Nikkei gave prominent coverage to Turing's Series A (¥15.27 billion), highlighting Japan's unique homegrown approach to end-to-end autonomous driving via multimodal generative AI.

The Ministry of Economy, Trade and Industry (METI) published ISO 23792-1:2026 and ISO 23792-2:2026 in March 2026 — international standards for autonomous driving system requirements and lane-change testing methods, developed under Japanese leadership.

The National Public Safety Commission: The April 2023 revision to the Road Traffic Act already introduced Level 4 (specific automated driving operations) as a permit-based system, applicable to remotely monitored transportation services for buses, taxis, and trucks.

4.4 Analyst Reports

Goldman Sachs Research:

• The robotaxi market is projected to grow at a CAGR of approximately 90% from 2025 to 2030
• In 2026 new vehicle sales, only 1–2% will include Level 3 capabilities; by 2030, 10% of new vehicles will feature Level 3 and approximately 2.5% Level 4
• Gross margins for vertically integrated AV operators are forecast to reach 40–50% within 3–5 years
• Total gross profit of the U.S. AV market: approximately $3.5 billion (approx. ¥525 billion) in 2030

Morgan Stanley Research:

• Adoption rate of partial-to-full autonomous driving: 8% (2024) → 28% (2030)
• Market opportunity: $200 billion (approx. ¥30 trillion) in 2030; $300–400 billion (approx. ¥45–60 trillion) in 2035
• China: approximately 60% of passenger vehicles to feature smart driving by 2030 (approximately 15 million units)

McKinsey Center for Future Mobility (survey of 91 decision-makers, 2025):

• ADAS/AD technology to contribute $300–400 billion (approx. ¥45–60 trillion) to the passenger vehicle market by 2035
• Large-scale global robotaxi deployment projected for 2030 (revised from 2029)
• Current state: more than 700,000 fully automated robotaxi rides per week worldwide (over 450,000 in the U.S., over 250,000 in China)

S&P Global Mobility — Projected U.S. road composition in 2035: approximately 87 million Level 2 vehicles, approximately 45 million Level 2+, approximately 3.5 million Level 3, and approximately 1.7 million Level 4

5. Latest Developments in the Regulatory Environment

5.1 United States

Federal Level:

• April 2025: Transportation Secretary Sean Duffy directed NHTSA to establish a regulatory framework based on three principles: "safety first, unleash innovation, and promote commercial deployment"
• AV STEP Program: A voluntary framework that streamlines the exemption process from FMVSS (Federal Motor Vehicle Safety Standards). First exemption issued in August 2025
• September 2025: Three rulemakings launched to modernize FMVSS (for vehicles without traditional operating controls)
• March 2026: NHTSA proposes amendments to collision avoidance standards for autonomous vehicles
• Four major bills (2025): AV Accessibility Act, AV Safety Data Act, Autonomous Vehicle Acceleration Act, AMERICA DRIVES Act

State Level: In just the first few months of 2025, 25 states introduced 67 new AV-related bills. California mandates a $5 million insurance bond, while Texas is rapidly emerging as a hub for AV trucking.

5.2 European Union

• January 1, 2026: Level 3 highway driving at up to 130 km/h legalized through amendments to UN Regulation No. 157 and harmonization of EU road traffic laws
• DSSAD (Data Storage System for Autonomous Driving) mandated for all Level 3 vehicles — timestamped logs of system activation, takeover requests, and driver operations accessible to insurers and law enforcement
• Regulatory sandboxes and autonomous driving corridors to be established in 2026

5.3 Japan

• Level 4 Permit System: Already introduced as "specified autonomous driving operations" under the amended Road Traffic Act enacted in April 2023
• RoAD to the L4 Project: Aims to achieve social implementation of Level 4 autonomous trucks from fiscal year 2026 onward
• New Form of Capitalism Implementation Plan 2025: Targets deployment of unmanned autonomous mobility services at 100+ locations by fiscal year 2027 (brought forward from the previous fiscal year 2030 target)
• July 2025: Commercial line-haul autonomous truck operations launched between Kanto and Kansai regions (Level 2, targeting Level 4 by 2027)

5.4 China

• BAIC Arcfox a-S and Changan Deepal SL03 obtained China's first Level 3 road test approvals by end of 2025
• February 2026: MIIT published a draft mandatory safety standard for advanced autonomous driving (scheduled to take effect July 2027). Requires Level 3 systems to independently execute minimum risk maneuvers when the driver is unresponsive — effectively bringing L3 closer to L4 in regulatory terms
• "Vehicle-Road-Cloud Integration" pilot projects underway in 20 cities

6. Japan's Autonomous Driving Ecosystem

6.1 Tier IV

A leading Japanese autonomous driving startup and developer of "Autoware," the world's largest open-source autonomous driving software.

• Announced an end-to-end Level 4 architecture in July 2025; deployed in over 50 locations nationwide
• Selected by NEDO to develop a large-scale AI data platform for autonomous driving
• Partnering with Isuzu to deploy Level 4 autonomous buses equipped with the NVIDIA DRIVE Hyperion platform
• Participating in the Ministry of Defense's unmanned security vehicle program
• Launching public-sector shuttles using Autoware and a robotaxi reference design in November 2025

6.2 Turing

• Raised ¥15.27 billion in Series A funding in November 2025 (co-led by JIC Venture Growth Investments and Global Brain)
• Received an additional ¥3.2 billion from GMO Internet in February 2026; cumulative fundraising for FY2025 totals approximately ¥24 billion (~$160 million USD)
• A Japan-based startup developing end-to-end autonomous driving powered by multimodal generative AI

6.3 Honda

• Originally planned a driverless ride-hailing service in Tokyo using the Cruise Origin through a joint venture with GM/Cruise (targeting early 2026), but revised plans following GM's withdrawal
• Shifted to a new approach: partnering with existing taxi operators (Kokusai Motorcars, Teito Motor Transportation) to deploy AV taxis on Tokyo expressways

6.4 Toyota / Woven City

• Woven City: Officially opened at the foot of Mt. Fuji on September 25, 2025; Phase 1 construction complete
• Approximately 100 initial residents (Toyota/WbyT staff and their families); public participation scheduled from FY2026 onward
• The e-Palette autonomous electric vehicle platform serves as the primary mode of transport; an underground autonomous logistics network is also in place
• e-Palette EV Minibus: Now available for external sale (manual driving mode, ¥29 million); Level 4 system deployment planned for FY2027

6.5 BOLDLY (formerly SB Drive / SoftBank subsidiary)

• Yokohama Pilot (November 2025 – January 2026): Autonomous driving operation with five Nissan Serena vehicles in the Minato Mirai district, with remote monitoring center "PLOT48" in operation
• Target for commercialization: FY2027 (April 2027)

---

7. Investment Round List (Major Deals, 2025–2026)

---

8. New Developments in Insurance and Liability Frameworks

With the commercialization of autonomous vehicles, the insurance and liability framework is being rapidly restructured.

Liability Shift Model:

At Levels 3–5, responsibility for accidents occurring while the system is in operation shifts from the driver to OEMs/software providers. Four major regimes (driver liability, system liability, manufacturer/operator liability, and mixed liability) are emerging globally.

Regional Trends:

• United States: California is considering raising minimum liability coverage requirements. NHTSA's AV framework simplifies incident reporting to a 5-day window.
• EU: With mandatory DSSAD requirements, manufacturer liability applies when the system is operating within the ODD (Operational Design Domain). France mandates black box recorders for all autonomous vehicles (from 2025).
• Germany: Compulsory liability insurance for Level 4 and above vehicles takes effect in 2025.

Safety Data (as of November 2025):

• 5,202 AV accidents reported in the United States. 451 injuries, 65 fatalities.
• However, zero human deaths caused by ADS (Automated Driving System) vehicles from 2021 to 2025 (the sole recorded case, 0.1% of all accidents).
• Waymo: 1,429 accidents reported (July 2021 – November 2025). 117 injuries, 2 fatalities.

---

9. Future Outlook — When and What Will Happen

Within 2026 (High Confidence)

• April 2026: Tesla Cybercab mass production begins
• Q2 2026: Aurora Innovation eliminates partner-required monitors and moves to fully driverless truck operations
• Summer 2026: Zoox × Uber commercial robotaxi service launches in Las Vegas
• Late 2026: Waymo's new "Ojai" robotaxi enters commercial deployment. London expansion (Waymo's first international rollout)
• End of 2026: Waymo targets 1 million weekly trips. Aurora deploys 200+ trucks. WeRide launches fully driverless service in Switzerland

2027 (Medium Confidence)

• Q1 2027: NVIDIA DRIVE dual Thor Level 3 highway variant goes on sale
• 2027: Tesla AI5 chip mass production begins. Japan targets commercialization of Level 4 autonomous trucks
• FY2027: Japan targets 100+ locations for unmanned autonomous mobility services. Toyota e-Palette Level 4 deployment
• Mid-2027: Zoox × Uber expands to LA

2028–2030 (Medium–Low Confidence)

• 2029: Pony.ai targets profitability
• 2030: McKinsey projection of large-scale global robotaxi deployment. Goldman Sachs projects North American AV market reaches $7 billion annual revenue. 10% of new vehicles equipped with Level 3 (Goldman Sachs)
• 2030: Morgan Stanley projects AV market reaches $200 billion (approx. ¥30 trillion)

2035 and Beyond

• Morgan Stanley projection: AV market reaches $300–400 billion (approx. ¥45–60 trillion)
• McKinsey projection: ADAS/AD contributes $300–400 billion to the passenger vehicle market
• S&P Global: Approximately 1.7 million Level 4 vehicles on U.S. roads

10. Investment Themes and Areas of Focus from a VC Perspective

When consolidating the investment themes in the autonomous driving sector that Silicon Valley's VC community is watching, the following picture emerges.

1. Concentrated Investment in Platform Players:

As Waymo's $126 billion valuation demonstrates, capital is concentrating on players capable of operating full-stack services under a "winner-takes-all" logic. With Cruise's exit and Motional's weakening, the U.S. market has effectively narrowed to three players: Waymo, Tesla, and Zoox.

2. High Valuations for "Shovel Businesses":

Tool and infrastructure companies like Applied Intuition ($15 billion) are growing rapidly. The autonomous driving "shovel business" (selling shovels during the gold rush) is valued as a more risk-diversified investment than individual robotaxi operators.

3. AI-First vs. Legacy Approaches:

Companies pursuing "no HD maps, end-to-end AI" approaches—such as Wayve (valued at $8.6 billion, AI-first)—tend to command higher valuations than traditional modular-approach companies. Interest in the VLA (Vision-Language-Action) paradigm is rising sharply.

4. Commercialization of Autonomous Trucking:

As demonstrated by Aurora Innovation (commercial operations launched), Kodiak Robotics (SPAC listing), and Gatik (fully driverless delivery launched for Fortune 50 retailers), autonomous trucking is achieving commercial viability ahead of robotaxis.

5. China Risk and Decoupling:

The reality that four Chinese companies control 89% of the LiDAR market is recognized as a supply chain risk. Luminar's decline has exposed the vulnerability of the U.S. LiDAR industry. Meanwhile, the Nasdaq listings of Pony.ai and WeRide are drawing attention as "indirect investment routes into Chinese AD companies."

6. Uber's Platform Strategy:

Uber's strategy of not developing its own AVs but instead becoming a robotaxi dispatch platform through partnerships with Waymo, Zoox, Wayve, Nuro, Momenta, and others is being evaluated by VCs as "the AV position with the highest risk-adjusted returns."

---

11. US-China Comparison: Two Autonomous Driving Ecosystems

China leads the U.S. in miles driven and LiDAR supply chain dominance, while the U.S. (Waymo) holds the advantage in valuation and weekly ride counts. However, the large-scale system failure in Wuhan starkly illustrated the risks of rapid scaling and sent a sobering signal across the entire industry.

---

Impact on the Industry

By 2026, autonomous driving has fully shifted from the question of "is it technically possible?" to "how can it scale safely and sustainably?" Waymo's $126 billion valuation signals the company's transformation from Alphabet's "experimental project" into an independent tech giant.

The large-scale outage of Baidu Apollo Go in Wuhan exposed the vulnerabilities of cloud-dependent architectures and renewed awareness of the importance of edge computing. Advances in edge processing power — such as NVIDIA's Thor chip (2,000 TOPS) and the Mobileye EyeQ Ultra — are key to safety and reliability.

Chinese companies' dominance of the LiDAR market (89%) highlights the geopolitical dimension of autonomous driving. Western AV manufacturers are growing increasingly reliant on Chinese-made LiDAR, a situation made more precarious by Luminar's decline. This represents a structural risk analogous to TSMC dependence in semiconductors, and VCs are watching "LiDAR supply chain diversification" as the next investment theme.

Japan has positioned 2025 as a pivotal year of transition "from demonstration to implementation," accelerating the societal deployment of Level 4 autonomy. The open-source ecosystem centered on Tier IV's Autoware, along with Turing's end-to-end approach, could become Japan's unique sources of technological differentiation. Whether Japan achieves its target of "unmanned autonomous driving at 100 locations by fiscal year 2027" will determine its international competitiveness in the autonomous driving industry.

---

References: Waymo official blog (various announcements, 2025/2026), Goldman Sachs Research (AV market forecast), Morgan Stanley Research (autonomous driving industry growth analysis), McKinsey Center for Future Mobility (2025 survey report), CNBC, TechCrunch, and The Information various reports, NHTSA AV framework, DIGITALEUROPE autonomous driving strategy recommendations, Ministry of Economy, Trade and Industry ISO 23792 announcement, Hesai Technology and RoboSense various IR materials, Applied Intuition official blog, NVIDIA CES 2026 announcement, Mobileye CES 2026 press conference, PitchBook and CB Insights investment data, Yole Group LiDAR market analysis, 5GAA C-V2X roadmap, S&P Global Mobility forecast, Allianz Motor Day 2025 report, World Economic Forum autonomous driving roadmap (April 2025), IDTechEx autonomous driving software market report (2026)