Tesla: Vertical Integration + Physical AI Build Multiple Flywheels, Robotaxi and Optimus as Core Long-Term Value Drivers

Overview

This research report systematically evaluates Tesla's (TSLA) long-term strategic value as a leader in 'physical AI,' focusing on its breakthrough potential in three emerging TAMs: Robotaxi, Optimus humanoid robots, and energy storage. The report argues that Tesla's core moat is not a single technology, but its full-stack vertical integration capability (from lithium mining to AI chips), which generates multiple self-reinforcing flywheel effects (e.g., factories serving as both product delivery sites and Optimus training grounds). Although current valuations reflect optimistic expectations, the rating remains Neutral with a target price of $475, implying approximately 13% upside, as short-term profitability has not yet materialized.

Core views

The report's core viewpoints are divided into three main parts: **1. Vertical Integration is the Underlying Flywheel Engine**: Tesla continues to extend vertically in battery materials (the only lithium refiner in North America), 4680 cells, LFP batteries, semiconductors (TeraFab JV), Gigacasting, and Unboxed manufacturing. This not only reduces COGS and enhances pricing power but also builds a closed loop of cross-business data and technology reuse—for example, using car factories to test Optimus verifies industrial-grade reliability while feeding back cost reductions to automotive production lines. **2. Robotaxi and Optimus Form Dual Main Growth Narratives**: - Robotaxi has entered early commercialization stages (Austin, Dallas, Houston). Cybercab is a dedicated platform with a target cost below $30,000, significantly lower than Waymo vehicles ($150,000–$200,000), offering substantial unit economics advantages; - Optimus represents a longer-cycle, higher-barrier disruptive opportunity. Its greatest structural advantage lies in being 'its own first and largest customer,' allowing it to bypass the long sales cycles and ROI validation challenges faced by traditional robot manufacturers, directly accumulating data, iterating hardware, and compressing costs through internal factory deployments. **3. Energy Storage is a Robust Growth Pillar**: Megapack is evolving from tool-type hardware to financial assets. Autobidder software enables real-time participation in electricity market bidding, while Powerhub transforms it into virtual power plant nodes. Combined with IRA act subsidies, explosive power demand from data centers, and rigid storage gaps due to aging grids, this business faces intensified competition but remains one of the company's fastest-growing and highest-margin hardware segments thanks to its software and upstream integration capabilities.

Analysis framework

The report adopts a three-layer analytical framework: 'TAM—Flywheel—Execution Risk': First, independent TAM calculations are performed for each emerging business (Robotaxi, Optimus, Energy Storage), covering dimensions such as geographic distribution, penetration curves, cost reduction paths, and regulatory access rhythms; Second, 'Feedback Loops' are identified and quantified, such as FSD data feeding back into Robotaxi algorithms, expanding Robotaxi fleets driving down marginal costs and promoting FSD subscription adoption, and Optimus deployment reducing labor costs in car factories and releasing more R&D resources; Finally, returning to fundamentals, long-term value realization is verified through financial models—the report explicitly states that the EPS inflection point is expected after 2028, with EPS potentially reaching $7.50 by 2030 (CAGR >50%). Therefore, current valuations rely primarily on long-term visions rather than short-term performance, while short-term stock price fluctuations will be dominated by key milestones for Robotaxi and Optimus (such as Cybercab S-curve ramp-up, Optimus Gen 3 mass production, and FSD v15 implementation).

Methodology notes

• Industry/Industrial Analysis FrameworkSupply and Demand Framework

  The report's calculation of Robotaxi TAM is based on a VMT (Vehicle Miles Traveled)-driven supply and demand model, considering urban travel demand, price elasticity, waiting time thresholds (<5 minutes), and the amplifying effect of fleet density on network effects.

  This method treats Robotaxi as a service. Its market size depends not only on the number of vehicles but also on whether incremental travel demand can be stimulated at sufficiently low prices and short waiting times (the Jevons Paradox), achieving a three-level leap from 'replacing ride-hailing' to 'replacing private cars' and then to 'replacing public transport.'

• Industry/Industrial Analysis FrameworkVolume-price decomposition

  The economic feasibility analysis of Optimus uses a volume-price comparison method of 'unit task cost vs. hourly labor wage,' combining regional wage levels and robot deployment cost predictions to determine commercialization timelines across different industries and regions.

  This method does not simply look at the robot's selling price but calculates its hourly operating cost (including depreciation, maintenance, and energy consumption) and compares it with the labor cost of corresponding positions. Economic feasibility truly holds when the robot's cost is below a certain proportion of the hourly wage (e.g., 60%-70%). This explains why China's manufacturing sector becomes the breakthrough point first, while high-end services in Europe and the US may lag behind for years.

• Industry/Industrial Analysis FrameworkUpstream-Midstream-Downstream Transmission in the Industry Chain

  The report repeatedly emphasizes the bidirectional transmission relationship between upstream (lithium refining, cell manufacturing), midstream (vehicle/robot integration, software stack), and downstream (fleet operations, user subscriptions, distributed inference).

  For example, self-built LFP cell capacity upstream not only reduces Megapack costs but also alleviates FEOC compliance pressures under US IRA regulations; midstream advancements in FSD visual algorithms directly improve downstream Robotaxi safety ratings and regulatory approval speeds; meanwhile, the edge computing network formed by millions of vehicles downstream provides low-cost, high-concurrency infrastructure for midstream AI model training. This transmission makes vertical integration not just a cost tool but a technology accelerator.

• Cycle and Prosperity FrameworkJevons Paradox

  The report frequently cites the Jevons Paradox (efficiency improvements lead to increased total resource consumption) to explain the potential amplifying effect of Robotaxi and Optimus on total transportation volume and industrial labor demand.

  Traditional thinking suggests autonomous driving will reduce vehicle ownership. However, the report points out that lower travel costs will stimulate a large amount of previously suppressed demand (such as night travel, inter-city commuting, and elderly travel), leading to an increase in total VMT. Similarly, robots lowering unit labor costs in factories may encourage enterprises to expand capacity and add production lines, thereby increasing total procurement demand for robots. This paradox is a key perspective for understanding their TAM expansion logic.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• TSLA.US

  Core asset, carrier of all TAMs and flywheel effects

  Strengths

  World's largest real-world driving dataset (~10 billion miles), leading AI training infrastructure (Cortex cluster), unparalleled depth of vertical integration (from lithium mining to AI chips), organizational execution history record (Giga Shanghai 12-month投产)

  Weaknesses

  Brand damaged by CEO's politicized behavior (brand value dropped 36% in 2026), Cybertruck deliveries below expectations, HW3 owners facing mandatory hardware upgrades, China-US geopolitical tensions affecting Chinese robot market access

  Comparison

  Compared to Waymo (strong perception, weak cost), Figure AI (strong industrial scenarios, weak consumer ecosystem), Unitree (cost leadership but insufficient technical depth), Tesla is the only full-stack player possessing scale, cost, software, and ecological synergy

  Risks

  Regulatory approvals falling short of expectations, delays in Optimus mass production schedule, public opinion crises triggered by Robotaxi safety incidents, delays in AI5 chip mass production, personal risk events involving Musk

Key data

• 2035 SoTP Valuation$3.9TSum of five TAMs: Automotive $400B, FSD Subscription $300B, Energy Storage $80B, Robotaxi $1.6T, Optimus $700B, Infrastructure Licensing $800B
• Robotaxi Target Cost$0.30/mileLong-term goal for Cybercab; current industry average is approx. $2.00–3.00/mile
• Optimus Mass Production Cost Target<$30K/unitTarget under scale production (>1 million units/year); Gen 3 aims for a 22-degree-of-freedom hand design
• Global Energy Storage Capacity Forecast for 20351,231 GWhApprox. 4x growth from 2025 (701 GWh), CAGR 29%
• 2030 EPS Forecast$7.50250% growth from $2.15 in 2027, CAGR >50%

Impact & implications

This report implies for the market: Tesla has transcended being merely an EV manufacturer, evolving into a 'physical intelligence platform company' centered on AI, with hardware as the carrier and data as fuel. Its value is no longer determined solely by vehicle sales but depends on non-traditional metrics such as Robotaxi network density, Optimus penetration rate in factories, and Megapack installation volume in data centers. Investors need to switch analytical paradigms from 'how many cars sold' to 'how many autonomous miles driven,' 'how many robots deployed,' and 'how many MWh of virtual power plants dispatched.' Meanwhile, the report warns that if any link (such as delayed regulatory approval, insufficient hand reliability for Optimus, or loss of control over Cybercab manufacturing costs) experiences significant deviation, the entire valuation system may be re-evaluated.

Risks

• Regulatory approval is the biggest bottleneck for the commercialization of Robotaxi and Optimus, especially against the backdrop of lacking unified federal standards and fragmented interstate policies in the US
• Musk's political activities continue to damage the brand, resulting in an estimated loss of 1–1.3 million potential sales from 2022–2025
• HW3 owners cannot achieve fully autonomous driving and require hardware upgrades, potentially triggering user dissatisfaction and trust crises
• The supply chain for Optimus core components (such as harmonic reducers, high-precision hand actuators) is not yet formed, posing risks to mass production ramp-up
• AI5 chip mass production progress lags behind original plans; if it fails to ramp up as scheduled by mid-2027, it will drag down the performance iteration of Robotaxi and Optimus

What to watch

• Speed of Cybercab S-curve ramp-up (starting from H2 2026)
• KPIs for official release of Optimus Gen 3 and initial internal factory deployments (uptime, cost displacement)
• Whether FSD v15 achieves large-scale unsupervised deployment by end of 2026
• Legislative progress of the US Motor Vehicle Modernization Act (raising the federal exemption cap from 2,500 to 90,000 vehicles)
• Actual implementation pace of LGES Lansing and Samsung SDI LFP cell supply agreements