Robotaxi and FSD Monetization: Vision vs. Reality

Bullish investor: Tesla has crossed the most important threshold — paying customers riding in fully driverless cars — and the ramp from here is exponential.

• Tesla launched paid, fully unsupervised robotaxi rides in Austin with zero notable safety incidents across 380,000+ miles as of FY26Q2. The service has expanded to seven markets, with management reporting week-over-week growth exceeding 10% in unsupervised miles. This is no longer a promise; it's a live, scaling operation.
• FSD attach rates are accelerating: 55% of North America deliveries in FY26Q2 had FSD at delivery, and active paid subscribers reached 1.48M globally in FY26Q2, up 51% YoY from FY26Q1's 1.28M. Management reports customers now say they're buying FSD with a car attached, not the other way around.
• The transition to subscription-only FSD creates a recurring revenue stream. At roughly 1.5M subscribers at ~$99/month, this is already approaching a $1.8B+ annualized revenue run rate at near-100% gross margin — and penetration of the installed fleet remains low, around 12%.
• The Cybercab entered production in FY26Q2 with economics purpose-built around a sub-$0.30 cost-per-mile target, which management says will be transformative once at scale. Cybercab production ramping alongside the existing Model Y fleet gives Tesla a dual-track scaling strategy.

Bearish investor: Tesla's robotaxi business is real but still very small, and the path to financial materiality is longer and bumpier than management implies.

• As of FY26Q2, Tesla has accumulated 380,000 cumulative unsupervised miles across six cities — still a very small number. Despite launching over a year ago in Austin, the robotaxi fleet remains in the dozens-to-hundreds of vehicles range, not the thousands needed to move the financial needle. Meanwhile, operating income declined 57% YoY to $398M in FY26Q2, while the company is spending billions to build the fleet.
• The subscription-only FSD transition, while strategically correct, creates an acknowledged near-term margin headwind that has already shown up. Automotive gross margins excluding credits fell from 19.2% in FY26Q1 to 16.3% in FY26Q2.
• Hardware 3 vehicles — roughly 30% of the fleet — cannot run unsupervised FSD and require hardware and camera replacements. Tesla has acknowledged needing to set up "micro factories in major metropolitan areas" to do these upgrades efficiently. The cost and timeline of this program have not been quantified and represent a real overhang on fleet scalability.
• FSD and robotaxi are still pending regulatory approval in Europe and China — two of Tesla's largest markets — which suppresses attach rates and fleet scale in a major portion of the global installed base.

Optimus: Transformative Product or Overpromised Science Project?

Bullish investor: Optimus 3 is the right design, the ramp has started, and Tesla has capabilities no competitor can match.

• Tesla is the only company with all three requirements for a useful, scalable humanoid robot: real-world AI (developed through billions of FSD miles), advanced electromechanical engineering (particularly hand dexterity), and mass manufacturing expertise. Management is emphatic that no other company outside China has meaningful competition, and Tesla believes it is ahead of Chinese competitors on AI and dexterity.
• The same end-to-end neural network architecture powering FSD — pixels in, controls out — is being directly applied to Optimus. The team, the data flywheel, and the training infrastructure are all shared, giving Tesla a head start that would take competitors years to replicate.
• Optimus 3 is already walking 24/7 in Tesla's Palo Alto offices, guiding visitors, and performing factory tasks. Production has begun ramping at Fremont, with a second Optimus factory being built at Gigafactory Texas. Management has guided to 1M units/year at Fremont as the long-term target.
• At scale production of 1M+ units/year, management estimates production cost below $20,000 per unit, implying massive profit potential for a product with virtually unlimited demand in the labor market.

Bearish investor: Optimus is a genuinely hard problem with an entirely new supply chain, and production timelines will be far slower than management implies.

• Every component in Optimus is new — there is no existing supply chain for any of the 10,000+ unique parts. Management itself says production will move at the speed of the "slowest, least lucky" component, making planning nearly impossible. This was already demonstrated when Chinese rare earth magnet export restrictions stalled actuator production in early FY25.
• The Optimus production S-curve will be "quite flat and long" per Musk in FY26Q2 — a candid admission that meaningful volumes are years away. Converting the Fremont Model S/X line alone takes months just to dismantle and reinstall. Optimus 4, targeting 10M units/year, requires a still-unannounced TerraFab and an even more vertically integrated supply chain that doesn't exist yet.
• Chinese humanoid robot companies are advancing rapidly with the full weight of China's manufacturing scale and AI capabilities behind them. Management acknowledges China will be "the toughest competition by far."
• Despite enormous hype, Optimus has generated essentially zero revenue to date and will not be a meaningful revenue contributor in the near term. Investors are being asked to fund a multi-year, capital-intensive science project with highly uncertain timing.

Competitive Position in Core EVs vs. Chinese and Legacy Automakers

Bullish investor: Tesla's EV position is stabilizing, FSD is becoming a real differentiator, and no competitor has the full package.

• Model Y remains the best-selling car of any kind globally. FY26Q2 saw record Q2 deliveries of 480,126 units, up 25% YoY, with Tesla exiting the quarter with its largest order backlog since 2023. This is not a company losing the EV race.
• FSD is becoming a genuine demand driver that competitors cannot replicate. Management notes customers are coming into stores specifically for FSD, and in markets where FSD is approved, the effect on demand is visible. No other automaker has anything close to Tesla's autonomy capability, creating a durable moat that strengthens with every mile driven.
• Tesla's USMCA-compliant supply chain (~85% on a weighted average basis) and vertical integration give it a structural tariff advantage over essentially all competitors, which is increasingly important in the current trade environment.
• Cost per vehicle below $35,000 as of Q4 FY24 gives Tesla a cost structure that most legacy OEMs and Chinese EV entrants struggle to match outside of China.

Bearish investor: Automotive margins are structurally under pressure, and Tesla is losing pricing power to better-funded Chinese competition.

• Automotive gross margin excluding regulatory credits has declined from a peak of ~29% in FY21-22 to 17.8% in FY25 and 16.3% in FY26Q2. Price cuts have consistently outpaced cost reductions for three straight years. There is no sign this trend is reversing.
• The $7,500 IRA EV consumer tax credit was eliminated in Q3 FY25, effectively raising the price of Tesla vehicles for U.S. buyers by $7,500 with no offsetting benefit. Regulatory credit revenue — essentially free money — collapsed 28% in FY25 to $2.0B and is unlikely to recover under current policy.
• In China, Tesla faces BYD and a growing array of well-funded domestic EV brands competing aggressively on price and features. Tesla's FY25 deliveries declined 9% YoY to 1.64M while BYD continued to grow. Tesla's brand headwinds from Musk's political activities impacted certain markets in FY25, per management's own acknowledgment.
• Battery pack supply has been the primary production constraint in FY25 and FY26Q1, capping Tesla's ability to capitalize on demand recovery even when order trends improve.

Capital Cycle: $25B+ CapEx With Uncertain Returns

Bullish investor: Tesla is making the most ambitious and potentially most valuable industrial investment in modern history, and the balance sheet can support it.

• Tesla is simultaneously launching six new production lines, building AI compute infrastructure, and funding semiconductor R&D from a starting position of $44B in cash and $14.8B in FY25 operating cash flow. The company can fund this internally without issuing equity.
• Every major tech company is spending at similar or greater scale on AI infrastructure. Tesla's investments span physical AI (robots, autonomous vehicles), energy storage, and semiconductor design — all of which management argues have better returns than pure cloud infrastructure spending.
• Management argues they are being capital-efficient — acting as general contractor on all construction, minimizing waste, and prioritizing productive assets. The FY26Q2 $5.8B quarterly CapEx is going into factories that will produce real physical products with clear demand.
• If even a fraction of the Optimus and robotaxi opportunity materializes, the $25B+ CapEx cycle will look extraordinarily cheap in hindsight.

Bearish investor: Tesla is in a speculative capital cycle that has turned free cash flow negative with no clear near-term payback.

• Free cash flow turned negative in FY26Q2 at -$1.1B, and management has guided for CapEx to "grow for the next two to three years" beyond the $25B+ FY26 guide, with solar cell manufacturing and the TerraFab semiconductor fab not yet included in that number. This cycle could extend to $30B+/year of spend.
• Tesla is now actively seeking to borrow up to $30B in debt facilities, layering financial leverage onto a business generating GAAP operating margin of just 1.4% in FY26Q2. This is a challenging combination.
• The $2B investment in xAI and deepening SpaceX entanglement (TerraFab partnership, Digital Optimus, Starlink integration, $2B SpaceX equity investment in FY26Q1) raise legitimate questions about whether Tesla's capital is being deployed in Tesla shareholders' best interests.
• Tesla's own risk disclosures warn that CEO performance award targets may not align with products that generate sufficient financial returns — an unusual acknowledgment that management's vision and shareholder returns may diverge.

Energy Storage: Margin Durability Under Competitive Pressure

Bullish investor: Energy storage is a secular growth business driven by AI data center demand and grid modernization, with Tesla uniquely positioned to benefit.

• Energy deployments grew 49% YoY to 46.7 GWh in FY25, and management expects 2026 deployments to exceed 2025. The backlog is robust and globally diversified. AI data centers alone are creating enormous demand for fast-acting battery storage to smooth training run power fluctuations — a specific use case where Megapack with its Autobidder software is differentiated.
• Megapack 3 — with improved unit economics — is entering production in FY26, and a third Megafactory near Houston is ramping. Megapack 3 should drive meaningful cost reduction and help offset ASP compression.
• The domestic LFP cell manufacturing facility in Nevada began commissioning in early FY26. As this scales, it reduces Tesla's dependence on tariff-exposed Chinese cells, which currently add $200M+ per quarter in cost.
• Even at the guided "mid- to low 20%" normalized gross margin range, Energy is a high-growth, double-digit margin business that is a meaningful and expanding contributor to Tesla's overall profitability.

Bearish investor: Energy margins have already compressed sharply and the structural headwinds are intensifying.

• Energy gross margin contracted from a one-time peak of 39.5% in FY26Q1 (inflated by ~$250M in tariff reversals) to 20.4% in FY26Q2, which also included a ~$240M warranty true-up. Normalizing for both, underlying margins may be well below 20% already. Management has guided the business to "mid- to low 20%" going forward — a significant step down from the 29.8% achieved in FY25.
• Competition from CATL, LG Energy Solution, and other integrators is intensifying on price. Management has explicitly acknowledged ASPs are declining. The entry-level economics of large-scale battery storage are commoditizing quickly.
• Energy deployments are inherently lumpy — FY26Q1 saw a 38% sequential decline to 8.8 GWh before rebounding to 13.5 GWh in FY26Q2. This makes reliable financial forecasting difficult and introduces quarterly earnings volatility that can mask underlying trends.
• Most LFP cell inputs remain China-sourced, with tariffs adding over $200M per quarter in FY25. U.S. LFP manufacturing covers only a fraction of demand, and building out domestic supply will take years and substantial additional capital.

Chip Supply Chain as the Long-Term Growth Constraint

Bullish investor: Tesla's AI5 chip and TerraFab initiative are proactive and essential investments that will remove a critical bottleneck before it becomes a crisis.

• AI5 taped out in April FY26 and is already in volume production at TSMC (Arizona) and Samsung (Texas) — all U.S.-based fabs. Management claims AI5 will be 40x better than AI4 on key metrics and potentially the best edge computing AI chip in the world, designed specifically for Tesla's use cases.
• Tesla's AI is uniquely memory-efficient — management claims Tesla's intelligence density per gigabyte exceeds any other AI company by an order of magnitude. This gives Tesla a meaningful buffer against the memory constraints that limit other AI deployments.
• The research semiconductor fab at Gigafactory Texas will integrate lithography mask production, logic, memory, and packaging under one roof — a first-of-its-kind facility that enables rapid iteration cycles and could yield breakthrough chip architectures.
• Partners Samsung and TSMC are both investing billions in U.S. fabs partially dedicated to Tesla's future projects, with Micron providing meaningful memory allocation. This is not a solo effort.

Bearish investor: Chip supply will become Tesla's primary growth constraint within three to four years, and TerraFab is still years from meaningful production.

• Management has explicitly stated that chip production — specifically AI logic and memory — is likely to become the primary constraint on Tesla's growth in three to four years. There are currently zero advanced memory fabs at scale in the United States, and even maximum output from TSMC, Samsung, and Micron may be insufficient for Tesla's FY28-FY29 plans.
• The TerraFab initiative is in very early stages. The Gigafactory Texas research fab is a "few thousand wafer starts per month" R&D facility — not remotely production-scale. The large-scale TerraFab involves SpaceX as lead investor and has not been fully announced. The location hasn't even been disclosed yet.
• AI5 is being prioritized for Optimus and data centers rather than vehicles, because AI4 is sufficient for unsupervised FSD. But the ~30% of the Tesla fleet still on Hardware 3 cannot run unsupervised FSD at all and requires hardware and camera replacement — a massive logistical and capital burden that management has not fully quantified.
• Tesla is asking investors to fund a semiconductor fab — one of the most capital-intensive and technically complex industrial undertakings in existence — on top of everything else in the current CapEx cycle. This is an area where Tesla has no track record, and the risk of significant cost overruns or delays is high.