1. introduction: The Oppenheimer Moment in the Mobility Industry
in 2025, the global automotive industry is at a crossroads of opportunity and existential crisis that goes beyond a simple shift in powertrain (from internal combustion engines to electric vehicles). just 100 years after humans shifted from the horse to the car (Ford Model T), we are witnessing a massive civilizational shift in which the very act of driving is being delegated to machines. at the center of this shift is the robotaxi. it's not just a driverless mode of transportation; it's the culmination of a "moving artificial intelligence (AI) revolution" that is reshaping the spatial fabric of cities, revolutionizing energy consumption patterns, shaking up the job market for millions, and even determining the technological hegemony of nations.
on the foggy hillsides of San Francisco, Waymo's Jaguar I-PACEs carry passengers without a driver, making more than 250,000 paid trips a week. meanwhile, in Texas, Elon Musk unveiled the Cybercab, a car that completely removes the steering wheel and pedals, revealing his ambition to disrupt the global transportation market for less than $30,000. where does the Korean market stand now that mobility leaders such as the U.S. and China have moved beyond the commercialization stage of autonomous driving and are in a race for 'economies of scale-up'?
despite having one of the world's leading automakers (Hyundai Motor Group) and tech companies (Naver, Kakao), South Korea is stuck in a triple bind: the regulatory trauma of the 'No Ride' law, the archaic legacy of the taxi licensing system, and a rapidly aging population. the Bank of Korea's recent report proposing a drastic restructuring of the taxi licensing system, including the forced purchase of taxi licenses, suggests that the country's mobility ecosystem has reached a critical mass that can no longer be delayed.
this report analyzes the technological and business strategies of Waymo and Tesla, the two major players in the global robo-taxi market, under a microscope, and dissects the structural dilemma facing South Korea in the midst of this battle. furthermore, the strategic pivot of Hyundai Motor Group and the government's policy responses will be analyzed in detail to provide a concrete roadmap for South Korea to make a soft landing without losing its technological sovereignty in the coming era of fully autonomous driving.
2. the Great Clash of Tech Philosophies: Deterministic Safety vs. Probabilistic Scaling
the race for robotaxi supremacy is not just a battle for market share between companies. it's a giant experiment and war between two different technological philosophies for enabling autonomous driving: Waymo's multi-sensor fusion-based deterministic approach and Tesla's vision-only probabilistic approach.
2.1 Waymo: eyes of hardware and brains of maps
waymo, a subsidiary of Alphabet, has made safety a non-negotiable priority since the days of Google's self-driving project. their strategy centers on creating a controllable environment through expensive sensor equipment that transcends human cognitive capabilities and pre-built HD maps that are perfect for the task.
2.1.1 Sensor Fusion: the Aesthetics of Over-engineering
waymo's sixth-generation autonomous driving system uses sensor fusion technology, which combines LiDAR, radar, cameras, and even audio sensors.
LiDAR's role: it fires a laser to scan the three-dimensional shape of the surrounding environment with millimeter precision. this allows it to measure the exact distance to objects, even at night or in backlit conditions. elon Musk has dismissed it as an "expensive and unnecessary crutch," but for Waymo, lidar is a key safety feature that compensates for human visual limitations.
cost dilemma and innovation: Early self-driving car sensor suites cost as much as $75,000 per unit, but costs have dropped dramatically as the technology has matured. by 2024, Weimo's sensor suite is estimated to cost around $12,700, and with the advent of low-cost lidar from China, this price is coming down even faster. this proves that Weimo is no longer a 'lab technology' but a commercial product that can be mass-produced.
2.1.2 The paradox of geo-fencing and HD maps
weimo's vehicles drive on high-definition maps (HD Maps) that are scanned in centimeters (cm) in advance. the vehicles determine their location not by GPS, but by comparing the map to the surrounding terrain.
advantages: The vehicle already knows where the lanes are, the height of traffic lights, and the curvature of the road, so the computational load is low and the driving is very stable. predictable behavior even at complex intersections.
cons: It's impossible to drive where there are no maps. expanding to new cities requires months of mapping and millions of dollars in upfront investment. this is the biggest bottleneck that physically limits the rate of expansion of Weimo.
2.1.3 Safety data: statistical proof beyond humans
weimo has a huge amount of driving data to statistically prove that their system is safer than humans. according to Weimo's published safety reports, the company has seen an 80% reduction in injury-causing accidents and a 91% reduction in serious injury and fatal accidents compared to human drivers.
specifically: intersection accidents are down 96%, pedestrian collisions are down 92%, and accidents serious enough to deploy airbags are down 79%. this suggests that Weimo's conservative driving algorithms react much faster and more accurately than humans in unexpected situations. global reinsurers like Swiss Re recognize that Weimo's accident claim rates are significantly lower than human drivers.
2.2 Tesla: gambling on visual intelligence and giant neural networks
tesla's strategy is to minimize hardware and maximize the intelligence of its software (AI). it's an approach comparable to how smartphones got rid of physical keypads in favor of software keyboards.
2.2.1 Vision-only: Emulating the human eye
tesla's Full Self-Driving (FSD) relies solely on visual information from eight cameras. lidar and radar have been eliminated altogether.
economics: the hardware cost of a camera-based system is only about $400 per unit. this is one-thirtieth the cost of the Waymo sensors, a fundamental competitive advantage that allows Tesla to ship millions of vehicles per year with autonomous driving hardware and collect data.
thetechnical challenge: Reconstructing three-dimensional space from two-dimensional images with no ranging sensors makes AI's inference capabilities absolutely critical. there is a constant risk of the camera being neutralized (blinded) in adverse visual conditions such as fog, heavy rain, or direct sunlight.
2.2.2 End-to-End Neural Networks
starting with Tesla's FSD v12, the company abandoned rule-based coding in favor of an end-to-end approach where a single giant neural network handles everything from video input to control output. this is how AI learns to drive itself by watching tens of billions of miles of driving footage.
2.2.3 Cybercab and the $30,000 Impact
in October 2024, Tesla unveiled the Cybercabat its "We, Robot" event. it is a two-seater vehicle with no steering wheel and no pedals, and is targeted for mass production in 2027.
price disruption: The target price is less than $30,000, and the company promises to bring operating costs down to $0.2 per mile.
wireless charging: by eliminating charging ports and introducing inductive wireless charging, the company is heralding a fully unmanned operating system, with robots performing everything from cleaning to charging.
2.2.4 The Shadow of Safety: Sleeping Monitors and Regulatory Walls
tesla claims that Autopilot is nine times safer than a human driver, but critics argue that this may be an optical illusion based on data that is heavily weighted toward highway driving.
recent issues: a safety officer in a Tesla car being tested in San Francisco and elsewhere was caught falling asleep for nearly an hour while driving. this shows that Tesla's driver monitoring system (DMS) is not perfect, and that reliability issues have not yet been resolved to allow it to go fully unsupervised. ongoing investigations by the National Highway Traffic Safety Administration (NHTSA) are also a risk factor for Tesla.
3. market dominance comparison: Who has the future?
3.1 Waymo's territorial expansion: a proven commercial model
in 2025, Weimo is out of the 'experiment' and into the 'business'.
service area: it has expanded to all of Phoenix, San Francisco, Los Angeles (LA), Austin, and now Atlanta. in San Francisco, the company is allowed to drive throughout the city and on freeways, effectively competing directly with Uber.
fleet size: About 2,500 vehicles in service, with more than 250,000 paid rides per week. this means it's starting to generate hundreds of billions of won in annual revenue.
partnerships: By partnering with Uber to make Waymo hailable from the Uber app, the company is quickly absorbing existing ridesharing customers.
3.2 Tesla's potential: the shadow of a giant fleet
tesla does not currently operate a commercial robotaxi service. however, the company's fleet of more than 6 million vehicles around the world are potential robotaxi candidates.
theTesla Network: Musk is sticking with his idea of a "Tesla Network" where privately owned Tesla cars are monetized as robotaxis when not in use. if realized, Tesla would quickly have millions of robotaxis, giving it economies of scale that Waymo can't match.
分类waymotesla current Status fully driverless commercial service (paid) FSD beta (supervision required), testing phase number in operation approximately 2,500 vehicles 0 (potentially millions) major locations SF, LA, Phoenix, Austin, Atlanta worldwide (as soon as software is deployed) business Model B2C direct operations + Uber affiliate B2C sales + P2P network (planned) regulatory Approvals california CPUC approved unregulated (requires safety personnel on board)
4. structural crisis in the Korean market: mobility Stuck in the Galapagos
while the global market is riding the giant wave of autonomous driving, South Korea is stuck in a calm but slowly simmering pot. worse than the technological gap are the layers of regulations that stifle innovation and a taxi industry structure that is on the verge of collapse.
4.1 The specter of the 'No Ride' law and five lost years
the Passenger Carrier Business Act Amendment Bill, or "Tada Ban Act," passed by the National Assembly in 2020, left an irreparable scar on the Korean mobility industry.
how ithappened: Tada Basic, operated by VCNC, exploded in popularity by utilizing 11-seat rental cars to arrange drivers. however, after a fierce backlash from the taxi industry and a spate of alter egos, the National Assembly passed a bill restricting the use of 11-seater rental cars to "tourism purposes" and requiring rental times to be at least six hours.
closing a legal loophole: The bill effectively outlawed Tada's core business model, forcing the company to shut down its service. the Constitutional Court later ruled the bill unconstitutional, driving a wedge.
butterfly effect: the case sent a strong signal to the South Korean mobility market that "innovation that threatens an established industry is not allowed." global companies like Uber were blocked from entering, and local startups were forced to retreat to license-based franchised taxi businesses (like Tada Lite). As a result, South Korea missed a golden opportunity for data accumulation and AI learning.
4.2 The taxi industry's demographic cliff
the taxi industry, which has been protected from disruption, is now crumbling from within. the most serious problem is the aging of drivers.
the statistics warn: as of 2024, 72%of private taxi drivers will be over the age of 60. drivers in their 70s are common, and there are even some in their 90s.
causes of the late-night taxi crash: Older drivers tend to avoid driving at night due to poor eyesight and physical fitness. this is the underlying cause of the late-night taxi rush in big cities like Seoul. younger people are also unwilling to work long hours and low incomes, so they have moved into the delivery industry.
safety concerns: The rate of traffic accidents among older drivers is rising rapidly every year. statistics showing that drivers over the age of 65, who make up only 12% of all taxi drivers, cause 20% of all accidents are fueling public anxiety.24 Seoul is considering restricting the ability of drivers over the age of 75 to retain their licenses, but this is facing strong opposition from senior drivers who are the breadwinners.
4.3 The dilemma of the 120 million won license
in South Korea, a private taxi license is more than just a business license; it's a driver's retirement fund and a retirement asset.
asset value: The market price of a private taxi license in Seoul is around 120 million won. there are about 250,000 taxis nationwide, and the total value of these licenses is in the trillions of won.
the source of resistance: Once self-driving taxis are introduced, the scarcity of taxi licenses will disappear and prices will plummet. in New York City, the price of a medallion taxi plummeted by more than 90% from $1 million (approx. KRW 1.3 billion) to less than $100,000 (approx. KRW 130 million) after the introduction of Uber, leading to bankruptcies and suicides among drivers.26 The Korean taxi drivers' united opposition to the introduction of robo-taxis is not just a "battle for a bowl of rice," but a "war for the defense of all property.
5. hyundai Motor Group's agony and strategic pivot
hyundai Motor Group, South Korea's leading automobile company, made a bold decision to survive in the midst of this internal and external competition. it's called "sleeping with the enemy.
5.1 Motional's limitations and lessons learned
in 2020, Hyundai Motor established Motional, a company specializing in autonomous driving, in a joint venture with Aptiv. the goal was to secure its own Level 4 autonomous driving technology and dominate the robotaxi market.
the reality: The difficulty of developing fully autonomous driving technology was much higher than expected. despite huge R&D expenditures, commercialization was delayed. the company abandoned its 2024 commercialization goal and pushed it back to 2026, shut down pilots in Las Vegas and Santa Monica, and made massive workforce cuts.
thelesson: This highlights the challenge of fully internalizing even software-driven autonomous AI technologies for automakers (OEMs).
5.2 The Waymo Alliance: The Autonomous Driving Foundry Strategy
in October 2024, Hyundai announced a partnership with Waymo in a big way.
what: Hyundai's electric car, the Ioniq 5,will be equipped with Waymo's sixth-generation self-driving system (Waymo Driver) to serve as a Waymo One service vehicle.
where: The vehicles will be produced at Hyundai Motor Group Metaplant America (HMGMA), a dedicated EV factory built by Hyundai Motor Group in Georgia.
strategic implications: This signals Hyundai's shift to a "foundry strategy" that capitalizes on its position as a hardware platform provider rather than insisting on developing autonomous driving software in-house. Like TSMC outsourcing chips to Apple, Hyundai will maintain its market dominance by supplying the best hardware (vehicles) to the best software company (Weimo). this is in contrast to Tesla's pursuit of the "Apple model" where it monopolizes both hardware and software.
6. future scenarios and survival roadmap for the Korean market
6.1 The Bank of Korea's Shock Therapy: "Kill the License with Money"
in a recent report, the Bank of Korea warned that the current incremental approach will kill the Korean mobility industry and proposed a radical restructuring plan.
create a license buyback fund: a buy-back program should be implemented, whereby taxi licenses are purchased by the government or a public fund and burned. this could be funded through a 10 percent fee on taxi fares or a 1,000 won per call levy.
consumer benefits: The report estimates that converting 7,000 taxis, or 10% of Seoul's fleet, to autonomous taxis would increase consumer benefits by about KRW 160 billion per year through lower fares and supply efficiencies.
political challenges: The proposal is benchmarked on Australia's taxi license buyout (39% of market value compensation), but requires high-level political leadership to bridge the gap between the level of compensation drivers demand and the amount of fare increase the public is willing to accept.
6.2 Current autonomous driving experiences: The reality of Gangnam and Sangam
seoul has piloted autonomous taxi services in Gangnam (Loboride) and Sangam. however, the actual user experience is still more 'experiment' than 'future'.
user experience (UX): in 2025, self-driving taxis in Gangnam are limited to late-night hours (11 p.m. to 5 a.m.). users report feeling "motion sick" or "frustrated" due to sudden stops, frequent lane change failures, and low speeds that disrupt the flow of traffic around them.
the presence of a lifeguard: Current laws require a lifeguard to be on board, so it's not a Weimo-like experience with complete privacy. fares are free or very low, but the lack of technological maturity makes it premature to replace public transportation.
6.3 In 2027, South Korea's fate will be decided
the Ministry of Land, Infrastructure, and Transport has targeted 2027 as the year of commercialization of Level 4 autonomous driving. however, given the state of readiness of the technology and institutions, this goal is highly challenging.
galapagos risk: If South Korea sticks to regulatory barriers to prevent Waymo or Tesla from entering the market, and domestic technology alone is not competitive, South Korean consumers will be stuck with an antiquated taxi system that is expensive and inconvenient.
therise of hybrid models: A realistic alternative is likely to be the reverse importation of Waymo-equipped robo-taxis produced by Hyundai Motor Company into the Korean market, or domestic IT platforms (Kakao, T-Map) bringing in and servicing foreign autonomous driving solutions.
7. conclusion: No right to resist change
the robotaxi revolution has already begun. weimo has proven the 'technology' and Tesla is trying to disrupt the 'cost'. the competition between these two giants is redefining the standards of the global mobility market. south Korea doesn't have much time left. it should not repeat past mistakes like the 'No Ride Law'.
the government must lead the way in social compromises, including the purchase of taxi licenses, and companies like Hyundai must leverage their hardware manufacturing capabilities to become key partners in the global autonomous driving ecosystem. most importantly, consumer choice is key. freedom of movement - safe, affordable, and available at any time - is a fundamental right in the Fourth Industrial Revolution. how to secure the retirement of aging taxi drivers while opening the door to innovation for future generations is the most urgent and critical challenge for Korean society in 2025.