Tesla (TSLA) Q4 2025 earnings review
Margins Hit 2-Year High, But Tesla Bets the Farm on Autonomy
Tesla delivered a tale of two companies in Q4. The legacy auto business shrank—deliveries fell 16% YoY to 418K units and automotive revenue dropped 11%—yet gross margin surged to 20.1%, the highest since Q1 2023. The secret: a better regional mix (APAC/EMEA strength), improving auto margins ex-credits (17.9% vs 13.6% a year ago), and a record-breaking Energy segment ($1.1B gross profit). But the real story is what comes next. Management announced Model S/X discontinuation, a $2B xAI investment, plans for a semiconductor TerraFab, and CapEx exploding from ~$8.5B to >$20B in 2026. Tesla is betting everything on becoming a robotics and AI infrastructure company. The first unsupervised robotaxi rides with no safety monitor or chase car are now happening in Austin.
🐂 Bull Case
Unsupervised robotaxi rides are now live in Austin with no safety monitor and no chase car. Management expects to operate in dozens of major US cities by end of 2026. FSD subscriptions reached 1.1M (up 38% YoY), and the transition to subscription-only will create recurring revenue. If autonomy scales as planned, the entire vehicle fleet becomes a robotaxi network asset.
Automotive gross margin ex-credits jumped 430bp YoY to 17.9%—the best since Q2 2023. This came despite lower volumes, indicating genuine cost improvements and favorable mix. Energy gross profit hit a record $1.1B. Total gross margin at 20.1% is the highest in two years, proving Tesla can be profitable without relying on regulatory credits.
Energy storage deployed 14.2 GWh in Q4 (record) and 46.7 GWh for 2025 (up 49% YoY). Revenue grew 27% YoY to $12.8B. With Megapack 3 and Megablock launching in 2026, plus AI data center demand, this segment is becoming a material profit driver with $2.7B gross profit for the year.
🐻 Bear Case
CapEx is exploding from ~$8.5B to >$20B in 2026—and that excludes potential solar and semiconductor fabs. Tesla is simultaneously ramping six factories (lithium refinery, LFP, CyberCab, Semi, Megafactory Houston, Optimus). Management acknowledges this is unprecedented and could strain cash despite the $44B balance sheet. Free cash flow in 2025 was only $6.2B.
Vehicle deliveries fell 9% YoY in 2025 to 1.64M units—the first annual decline in Tesla's history. Q4 deliveries of 418K were down 16% YoY and 16% QoQ. Model S/X discontinuation removes ~50K annual units. While CyberCab is coming, it won't produce meaningful volume until late 2026 at earliest, leaving a gap.
GAAP net income fell 61% YoY in Q4 to $840M and 46% for the full year to $3.8B. EPS of $0.24 was down 60% YoY. While some of this is Bitcoin mark-to-market and FX, operating income also fell 11% YoY despite the margin expansion. OpEx surged 39% YoY as R&D and AI spending accelerate.
⚖️ Verdict: ⚪
Neutral with positive bias. The margin improvement is genuinely impressive and the autonomy thesis is finally showing real-world proof points with unsupervised robotaxi rides. However, the sheer scale of investment ($20B+ CapEx, six factory ramps, potential TerraFab) creates significant execution risk. Tesla is transitioning from a car company to an AI/robotics infrastructure company—a bet that could pay off enormously or strain the balance sheet. The stock requires conviction on autonomy; the legacy auto business alone doesn't justify current valuations.
Key Themes
Unsupervised Robotaxi Now Live—The Autonomy Bet Becomes Real
In January 2026, Tesla began removing safety monitors from customer robotaxi rides in Austin—with no chase car following. This is the first time any company has achieved fully unsupervised, paid autonomous rides in a non-geofenced environment using a mass-market vehicle platform. Musk stated they have 'well over 500' robotaxi vehicles operating between Austin and the Bay Area, with plans to double monthly. The goal is to reach 'probably somewhere between a quarter and half of the United States by the end of the year, pending regulatory approval.' If federal preemption passes, this accelerates dramatically. Each successive FSD release will reduce driver monitoring requirements. Management expects personally-owned Teslas to join the fleet starting in 2026, potentially earning owners more than their lease payments.
Model S/X End of Life—Fremont Converts to Optimus Factory
Musk announced Model S and X will end production next quarter after over a decade. The Fremont S/X line will convert to an Optimus factory with a target of 1 million humanoid robots per year. This is a defining strategic pivot: Tesla is explicitly choosing to exit the premium sedan/SUV segment to build robots. Management called it 'time to bring the SX programs to an end and shift to an autonomous future.' For context, S/X combined delivered ~51K units in 2025, representing ~3% of volume but occupying valuable Fremont floor space. The Optimus Gen 3—'the first design meant for mass production'—will be unveiled in Q1 2026, with production planned before end of 2026.
$20B+ CapEx Year—Six Factory Ramps Simultaneously
Tesla's CapEx is exploding from approximately $8.5B in 2025 to over $20B in 2026. CFO Taneja listed six factories ramping simultaneously: lithium refinery (Corpus Christi), LFP cell factory (Nevada), CyberCab (Texas), Semi (Nevada), Megafactory Houston, and Optimus (Fremont). On top of this, significant investment goes to Cortex 2 AI compute infrastructure and existing factory expansions. Critically, this $20B+ figure excludes potential solar cell manufacturing and the proposed semiconductor TerraFab. Management acknowledged funding will initially come from the $44B cash position, with robotaxi fleet financing from banks once cash flows are proven. This is the most capital-intensive period in Tesla's history.
TerraFab Vision—Tesla Building Its Own Chip Fab
Musk revealed plans for a 'TerraFab'—a large-scale semiconductor fab integrating logic, memory, and packaging, built domestically in the US. He cited supplier constraints and geopolitical risk as motivations: 'when I look ahead at what's the limiting factor for Tesla growth... three or four years out, I think it actually is chip production.' He noted there are 'literally zero' advanced memory fabs at scale in the US. The AI5 chip (production 2027) and AI6 (targeting 2028) are being designed in-house, with AI5 expected to be '50x better than AI4.' Musk acknowledged 'fabs are really hard' but said Tesla builds hard things 'out of desperation.' This would be a multi-billion dollar commitment not yet included in the $20B CapEx guidance.
Energy Segment Hits Record After Record
Energy storage deployed 14.2 GWh in Q4 (record) and 46.7 GWh for 2025 (up 49% YoY). Revenue grew 27% YoY to $12.8B with gross profit reaching a record $1.1B in Q4 alone. For the full year, Energy gross profit was $2.7B. The segment now contributes 13% of total revenue and is growing faster than the auto business. Megapack 3 and Megablock production begins in 2026 at Megafactory Houston (targeting 50 GWh capacity). The Powerwall network now exceeds 1 million installed units and supported 89,000 Virtual Power Plant events in 2025. However, CFO Taneja warned of expected margin compression from 'increased low-cost competition... policy uncertainty, and the cost of tariffs.'
Automotive Margin Recovery Despite Volume Decline
Automotive gross margin excluding regulatory credits improved to 17.9% in Q4, up from 15.4% in Q3 and 13.6% a year ago. This 430bp YoY improvement came despite a 16% decline in deliveries. Management attributed this to favorable regional mix (proportionately more APAC/EMEA deliveries where FSD isn't available yet) and cost improvements. Total automotive gross profit was nearly flat sequentially at $4.9B despite the delivery decline. The new Model Y variants ramped globally, and Tesla now produces battery packs with internally-manufactured 4680 cells. However, the transition to FSD subscription-only 'will impact automotive margins in the short term' as high-margin upfront purchases phase out.
$2B xAI Investment—Related Party Transaction
Tesla announced a $2B investment in xAI's Series E round at 'market terms consistent with those previously agreed to by other investors.' A framework agreement was also signed for 'potential AI collaborations between the companies.' Musk justified this as shareholder-driven: 'We just had a lot of investors ask us to do this.' The stated rationale is that Grok can optimize large autonomous fleet management and Optimus robot coordination—acting as an 'orchestra conductor.' While there may be legitimate synergies, this is a significant related-party transaction that uses Tesla cash to invest in another Musk-controlled entity. The investment is subject to regulatory approval with expected Q1 2026 closing.
CyberCab Production Begins April 2026
The purpose-built Cybercab autonomous vehicle—a two-seater with no steering wheel or pedals—will begin production in April 2026. Musk expects CyberCab to eventually represent 'several times more' volume than all other Tesla vehicles combined, given that 90% of vehicle miles traveled are with two or fewer passengers. The vehicle is optimized for 'minimum cost per mile' and higher duty cycles (50-60 hours/week vs 10-11 for personal vehicles). The 'unboxed' manufacturing method targets revolutionary efficiency with an eventual goal of one vehicle every five seconds. Near-term, CyberCab will seed Tesla's owned robotaxi fleet before eventually being sold to third parties.
Operating Expenses Surge 39% YoY
Operating expenses reached $3.6B in Q4, up 39% YoY and 5% sequentially. R&D grew 40% YoY to $1.8B as AI, Optimus, CyberCab, and Semi development accelerated. SG&A grew 26% YoY to $1.7B. Stock-based compensation surged to $954M in Q4, up from $579M a year ago, driven by AI employee retention and the 2025 CEO Performance Award (Operation Maestro) becoming probable. Restructuring charges of $162M added further pressure. Management explicitly stated this elevated spending 'will continue for the full year 2026' as Tesla invests in the AI and robotics transition.
Vehicle Deliveries: First Annual Decline in Company History
Tesla delivered 1.64M vehicles in 2025, down 9% YoY—the first annual decline ever. Q4 deliveries of 418K were down 16% YoY and 16% QoQ after Q3's 497K record. Production of 434K also declined. Model 3/Y deliveries fell 7% YoY while 'Other Models' (S/X/Cybertruck) fell 40%. Days of supply increased to 15 from 12 a year ago. Management attributed Q4 weakness to Q3 demand pull-forward (ahead of higher consumer credit cliff), but also noted record deliveries in smaller markets (Malaysia, Norway, Poland, Saudi Arabia, Taiwan) and ended 2025 'with a bigger backlog than in recent years.' The focus has clearly shifted from volume to autonomy.
Mission Update: 'Amazing Abundance'
Musk updated Tesla's mission from sustainable energy to 'amazing abundance,' signaling a philosophical shift. He envisions 'a future of universal income, not universal basic income, but universal high income' driven by AI and robotics. This framing positions Tesla not just as a car or energy company, but as an infrastructure provider for post-scarcity economics. The company explicitly stated it is transitioning 'from a hardware-centric business to a physical AI company.' Whether this vision is achievable or simply marketing for the investment thesis, it represents a significant messaging evolution and sets expectations for what success looks like.
GAAP Net Income Down 46% YoY Despite Margin Gains
Full-year GAAP net income fell to $3.8B from $7.1B in 2024, a 46% decline. Q4 net income of $840M was down 61% YoY. While gross margin improved, operating income still fell 38% YoY for the full year due to the 23% increase in operating expenses. Below the line, Q4 was hit by a $592M loss in 'Other income' driven by Bitcoin mark-to-market (BTC fell 23% in Q4) and unfavorable FX on intercompany borrowings. Non-GAAP EPS of $1.66 for 2025 was down 28% YoY. The disconnect between margin improvement and earnings decline reflects the massive investment phase Tesla is entering.
Other KPIs
FCF surged 74% YoY from $3.6B in 2024, despite the earnings decline. Operating cash flow of $14.7B was essentially flat YoY, while CapEx fell 25% to $8.5B as spending was deferred to 2026. Q4 FCF of $1.4B was down 30% YoY but positive for the fifth consecutive quarter. This cash generation supports the $44B balance sheet heading into the massive 2026 investment year. However, with CapEx guided to >$20B, FCF will likely turn negative in 2026 unless operating cash flow improves significantly.
Record cash position, up 21% YoY and $7.5B higher than year-end 2024. The sequential increase of $2.4B was primarily from positive FCF. Total debt (recourse + non-recourse) was $8.2B at year-end, giving Tesla a net cash position of ~$36B. This liquidity is essential given the >$20B CapEx plan for 2026. Management indicated they will use internal resources initially, then leverage the robotaxi fleet's cash flows for bank financing once proven. The xAI investment ($2B) and any TerraFab commitment would further draw on this balance sheet.
Adjusted EBITDA declined 4% YoY but margin improved 17bp to 16.7% from 16.9% a year ago. For full year 2025, EBITDA was $14.6B (15.4% margin), down 9% YoY. The margin trajectory shows improvement: Q1's 14.6% → Q2's 15.1% → Q3's 15.0% → Q4's 16.7%. This metric excludes SBC, depreciation, and Bitcoin volatility, providing a cleaner view of operating performance. The improvement supports management's claim that underlying profitability is strengthening even as GAAP earnings decline.
Services gross profit improved YoY driven by Parts Sales and Supercharging. However, margin compressed from 10.5% in Q3 to 8.8% in Q4 due to higher employee costs at service centers in preparation for fleet growth. Notably, robotaxi business costs are now included in this segment, though 'not material' at current scale. The Supercharger network grew to 8,182 stations (+17% YoY) and 77,682 connectors (+19% YoY). Tesla Insurance expanded to Florida, with FSD usage potentially offsetting subscription costs through premium discounts.
Regulatory credits declined 22% YoY from $692M but increased 30% QoQ from $417M. For full year 2025, credits totaled $1.99B, down 33% from $2.98B in 2024. This decline reflects tighter emission standards reducing the value of credits and increased EV competition generating more supply. CFO Taneja previously warned that IRA-related policy changes could reduce credits further. Excluding credits, automotive revenue fell 14% YoY, highlighting the underlying demand challenge.
Guidance
Massive acceleration from approximately $8.5B in 2025. The increase funds six factory ramps (lithium refinery, LFP, CyberCab, Semi, Megafactory Houston, Optimus), Cortex 2 AI compute expansion, existing factory capacity additions, and robotaxi/Optimus fleet expansion. Critically, this excludes potential solar cell manufacturing and the proposed semiconductor TerraFab. Management emphasized this is deliberate investment for 'an epic future' and will be executed in a 'capital-efficient manner.' This represents the most aggressive investment phase in Tesla's history.
Production of the purpose-built autonomous taxi begins in April. The vehicle has no steering wheel or pedals—'this car either drives itself or it does not drive.' Management expects CyberCab to eventually represent 'several times more' annual production than all other vehicles combined, given 90% of miles traveled involve two or fewer passengers. The manufacturing line targets revolutionary efficiency via the 'unboxed' method. Near-term volumes will follow a typical S-curve ramp.
Optimus Gen 3—'the first design meant for mass production'—unveils in Q1 2026. Production line installation is underway at Fremont (converted S/X line). Management targets eventual capacity of 1 million robots per year at Fremont alone. The S-curve ramp will be 'stretched out' because the entire supply chain is new. Optimus will learn tasks by observing humans, verbal description, or watching videos. Musk emphasized Tesla is the only company with all three key capabilities: dexterous hand design, real-world AI, and manufacturing scale.
Currently operating in Austin (unsupervised) and Bay Area (with safety driver). Planned 1H 2026 expansion includes Dallas, Houston, Phoenix, Miami, Orlando, Tampa, and Las Vegas. Management expects to cover 'somewhere between a quarter and half of the United States by the end of the year, pending regulatory approval.' Federal preemption for autonomous vehicles would accelerate rollout. The fleet will grow by allowing personally-owned Teslas to join starting in 2026.
Cortex 2 is under construction at Gigafactory Texas to more than double onsite compute (in H100 equivalents) by 1H 2026. Current capacity exceeds 100,000 H100-equivalent GPUs. Management is pursuing AI5 chip production in 2027 (targeting 50x improvement over AI4) and AI6 in 2028. The in-house chip strategy is explicitly designed to avoid being 'supplier limited' on AI logic and memory, which Musk sees as the key constraint in 3-4 years.
Megapack 3 and Megablock production begins at Megafactory Houston in 2026. Backlog remains 'strong' and 'well-diversified globally.' However, CFO Taneja explicitly warned of expected margin compression from 'increased low-cost competition impacts to market from policy uncertainty, and the cost of tariffs.' The energy business faces headwinds even as volume grows, though the record Q4 gross profit of $1.1B provides a strong base.