YK Research

Tesla, Inc. (TSLA)

The Convergence Machine. Robotics, Autonomy, and the SpaceX Combination.

Published: 7 April 2026 · YK Research

The Mispricing

Tesla is priced as an EV company in secular decline. Automotive revenue fell 10% YoY in 2025. Deliveries dropped 16%. Operating income halved over two years. JPMorgan has a $145 target. The P/E is 327x trailing earnings.

They’re right about the auto business. They’re wrong about what Tesla is becoming.

The market is valuing Tesla on the business it built (cars) and ignoring the three businesses it’s assembling: autonomous transport, humanoid robotics, and vertically-integrated space-industrial infrastructure. Each is pre-revenue or early-revenue. Each has a plausible path to revenue scales that dwarf the current auto business.

The SpaceX IPO, expected June 2026, creates a forcing function that could crystallize the convergence thesis in a way the market has never had to price before.

🔍
The mispricing isn't that Tesla is cheap. At $353 on trailing auto earnings, it's expensive. The mispricing is that the market treats autonomy, robotics, and the SpaceX combination as optionality worth approximately zero, when each is approaching commercial reality on a 12-24 month horizon.

Company Snapshot. Four Companies in One Ticker

Market Cap
$1.37T
Price
$353
P/E (Trailing)
327x
Revenue (2025)
$94.8B
Auto Revenue
$69.6B
Energy Revenue
$12.8B
Deliveries (2025)
1.64M
FSD Miles Driven
9B+
Robotaxi Paid Miles
700K+
2026 Capex
$20B+
Optimus Units (2025)
~150
Analyst Consensus
$403 (Hold)

1. Cars, The Decoy

The EV business is mature and margin-compressed. $69.6B automotive revenue in 2025, down 10% YoY. 1.64M deliveries. Gross margins under pressure from price cuts and competition, particularly from BYD in China and legacy OEMs in Europe. Regulatory credit revenue dropped 50% in Q2 2025.

This is the business Wall Street knows how to model. Analysts slap a PE multiple on auto earnings and declare Tesla overvalued. They’re right: the auto business cannot support a $1.37T market cap. But it’s increasingly the least interesting thing Tesla does.

Auto Revenue (2025)
$69.6B
YoY Change
-10%
Deliveries
1.64M
Op. Income (2025)
$4.4B

2. Energy, The Quiet Compounder

$12.8B revenue in 2025, up 27% YoY. Megapack deployments exceeded 60 GWh. Gross margins above 28%. Shanghai Megafactory hit 40 GWh annualized run rate by March 2026. Houston Megafactory (50 GWh initial capacity) ramps late 2026. Combined capacity by 2027: 120+ GWh across three gigafactories.

Energy Revenue
$12.8B
YoY Growth
+27%
Gross Margin
28%+
2027 Capacity
120+ GWh
🔍
This business alone, at current growth rates and margins, would justify a $100-150B standalone valuation as a pure-play grid storage company. It's growing faster than the auto business ever did at comparable scale, and it's doing it at higher margins.

The energy segment grew from 6% to 13% of Tesla’s revenue in two years. By 2027, with three Megafactories online and Megapack 3 shipping, this could approach $20B+ in annual revenue and 20% of Tesla’s total profits.

3. Autonomy, The Bet That’s Landing

Robotaxi: Live in Austin

Tesla launched driverless ride-hailing in Austin, June 2025. By February 2026: ~700,000 paid robotaxi miles. The Austin fleet has 30-40 Model Y vehicles operating fully autonomously, monitored remotely 24/7. 60+ vehicles spotted in Phoenix for expansion testing.

FSD Miles
9B+
Safety
7x Safer
Paid Robotaxi Miles
700K+
Crash Rate
1 per 5-7M mi

Cybercab: Production Started

Cybercab production began at Giga Texas. First unit off the line mid-February 2026. Drone footage from March 25 showed 36+ Cybercabs on factory grounds. Mass production ramping from April 2026, targeting hundreds weekly. Cost under $30K/unit. Late 2026 deployment in Austin and select US cities.

Cybercab Unit Cost
<$30K
Production Start
Apr 2026
2027 Target
2-4M/yr
Waymo Vehicle Cost
$150K+
🔍
At $0.50/mile operating cost (no driver), Tesla would undercut Uber/Lyft structurally while capturing 100% of the fare minus vehicle depreciation and energy. The ride-hailing TAM exceeds $200B globally.

FSD Regulatory Momentum

March 2026 NHTSA ruling cleared paths for unsupervised testing and robotaxi deployment. FSD v14.3 in testing for wide release. Expansion planned to 7+ US cities in H1 2026. Europe: supervised rollout advancing, national approvals pending.

4. Robotics, The Moonshot That’s Walking

Optimus Gen 3 prototypes are operational. Low-volume production begins summer 2026 at Fremont (converted from Model S/X lines). Initial deployment is internal: factory tasks, data collection, imitation learning. First B2B sales late 2026 at $100K+ per unit.

Height / Weight
5'8" / 125 lbs
Hand DOF
22
Carry Capacity
20 lbs
Deadlift
150 lbs
2026 Target
50K-100K units
B2B Price
$100K+
Scale Price Target
$20-30K
Long-Term Production
1M+/yr

The Competitive Landscape

Tesla Optimus
Self-funded

$20B 2026 capex. ~150 units shipped 2025. Manufacturing DNA from millions of cars.

Figure AI
$39B valuation

Sep 2025 raise. Zero revenue. Zero production. Pure venture bet.

Unitree (China)
5,000 shipped

$5,900-$13,500 price points. Volume leader. China’s structural cost advantage.

🔍
Musk claims Optimus could represent 80% of Tesla's future value. The humanoid robot TAM is projected at $38B by 2035 by consensus. Musk's target of 10M units/yr at $20-30K implies $200-300B in annual revenue. The truth is probably somewhere between the conservative analysts and Musk's fever dreams. Even the midpoint is a very large number.

AI Infrastructure: Cortex 2.0

The AI training cluster powering both FSD and Optimus. 250MW online April 2026, full 500MW by mid-2026. Powered by the new AI5 chip (5x memory bandwidth vs AI4), manufactured in-house at TERAFAB on the Giga Texas campus.

The Moat: Manufacturing + Data + Vertical Integration

Manufacturing

  • Builds own batteries, motors, inverters, chips
  • TERAFAB: 1 TW compute output/yr target
  • Fremont auto lines → robot production (unique conversion capability)
  • Muscle memory from millions of electromechanical products

Data Flywheel

  • 9B+ FSD miles (no competitor comes close)
  • Shared perception/planning stack for FSD + Optimus
  • Every Tesla on the road trains both products simultaneously
  • Imitation learning from factory robot deployments

Vertical Integration

  • TERAFAB: AI5 chips for FSD/Cybercab/Optimus + D3 for SpaceX
  • Cortex 2.0 serves three business lines
  • One battery cell line serves EVs, Megapack, Cybercab
  • Dramatically lower incremental capex per new product line

The SpaceX Catalyst: The $3 Trillion Convergence

This is the piece the market hasn’t priced at all.

SpaceX filed confidentially for an IPO on April 1, 2026. Codename: “Project Apex.” Target valuation: $1.75 trillion. 21 underwriting banks. $75 billion raise. Expected June 2026.

SpaceX Target
$1.75T
IPO Raise
$75B
SpaceX Revenue
$15-16B
SpaceX Profit
$8B

Why the Merger Is Coming

1. TERAFAB is already joint. The chip fab at Giga Texas serves both Tesla and SpaceX. AI5 and D3 chips manufactured on the same campus. Supply chain integration is already happening.

2. Optimus robots in SpaceX factories. Tesla’s robots are deployed at Giga Texas, which houses both operations. Cross-pollination of manufacturing talent is already a reality.

3. Validated mark-to-market. Once SpaceX is public, both companies have daily market-priced valuations. This eliminates the governance objection that has historically blocked combination.

4. The space-industrial frontier. Tesla builds robots. SpaceX provides transport. Lunar manufacturing becomes feasible. Mass drivers can ship processed ore at near-zero marginal cost. All you need is robots and cheap transport.

🔍
A combined Tesla-SpaceX entity at current valuations would be worth ~$3.1T, the 4th largest company globally. But the combination unlocks value greater than the sum of parts through shared chip fabs, shared AI infrastructure, shared manufacturing expertise, and access to the space-industrial frontier.
Chamath put the merger probability at '99.999%' on All-In Podcast, April 2026

The Catalyst Stack

Apr 2026

Cybercab mass production ramp at Giga Texas. Cortex 2.0 (250MW) online. Hundreds of Cybercabs/week target.

Jun 2026

SpaceX IPO. First validated public price for the Musk ecosystem. $75B capital raise. Institutional reallocation begins.

Summer 2026

Optimus Gen 3 low-volume production at Fremont. First robots deployed in Tesla factories at scale.

H2 2026

Houston Megafactory ramps (50 GWh). Megapack 3 deliveries begin. Energy segment approaches $20B run rate.

Late 2026

First Cybercab robotaxi rides. First B2B Optimus sales ($100K+/unit). Robotaxi expansion to 7+ US cities.

2027

High-volume Optimus (1M+/yr target). Cybercab 2-4M annual. Energy 120+ GWh. Tesla-SpaceX merger goes from speculation to board-level.

Valuation: Sum of Parts

The bear case treats Tesla as one business. The bull case treats it as four (or five, with SpaceX).

SegmentRevenueGrowthComparableMultipleImplied Value
Automotive$69.6B-10% YoYMature automaker1-2x rev$70-140B
Energy$12.8B+27% YoYHigh-growth utility5-8x rev$65-100B
Autonomy / RobotaxiPre-scaleLaunchedWaymo ($100-175B in GOOGL SOTP)Comp-based$100-250B
RoboticsPre-revenuePrototypingFigure AI ($39B, zero prod.)Comp-based$50-150B
SpaceX OptionN/AN/ANot in stock todayUnpricedTBD
Conservative (no SpaceX)
$285-640B
-53% to -79%

Auto business declines, new segments underdeliver. Market reprices to fundamentals.

Base (no SpaceX)
$450-800B
-42% to -67%

Energy compounds. Robotaxi scales modestly. Optimus stays internal. Stock overvalued near-term.

Bull (with SpaceX)
$800B-1.5T+
-42% to +10%

Convergence thesis plays out. SpaceX merger materializes. Multiple expansion on new businesses.

🔍
The thesis is that the catalyst stack shifts the market from pricing the bear/base case to pricing the bull case over the next 12-18 months. At the bull end, today's price is roughly fair. This is a momentum-and-catalysts bet, not a value bet.

What Breaks It

RiskSeverityProbabilityImpact on ThesisMitigant
Autonomy stallsHIGH30%Largest valuation chunk evaporates. Cybercab becomes stranded asset.9B+ miles, 7x safety edge, NHTSA regulatory win. Watch FSD v14.3 rollout.
Optimus is vaporwareHIGH25%Robotics valuation → near-zero. Fremont conversion wasted.Gen 3 walking. Fremont production line converting. Track unit counts quarterly.
China competition crushes marginsMEDIUM40%Both cars and robots undercut. Unitree at $5,900 vs Optimus at $20-30K.Tesla's data flywheel + vertical integration vs. China's cost advantage. Different markets initially.
SpaceX merger blockedMEDIUM20%Convergence thesis loses its most powerful catalyst.Even without merger, TERAFAB and cross-pollination continue de facto.
Musk brand toxicity spreadsMEDIUM35%EU sales already depressed. Could spread to US/China, accelerating auto decline.New businesses (energy, robotaxi) are less brand-sensitive than consumer cars.
Auto cash generation collapsesHIGH20%$20B capex for AI/robotics depends on auto cash flow. If deliveries < 1.4M, math breaks.Energy business provides growing alternative cash source. $9.5B balance sheet cash.

Position Sizing & Timing

Entry

Accumulate below $300. Q1 delivery miss + JPMorgan $145 target suggest more near-term downside. The stock has been below $300 within the last 12 months. Patience.

Size

3-5% of portfolio. High-conviction directional bet with massive uncertainty. Bimodal outcome: $600+ by 2028 or $150-200 on auto fundamentals alone.

Catalysts to Watch

  • Cybercab weekly production (target: hundreds/week by summer)
  • SpaceX IPO pricing and post-IPO trading
  • Optimus factory deployment count
  • Robotaxi city expansion beyond Austin
  • Energy segment quarterly revenue (inflection above $5B/quarter)

What Changes My Mind

  • FSD safety metrics deteriorating
  • Cybercab production delays beyond Q3 2026
  • Optimus production below 10,000 units by year-end 2026
  • SpaceX IPO pricing below $750B
  • Auto deliveries below 1.4M in 2026

This is not the thesis of a value investor. This is the thesis of someone who believes that manufacturing capability, data flywheels, and vertical integration compound nonlinearly when applied across multiple product categories simultaneously. Tesla is the only company in the world attempting to build cars, robots, grid-scale batteries, autonomous transport, and semiconductor fabs while its sister company builds rockets and a backup internet. The question isn’t whether that vision is ambitious. It’s whether any of it actually works. The next 18 months will tell us.