Robotaxi and FSD Scaling: Software Maturity vs. Commercial Readiness

Bullish investor: Tesla has crossed the threshold that most doubted it would ever reach — fully unsupervised commercial rides with zero safety incidents to date.

• Tesla is operating paid autonomous rides in Austin, Dallas, and Houston with no safety driver and no chase car. FSD accumulated over 6 billion supervised miles by Q3 2025, and the Robotaxi fleet has covered over 250K unsupervised miles in Austin alone with zero injury incidents.
• Active paid FSD subscriptions reached 1.28M globally in Q1 FY26, up 51% YoY, driven almost entirely by subscriptions. Churn is falling, and miles driven per subscriber are rising — both indicators that the product is genuinely improving.
• The core insight that most bears miss: Tesla is not building city-by-city LiDAR maps. It is building a generalized AI solution that, once validated in a few cities, can deploy anywhere the regulations allow. That scalability is structurally different from Waymo's model.
• Management expects Robotaxi revenue to be material in the second half of FY26. With a fleet of Hardware 4 vehicles already on the road and the Airbnb-style personal vehicle program launching next year, Tesla can scale the fleet without building all the cars itself.

Bearish investor: Tesla's Robotaxi is still in a fragile early stage, and the gap between "Austin pilot" and "commercially material" is enormous.

• As of Q1 FY26, Tesla has only a few hundred active Robotaxi vehicles across three cities, covering small geofenced areas. Going from this to millions of vehicles generating material revenue by the second half of FY26 requires an exponential ramp in safety validation, operational infrastructure (charging, dispatch, servicing), and regulatory approval — none of which is guaranteed.
• Management explicitly acknowledged that the biggest bottleneck is not safety but convenience: cars getting stuck, blocking intersections, and entering infinite loops. These are not trivial edge cases — they are the kinds of problems that need to be solved at scale across every new city before expanding.
• Unsupervised FSD for personal customer vehicles is expected only in Q4 FY26 at the earliest, and Hardware 3 owners — a large portion of the fleet — cannot participate without expensive camera and computer retrofits. Tesla has not disclosed the cost or timeline for this "micro-factory" retrofit program.
• FSD regulatory approval outside the U.S. is moving slowly: supervised FSD only reached the Netherlands in Q1 FY26, with broader EU approval pending and China broader approval still uncertain. This compresses the addressable fleet for the Airbnb-style model considerably.

Tesla vs. Waymo: Who Wins the Robotaxi Market

Bullish investor: Tesla's vision-only, hardware-light approach is structurally cheaper and more scalable than Waymo's sensor-laden approach — and the gap is widening.

• A Tesla vehicle used for Robotaxi costs roughly 20-25% of a Waymo vehicle by management's estimate, and Tesla produces millions of vehicles per year versus Waymo's limited fleet. Cost per mile is the only metric that matters at scale, and Cybercab management targets sub-$0.30 per mile over time.
• Waymo requires expensive HD maps of specific geographies that must be updated when roads change. Tesla's AI generalizes across geographies, which is why it could deploy FSD in China using mostly publicly available street video rather than custom mapping.
• Tesla's Supercharger network, service center footprint, and over-the-air software capability give it a unique operational infrastructure that a new entrant or even a well-funded Google subsidiary cannot replicate quickly.

Bearish investor: Waymo has a multi-year head start in commercial operations and is backed by Alphabet's essentially unlimited capital — and it is already operating safely at scale where Tesla is still validating.

• Waymo has accumulated millions of paid commercial trips across San Francisco, Phoenix, and other markets, with an established safety record. Tesla's unsupervised commercial fleet is in the hundreds of vehicles. In the time Tesla is validating, Waymo is compounding operational experience and brand trust.
• Tesla's vision-only approach has real tail risks: camera occlusion from sun glare (a still-open NHTSA investigation as of Q1 FY26), novel road conditions, and the risk that a single high-profile accident triggers multi-state regulatory setbacks simultaneously. Waymo's sensor redundancy is a cost disadvantage but also a reliability hedge.
• Waymo can expand its geographic footprint by deploying capital, not by waiting for software maturity. If Waymo proves the LiDAR model is reliable enough at commercial scale, it may capture the highest-value urban markets before Tesla's fleet reaches critical mass.

Cybercab and Optimus Production Ramp Execution

Bullish investor: Tesla has a track record of eventually delivering on hard manufacturing problems, and the Cybercab's "unboxed" process represents a genuine step-change in manufacturing efficiency.

• Cybercab entered production in April FY26, and the "unboxed" manufacturing process targets a cycle time of under 5 seconds per vehicle — roughly 6-7x faster than Tesla's best existing line. If achieved, this would significantly reduce per-unit cost and enable fleet-scale economics.
• The Cybercab is designed from the ground up for autonomy: no steering wheel, no pedals, optimized for a gentle duty cycle, built for high utilization. Management estimates 50-60 hours per week of use versus 10-11 hours for a driven car, which transforms the unit economics for fleet operators.
• Optimus 3 is described as near-human in appearance and movement, with a production line being installed at Fremont targeting 1M units per year. Tesla's combination of real-world AI, electromechanical manufacturing expertise, and ability to scale production is something management believes no other company outside China possesses.

Bearish investor: Tesla is simultaneously ramping three entirely new products — Cybercab, Optimus, and Megapack 3 — each with new supply chains and new manufacturing processes, and history argues strongly for delays.

• Musk himself said Cybercab's initial S-curve "will be very slow," and the "unboxed" process has never been used in auto history. Tesla is targeting a 5-second cycle time it has never achieved before. Optimism about a process that has never been proven at volume is speculative.
• Optimus production at Fremont requires dismantling the entire Model S/X line and installing an entirely new production line, which Musk estimated will take at least four months just for the physical transition. With 10,000+ unique components in Optimus and no existing supply chain for any of them, production ramp will be gated by the "least lucky, least confident" part — an unpredictable and potentially long constraint.
• Rare earth magnets from China remain a live single point of failure for Optimus actuators. China requires an export license for these materials, and Tesla is dependent on Chinese government approval to produce its most-hyped product.
• FY26 CapEx guidance was raised to over $25B, covering six simultaneous new factories. This is a commitment to a level of concurrent new program execution that even Tesla, with its track record, has never attempted.

Automotive Gross Margin Recovery

Bullish investor: Automotive margins are trending in the right direction, and the transition to FSD-as-a-service creates a high-margin recurring revenue layer on the existing fleet.

• Automotive gross margin excluding regulatory credits improved sequentially from 15.4% in Q2 FY25 to 17.9% in Q4 FY25 to 19.2% in Q1 FY26, driven by lower material costs, better regional mix, and fixed cost absorption improvement.
• Tesla's cost per vehicle has been driven below $35K, and the company is adding affordable new models to expand addressable volume, which should improve fixed cost absorption further as production ramps.
• FSD subscriptions at $99/month generate near-100% gross margin on an installed base of 1.28M paid subscribers in Q1 FY26. As unsupervised FSD unlocks, management believes the product value increases substantially, enabling higher pricing. If Tesla reaches 10M subscribers — one of the CEO award milestones — this would represent ~$12B/year in near-zero-cost revenue.

Bearish investor: The apparent Q1 FY26 margin improvement is partly one-time, and structural headwinds persist throughout FY26.

• The Q1 FY26 automotive gross margin of 19.2% included approximately $230M in one-time warranty true-downs. On a normalized basis, the improvement is less impressive.
• Tesla is transitioning FSD to subscription-only, which management acknowledged will pressure automotive margins in the near term, as the higher upfront revenue recognition under the purchase model is replaced by ratable monthly recognition.
• Interest rate subvention costs are recognized upfront, and management warned that if rates continue to rise, subvention will continue to pressure margins. Tesla cannot fully control this headwind.
• Regulatory credit revenue declined 28% to $2.0B in FY25 due to the OBBBA eliminating emissions penalties for other OEMs. This high-margin revenue stream is structurally impaired, not just cyclically weak, reducing the effective gross margin cushion that masked underlying automotive profitability weakness for years.
• Battery pack capacity is the stated primary production constraint, meaning Tesla cannot fully absorb fixed costs until pack capacity is resolved — a multi-quarter investment cycle.

Energy Storage: Long-Term Growth vs. Near-Term Margin Compression

Bullish investor: Megapack is a structurally growing, high-margin business, and AI-driven electricity demand is creating a demand increase that will outpace supply for years.

• Energy revenue grew 27% YoY to $12.8B in FY25, with deployments up 49% YoY to 46.7 GWh. Gross margins expanded to 29.8% for FY25. The Q1 FY26 figure exceeded 39.5%, even net of normalization.
• The demand driver is secular and accelerating: AI data centers are creating massive incremental electricity demand, and Megapack is one of the fastest ways to add dispatchable grid capacity. Management has described demand as consistently exceeding supply across all regions.
• Megapack 3 launching from the Houston Megafactory and Megablock add cost efficiency and product capability. Tesla also began producing Tesla-branded solar panels from its Buffalo factory in Q1 FY26. The integration of solar, storage, and software (Autobidder, Powerhub) creates a differentiated, stickier offering than pure-play storage competitors.

Bearish investor: Energy margins have likely peaked and face real structural compression from Chinese competition and tariffs.

• Management explicitly guided for energy margin compression in FY26 from low-cost Chinese competition (CATL, BYD) and tariffs on Chinese LFP cells. Nearly all of Tesla's battery cells for Megapack are currently sourced from China, and quarterly tariff exposure in the energy segment alone exceeds $200M.
• Energy deployments are lumpy: they fell 38% sequentially in Q1 FY26 to 8.8 GWh versus 14.2 GWh in Q4 FY25. This lumpiness creates significant quarter-to-quarter margin volatility and makes financial modeling difficult.
• Domestic LFP cell manufacturing in Nevada remains a fraction of Tesla's total installed Megapack capacity. Until domestic production scales, Tesla is price-taker on its primary input cost, and Chinese competitors can undercut on price with a structurally lower cost structure.

Capital Allocation at $25B+ CapEx: Return Profile and Execution Risk

Bullish investor: Tesla is in a once-in-a-generation investment cycle building the infrastructure for autonomous vehicles, humanoid robots, and AI chips simultaneously — and it has $44B of cash to fund it without external dilution.

• Tesla ended FY25 with $44.1B in cash and investments, generated $14.8B in operating cash flow, and has line of sight to debt financing against Robotaxi fleet cash flows. The company is not dependent on equity issuance to fund this cycle.
• The $25B+ FY26 CapEx covers the core enablers — Cybercab, Optimus, LFP cells, new Megafactory, AI compute — that management believes will drive an order-of-magnitude shift in revenue and margin structure beginning in the second half of FY26.
• Each investment is synergistic: AI5 chips power both Cybercab and Optimus; Megapack demand is driven by the same AI infrastructure growth that also drives FSD compute demand; vertical integration in batteries, chips, and materials builds a compounding cost advantage over competitors who buy components externally.

Bearish investor: Tesla is making a massive speculative bet across six concurrent programs during a period when core automotive revenue is declining — and management will be free cash flow negative for the rest of FY26.

• After generating $6.2B in free cash flow in FY25, Tesla guided for negative free cash flow for the remaining three quarters of FY26. The company is spending ahead of any proven revenue from the new programs, and any execution setback could extend the cash burn period.
• CEO compensation milestones — including 20M vehicle deliveries, 10M FSD subscriptions, 1M robots delivered, and 1M robotaxis in commercial operation — create a risk that capital is deployed to hit specific award thresholds rather than to maximize risk-adjusted returns. The board itself acknowledged in Tesla's own disclosures that the award goals "may not be indicative of the products or services that would generate the financial returns necessary to justify the award's market capitalization targets."
• The $2B equity investment in xAI is a related-party transaction where Musk acknowledged it was partly motivated by investor pressure rather than pure strategic logic. Combined with the SpaceX collaboration on TerraFab, Musk's multiple financial interests create conflicts of interest that are difficult for outside shareholders to disentangle.