Tesla's Q1 2026 Earnings Call

The check engine light is a dashboard warning that indicates the vehicle’s engine or emissions system has detected a fault. It can range from minor issues (like a loose gas cap) to problems that need prompt diagnosis.

ABS stands for Anti-lock Braking System. When it detects a problem, the ABS warning light can come on, indicating the system may not prevent wheel lockup during hard braking.

O'Reilly Veriscan is a diagnostic service/tool that reads vehicle warning codes and generates a report. The pitch here is that it reduces guesswork by providing verified solutions and can point you toward next steps.

ASE is the National Institute for Automotive Service Excellence, a widely recognized certification program for automotive technicians. “Master technician” typically indicates advanced, high-level credentials across multiple areas of vehicle repair.

Progressive is an insurance company that offers tools like “Name Your Price,” which helps customers choose coverage options aligned with a target budget. It’s presented as a way to compare policy choices more easily.

“Name Your Price” is a Progressive feature where you set a target monthly price and the system shows policy options that can fit that budget. It’s essentially a budgeting-first way to shop insurance.

The Toyota Sienna is a minivan, and the segment highlights it “back in the Sienna” with an available rear-seat entertainment system. Minivans are often chosen for family logistics because they make it easier to access the second and third rows.

The Toyota RAV4 is a compact SUV, and the segment mentions “slip into the RAV4 with available all-wheel drive.” All-wheel drive can improve traction in rain, snow, or uneven surfaces by distributing power to more wheels.

All-wheel drive (AWD) sends power to all four wheels, improving traction compared with two-wheel drive. Many AWD systems are optimized for normal driving but can react quickly when grip is reduced.

“Hardware 3” refers to Tesla’s onboard compute hardware used for driver-assistance and autonomy features. In this segment, Tesla is saying Hardware 3 lacks the capability to reach unsupervised FSD, meaning it can’t fully operate without driver monitoring.

Tesla is the electric-vehicle company hosting this Q&A webcast. The call is focused on Tesla’s quarterly results and forward-looking outlook for its EV and software businesses.

“Forward-looking statements” are predictions about future performance or plans that may not come true. The speaker notes that actual results can differ due to risks and uncertainties, including items discussed in SEC filings.

Autonomous driving refers to vehicle systems that can perceive the road and control the car with varying levels of driver assistance, from lane centering to full self-driving in limited conditions. In Tesla’s context, autonomous-driving progress is closely tied to onboard hardware and software updates.

“Capital expenditures” (capex) are large, upfront spending commitments—typically for factories, equipment, and major infrastructure. In an earnings call, capex is a key indicator of how aggressively a company plans to expand manufacturing capacity and scale production.

A battery powertrain is the full EV propulsion system centered on the battery pack, including how energy is delivered to the motors and managed by the vehicle’s control electronics. When Tesla mentions investing in battery powertrain, they’re signaling work on efficiency, performance, and cost improvements across the EV’s core energy system.

“AI software” in an EV context typically refers to software systems that use machine learning for functions like driver assistance, vehicle perception, and optimization of driving/charging behaviors. Tesla’s emphasis on AI software suggests it sees software as a major differentiator alongside hardware.

AI training is the process of teaching machine-learning models using large datasets and compute resources so they can perform tasks reliably. In an automotive company’s earnings discussion, “AI training” usually implies ongoing investment in data pipelines, training infrastructure, and model improvement cycles.

Chip design refers to developing the specialized semiconductors used to run vehicle computing tasks, such as perception, control, and AI workloads. For EV makers, in-house or closely managed chip design can improve performance, supply reliability, and cost over time.

Vehicle production refers to the rate at which a manufacturer builds cars, typically tied to factory capacity, component availability, and manufacturing process improvements. Earnings-call language about “laying the groundwork” for increased vehicle production usually signals plans to ramp output in future quarters/years.

Optimus is Tesla’s humanoid robot project. In the earnings-call context, Tesla is describing how it’s moving from releasing/advancing the robot to increasing internal production for testing, with a goal of making it useful beyond Tesla.

“Supervised full self-driving” refers to Tesla’s driver-assistance system where the car can handle more driving tasks, but a human supervises and remains responsible. Tesla is claiming it’s improving rapidly, implying better real-world capability and reliability.

Tesla’s Semi is the company’s electric heavy-duty truck. Here, Tesla is saying it’s about to begin production and that early output will be limited by the challenges of standing up a new supply chain and manufacturing process.

CyberCab is Tesla’s upcoming new vehicle product being introduced in this call. Tesla says it has just started production, and that ramping will be slow at first due to a brand-new supply chain before accelerating later.

An S-curve describes how production ramps over time: slow at first, then faster as processes stabilize, and finally leveling off. Tesla is using it to set expectations that CyberCab and Semi output will start low due to a new supply chain, then increase more quickly later.

Megapack is Tesla’s large-scale grid energy storage product. Tesla says demand is strong and that it plans to begin production of “megapack 3” later this year, tying it to the need for energy storage as electricity demand grows.

“Full self-driving” (FSD) is Tesla’s software suite aimed at automating driving tasks. In the call, Tesla discusses how newer FSD versions change the software architecture and safety behavior, with the goal of eventually enabling unsupervised operation where legal.

RoboTaxi refers to Tesla’s planned autonomous ride-hailing service using its self-driving software. The transcript links FSD version updates to RoboTaxi, implying the software roadmap is intended to support a commercial, driverless-style operation where permitted.

AIFO is referenced as the compute platform the next FSD version will run on. In this context, it suggests a shift in the hardware/software stack used for perception and decision-making, which can affect latency, safety, and scalability.

Tesla is claiming its FSD safety performance can exceed human driving safety, and that newer versions will raise that bar further. This is typically based on safety metrics and scenario analysis, but the exact definition and measurement method matter a lot when interpreting such claims.

Tesla’s Cybertruck is the company’s angular, stainless-steel pickup truck. In earnings-call discussions, it often comes up as an example of Tesla’s ability (or difficulty) to scale a new product to mass-market demand.

The segment raises a key risk in robotics and automation: if a robot’s sensors, training, or production feedback are transmitted back to the robot maker, buyers may worry about confidentiality and future competition. This is especially sensitive when the robot is built using data from a customer’s manufacturing process.

“FSD cameras” refers to the camera suite Tesla uses for its Full Self-Driving (FSD) driver-assistance and autonomy features. These cameras feed the vehicle’s computer so it can perceive the road and traffic and support functions like lane guidance and automated driving behaviors.

“Autonomy ready” is a marketing/engineering claim that a vehicle has the hardware and software foundation to support future self-driving capabilities. In Tesla’s ecosystem, this is often discussed alongside computer and sensor upgrades that change what the car can do over time.

“Hardware 2” is Tesla’s earlier-generation onboard computer for autonomy-related processing. The host references it to highlight that Tesla has previously made similar capability/roadmap claims across multiple hardware generations.

“14.3” is a specific Tesla FSD software release mentioned as arriving soon and claimed to deliver a major improvement. Listeners should treat it as a software update version rather than a vehicle model or hardware change.

“15” is presented as the next FSD software target after 14.3, with Tesla claiming it will be a step-change improvement. In practice, these version numbers typically correspond to iterative updates to driving behavior, perception, and planning.

“Rigorous validation” describes the process of proving the robotaxi software is safe enough to operate in public. It typically involves scenario testing, data review, and operational monitoring to reduce the risk of accidents as service expands.

“Starter production” usually means an initial manufacturing run—limited quantities to validate the production process, supply chain, and quality control. It’s a common phase between prototype and full-scale production ramp.

A “supply chain” is the network of suppliers, logistics, and manufacturing steps needed to make parts and assemble a product. When Tesla says the supply chain is “totally new,” it usually means new sourcing and processes that can slow early production until everything is running smoothly.

“Ramp up” refers to increasing production volume over time as manufacturing processes stabilize and yield improves. Early output is often limited while factories and suppliers work through learning curves and quality issues.

“Frame-by-frame analysis” refers to competitors studying released footage in detail to infer design choices, motion characteristics, and potentially software behavior. In product launches, this can drive companies to delay revealing details until closer to production to reduce the amount of actionable information leaked.

This is a product strategy of timing disclosures to minimize competitive copying and to ensure the technology is mature enough to represent the final product. It also manages expectations by avoiding early prototypes that may change significantly before mass production.

“Nearly fully baked” is a metaphor for reaching a mature, stable state where the design is largely finalized and risks are reduced. In manufacturing and engineering, this typically means fewer last-minute changes and more predictable performance.

“Programming” here refers to the robot’s software—its control logic, behaviors, and autonomy. The speaker’s point is that competitors can only see external appearance from public demos, not the internal software and system details that determine real capability.

“Tilt the scale” is used here as a metaphor for shifting competitive advantage or investor perception. The host is weighing whether Tesla’s public robot demonstrations provide enough evidence to influence how people judge Tesla versus other companies working on similar robotics.

This is an example of how competitors might respond to new information: if they believe Tesla’s approach is fundamentally different, they could discard most of their existing progress and restart. The host questions whether that drastic pivot would be rational based on limited public footage.

AI5 is Tesla’s next-generation AI chip they’re describing as having been “taped out,” meaning the design has been finalized for manufacturing. The speaker frames it as a top-tier inference chip, suggesting it’s optimized for running the autonomy/AI workload inside the vehicle.

An AI inference chip is designed to run trained neural networks to make real-time decisions. “Edge compute” means the car processes AI locally on the vehicle (near the sensors) instead of relying on cloud servers, which can reduce latency and improve responsiveness.

AI6 is mentioned as the next step after AI5, indicating Tesla is already designing the subsequent generation of its in-house AI chip. This matters because each new chip generation can improve performance, efficiency, and the complexity of what the vehicle can do locally.

Dojo is Tesla’s purpose-built AI training system used to process large amounts of driving data and improve its autonomy software. “Dojo 3” refers to the next generation of that training platform, implying more compute capacity and faster iteration on models.

A chip “fab” is a manufacturing facility for producing semiconductor chips. Building a research chip fab on the “Gigatexus” campus suggests Tesla is expanding its vertical integration—doing more of the chip development/manufacturing work closer to its broader AI and vehicle programs.

An EU-wide approval refers to regulatory permission that would allow Tesla’s FSD software to be offered across multiple European Union countries under a broader authorization. The call frames it as a step beyond individual country approvals, with timing dependent on regulators’ processes.

“Existing fleet” means Tesla vehicles already in customers’ hands. When approvals arrive, Tesla can push software updates to those cars, increasing usage without needing customers to buy a new vehicle immediately.

“Lumpy” means deployments don’t happen smoothly month-to-month; they cluster around customer project schedules and installation timelines. Tesla is tying variability to when customers actually deploy storage systems, which affects quarterly results.

“GW” here is a power capacity unit (gigawatts) used to describe the scale of energy storage deployments. In practice, energy storage systems are often discussed in terms of how much power they can deliver, and the call uses it to quantify quarterly deployment volume.

Gross margin is the percentage of revenue left after direct costs, before operating expenses. Tesla is highlighting that margins in the energy storage business were unusually high due to one-time benefits related to tariff accounting/recognitions.

Tariffs are taxes or import fees applied to goods crossing borders. Tesla says tariffs can have outsized impacts in energy storage because many battery cells are procured from China, so changes in trade costs can quickly affect costs and pricing.

“Normalized” results adjust for unusual or one-time items to show what performance might look like under typical conditions. Here, Tesla is separating one-time tariff-related benefits from the underlying trend, which they expect to be pressured by competition and tariffs.

Churn is the rate at which customers stop using a service. In the context of Tesla’s FSD subscription, it’s a metric for subscription retention—how many paid customers cancel over a given period.

The Tesla Model X is a fully electric, battery-powered SUV known for its family-friendly size and distinctive design. In podcast discussions, it often comes up because it represents Tesla’s lineup strategy and production timing, especially when hosts talk about whether certain models are being paused or restarted. That’s why it’s mentioned alongside questions about production plans and the status of other Tesla vehicles.

The “10,000 unique items” framing emphasizes how complex a new product launch can be when many different parts must be sourced, manufactured, and assembled correctly. In manufacturing terms, more unique components generally increases the number of potential bottlenecks during early production.

“Ramp production” is the process of increasing manufacturing output from early builds to steady, high-volume production. The transcript highlights a common manufacturing reality: output is limited by the slowest or most problematic component or process step during the ramp.

“Fine production” here appears to mean reaching stable, repeatable manufacturing output—where the process is mature enough that production is no longer dominated by debugging and rework. It’s essentially the point where the factory can consistently build units without major interruptions.

An assembly line is a manufacturing system where a product is built through a sequence of stations, each performing a specific task. When a company introduces a brand-new product and production line, the early ramp phase is often slow because every step and component has to work together reliably.

The Mercury Villager is a minivan model associated with the Mercury brand, typically used for family transportation and everyday hauling. In the podcast context you provided, it appears as part of a humorous or fictional scenario rather than a technical discussion of the vehicle. That kind of mention usually signals the car as a recognizable “everyday” minivan reference.

“Recurring revenue” means income that repeats over time, such as subscription fees or ongoing service charges. In this context, the hosts are discussing how autonomous services (like RoboTaxi) could generate ongoing money instead of one-time vehicle sales.

The transcript focuses on how Tesla qualifies “injuries” when reporting safety outcomes for unsupervised FSD/RoboTaxi operations. The hosts note that “no injuries” may still allow minor injuries, which can be defined differently depending on reporting standards.

The National Highway Traffic Safety Administration (NHTSA) is the U.S. agency that oversees vehicle safety and collects crash-related reporting. Referencing NHTSA implies the speaker is using a formal definition of what counts as an “accident” for reporting purposes.

The transcript highlights that certain infrastructure factors—like confusing lane markings or intersections with high crash histories—can be especially challenging for automated driving systems. These factors can increase the likelihood of edge cases where perception and decision-making struggle.

“Gradual” rollout means enabling advanced autonomy in stages rather than all at once. This approach helps the company monitor real-world performance, learn from edge cases, and adjust before broader deployment.

The idea is that autonomy performance depends heavily on local conditions—road design, signage, weather patterns, and traffic behavior. So companies may enable advanced driving features only after a specific area meets safety criteria.

This suggests Tesla (or its partners) may seek approval for limited, controlled robot-taxi operations before enabling the same capability for private customer vehicles. Taxi deployments can be easier to manage operationally and to regulate in specific areas.

“Hardware 4” is Tesla’s newer onboard compute platform used to run more advanced FSD capabilities. The speaker ties hardware 4’s higher memory bandwidth to enabling unsupervised FSD and later robotaxi operations.

Memory bandwidth is how quickly the onboard computer can move data to and from memory. The call frames it as a key limiting factor for AI workloads used in unsupervised FSD, especially for large transformer-style models that need lots of data throughput.

A trade-in is when Tesla offers a discounted value for a customer’s current vehicle hardware configuration in exchange for upgrading to a newer one. Here, Tesla describes a discounted trade-in for cars with AI4 hardware (as stated in the call) as part of the path toward hardware upgrades.

The speaker says Tesla will offer an upgrade that replaces the vehicle’s onboard compute computer (the FSD computer) to move from hardware 3 to hardware 4. This is more than a software update—it implies physical hardware replacement.

The transcript indicates that upgrading to hardware 4 also requires replacing the vehicle’s cameras. That suggests the camera hardware (sensor capability, interface, or resolution) must match the newer compute platform for the intended FSD performance.

“Micro factories” here refers to setting up small, localized production/assembly lines in major metropolitan areas to perform hardware upgrade work efficiently. The speaker contrasts this with doing upgrades only at traditional service centers, which they say would be too slow and inefficient.

A “robot taxi fleet” is a planned operation where vehicles drive themselves to provide ride-hailing service without a human driver. In the transcript, Tesla links converting vehicles to hardware 4 as a prerequisite for entering that fleet with unsupervised FSD.

“V14” is the version number of Tesla’s FSD software being discussed. The call says Tesla will release a V14 version for hardware 3 that is distilled from the V14 software intended for hardware 4, aiming to deliver similar features within hardware limits.

A “distilled” software version is a smaller or more efficient model derived from a larger one, designed to run on less-capable hardware. In this context, it’s how Tesla plans to bring many V14 features to hardware 3 despite its lower compute/memory resources.

“Park State” refers to starting the driving feature from the vehicle’s parked condition. The speaker claims that with the V14 hardware 3 release, users should be able to begin the drive from Park State and access the same feature set as hardware 4’s V14 (as described).

The hosts discuss the business and customer-impact tradeoffs Tesla faces when upgrading older cars: offering free hardware updates, issuing refunds, or moving customers into newer vehicles. This matters because replacing hardware can be expensive and logistically difficult (e.g., not feasible at every service center).

“Unsupervised self-driving” refers to autonomy that can operate without continuous human intervention or ongoing supervision. The speaker frames AI4 as being able to reach this level of capability, and claims it would exceed “human safety levels,” which is a key threshold in autonomy roadmaps.

The speaker suggests the AI4 compute hardware will age enough that it becomes a reason to keep upgrading systems, even if it’s still used in the factory. This highlights how AI roadmaps often depend on hardware lifecycle, not just software improvements.

SOC stands for “system on a chip,” meaning multiple computing components are integrated into a single chip package. The transcript ties SOC to memory configuration (32 GB per SOC) and compute/memory performance, indicating how the AI platform scales across its hardware.

“AI4.1” is presented as an incremental upgrade path from AI4, likely reflecting the RAM and performance improvements described earlier. The speaker also links timing to production readiness, suggesting a staged rollout rather than a single all-at-once replacement.

The speaker says the AI4 upgrade depends on Samsung completing “modifications” for production. That implies Samsung is involved in manufacturing or packaging the relevant compute/memory hardware, making it a key supply-chain dependency for Tesla’s AI platform timeline.

The speaker suggests AI5 may be optimized for robotics (Optimus) rather than for automotive use. That matters because robot workloads, power/thermal constraints, and sensor/compute needs can differ from what a car requires for driving and driver-assist.

A “safety driver” is a human in the vehicle who can take over immediately if the autonomous system fails or encounters an unexpected situation. The segment focuses on removing this role only after safety validation is strong enough.

A “QA fleet” is a controlled set of vehicles used for quality assurance and validation of autonomous-driving behavior. The segment contrasts using QA vehicles versus the broader customer fleet to gather metrics and accelerate safety validation.

“Scaling issues” refers to practical problems that appear when autonomous vehicles operate at larger volumes, not just in limited test scenarios. The hosts mention scenarios like blocking intersections and dropping people off at slightly incorrect locations, which require system improvements beyond basic driving.

The hosts discuss using Tesla’s customer-owned vehicle fleet as a data source to monitor performance and safety metrics for FSD. They mention the fleet accumulating massive mileage as part of the validation strategy.

V15 is mentioned as a potential later software version if V14.3 isn’t sufficient for large-scale unsupervised FSD. This frames the discussion as a roadmap of software releases tied to safety validation progress.

“Major architectural improvements” refers to fundamental changes in how the software system is designed—often affecting perception, planning, and safety mechanisms. The speaker argues these changes increase the probability of safety, so they want to validate and release them before scaling autonomy.

This describes a staged deployment approach: finish writing the software, validate it, release it, and only then expand to larger-scale unsupervised operations. It highlights the idea that safety validation should precede broad real-world exposure.

The “TerraFab” project is discussed as a joint effort involving Tesla and SpaceX to build semiconductor manufacturing capacity. The segment focuses on who handles research, scaled-up deployment, and governance approvals.

“WAFAs” is used as a production-rate metric for the fab—essentially how many wafer-equivalents (or wafer-related units) can be produced per month. The point is that they need a few thousand units monthly to confirm the process works in production conditions.

SpaceX is described as taking care of the initial phase of the scaled-up TerraFab. The segment also notes governance complexity requiring approvals involving both Tesla and SpaceX leadership.

The speaker explains that any intra-company (or cross-entity) matter has to be approved by both Tesla and SpaceX boards. This is a governance concept: major strategic decisions require independent director review and conflict-resolution processes.

Intel is being discussed as a partner for manufacturing technologies used in Tesla’s production. The key point is that Tesla plans to use Intel’s semiconductor process node (“14A”) for some core manufacturing steps, tying EV hardware to advanced chipmaking.

FST is referenced as a factor that can help drive vehicle sales, and the speaker mentions improvements to FST technology with “version 14.” In EV/tech discussions, this kind of acronym often refers to a software/feature system or platform update that can improve user experience, product capability, or production/launch efficiency.

The speaker describes a future lineup strategy where Tesla’s vehicles become autonomous and come in multiple sizes. This is a product-planning concept: instead of one model, the company would offer a range of autonomous-capable vehicles optimized for different use cases and passenger capacities.

The Tesla Roadster is Tesla’s high-performance electric sports car concept/line. In this segment, the hosts suggest it could be the last “manually driven” car, implying Tesla’s broader shift toward autonomy. The “New York” phrasing likely refers to a specific event or demo location rather than a different model.

Safety metrics are the measurable indicators used to judge whether an autonomous driving system is safe enough to expand operations. In this segment, they’re discussing metrics like miles per intervention, miles per accident, and fatality rates to quantify real-world risk.

“Miles per intervention” is a common autonomy safety/quality metric: how many miles the vehicle can drive before a human or system override/intervention is required. Higher miles per intervention generally indicates the system is handling more situations without needing help.

Simulators are software environments used to recreate driving scenarios so the autonomy system can be evaluated without waiting for rare events to happen in the real world. They’re often used alongside real-world data to stress-test edge cases and improve confidence before expansion.

Neural networks are machine-learning models that can learn patterns from data and make predictions for perception and decision-making in autonomous driving. Here, they’re referenced as part of how the system estimates what would have happened in a given scenario.

This segment focuses on what limits wider deployment of robotaxis/road taxis, emphasizing that many constraints are not “hard safety” problems but usability issues. Examples include cars getting stuck at signals, refusing maneuvers, and looping when construction blocks routes.

Waymo is an autonomous-driving company known for deploying robotaxis. In this segment, the speakers compare “road taxi” incidents and behaviors—like getting stuck due to crashes or planning loops—to issues they associate with Waymo’s system.

The speakers describe a common autonomous-driving failure mode: overly conservative behavior. When the system prioritizes maximum safety, it may refuse to proceed in ambiguous situations (like crossing railroads or handling signals), leading to “stuck” behavior that hurts usability even if it’s not strictly unsafe.

An “infinite loop” in autonomous driving is when the vehicle repeatedly tries the same maneuver (like turning into a road) but can’t complete it due to an obstacle or planning failure. The car then cycles behaviorally—often going around the block and trying again—until the system is corrected.

The General Motors EV1 was an early electric vehicle from GM, built to demonstrate that battery-electric cars could be practical for everyday driving. It’s frequently discussed in EV history because it became a landmark example of how early EV programs were handled and what happened to them over time. In a podcast, it may come up through a personal story about owning one and what that experience was like.

“Carbon footprint” refers to the total greenhouse-gas emissions a person, organization, or city is responsible for. In the context of the episode, the host is pointing to practical municipal actions to reduce emissions, like upgrading equipment rather than replacing everything at once.

This is a restatement of a common sustainability principle: replacing systems too aggressively can be wasteful because it creates additional manufacturing and disposal emissions. The host contrasts that with a “replace when it’s time” approach—choosing more efficient options as older equipment reaches the end of its useful life.

This describes “timing” as a key lever in decarbonization: the most environmentally sensible upgrades often happen when equipment naturally fails or reaches end-of-life. Pairing end-of-life replacement with higher-efficiency products reduces both emissions and unnecessary waste from premature replacement.

The idea that “how a car is powered is a political statement” refers to how EVs, hybrids, and gasoline vehicles are tied to policy debates about emissions, energy sources, and industrial strategy. It frames vehicle technology as more than transportation—also as part of broader cultural and regulatory conflict.