THE VALUATION PROBLEM

The table below captures something unusual: ten firms, all covering the same company and the same quarter, producing price targets that range from $125 to $2,600. The spread has little to do with Q1: on the reported numbers, the analysts broadly agree. What they disagree on is which business Tesla becomes, and how soon.

The disagreement comes down to assumptions that are currently unknowable: how fast FSD scales, when robotaxi reaches commercial viability, what margins Optimus achieves at volume. None of these has an operating history at commercial scale, so none can be checked against actual results. Small differences in those assumptions compound into a delta of hundreds of billions in estimated value.

JPMorgan illustrated this point in June: a new analyst inherited the bank's Tesla coverage and the same filings, then lifted the firm's price target on the stock from $145 to $475, a 227.6% jump that reflected changed assumptions, not changed facts.

One thing the ten firms do not disagree on: today's earnings cannot justify today's valuation. At $425.30 on roughly 3.2 billion shares, Tesla's market capitalisation is around $1.4 trillion. Against trailing twelve-month diluted GAAP earnings of $1.09 per share, that price is a trailing P/E near 390, meaning that at the current rate of profit it would take more than three and a half centuries to earn back today's price. Nobody pays that for the business Tesla is today.

The price only makes sense if earnings multiply many times over.

WHAT TESLA IS BUILDING

Which businesses, exactly, is the market paying for? The split can be drawn precisely.

The left column generated $22.4 billion in Q1. Tesla does not separately disclose how much the right column contributes. FSD revenue is reported inside the broader Automotive line, which means valuing it requires exactly the kind of guesswork this report is trying to highlight.

At 1.28 million FSD users and the current US subscription price of $99 per month, the quarterly run-rate sounds straightforward, but Tesla does not disclose how many of those users are upfront purchasers versus monthly subscribers, so even that estimate requires an assumption. The $99 price is not fixed either: it could rise as the product improves toward unsupervised capability, or fall to drive adoption in price-sensitive markets, and each direction moves the revenue trajectory and the margin impact differently.

All $22.4 billion of Q1 revenue sits in the left column: $16.2 billion automotive, $2.4 billion energy generation and storage, and $3.7 billion services and other. The stock price, by contrast, is overwhelmingly driven by expectations about the right column, which investors call the platform thesis: the bet that Tesla's value will ultimately be driven by software, autonomy, and robotics rather than vehicle hardware. The right column earns almost nothing today. Generous figures place FSD revenue in the low hundreds of millions per quarter, robotaxi revenue is negligible this year per Musk on the Q1 call, and Optimus is not yet commercially operational.

The platform thesis gets most of the attention, but every future business line depends on the cars being produced and delivered today. FSD subscriptions, robotaxi, Optimus, and energy at scale all draw on what the vehicle business currently generates: its cash flow, its manufacturing base, and, for the automotive software products, its installed fleet.

The auto hardware numbers in Q1 show how durable that cash-generating foundation is, and where its strength comes from.

Automotive revenue was $16.2 billion in Q1, up 16% year-over-year, representing 72.5% of total revenue. That headline rate flatters a softer underlying picture, since it runs off a depressed Q1 2025 base and auto revenue actually slipped sequentially from Q4; the more useful question is where the growth came from, because the strength was heavily concentrated by region. EMEA carried the quarter: CFO Vaibhav Taneja pointed to strength there, with deliveries in France and Germany growing over 150% quarter-over-quarter, Giga Berlin reached record output of more than 61,000 units, and the order backlog at quarter-end was the highest for any first quarter in over two years. Global deliveries still rose 6% year-over-year to 358,023, enough to leave the cash-generating base the platform rests on intact.

A solid foundation is not the same as a transformed business, and that gap is the whole valuation problem. Tesla remains a strong automotive franchise by any conventional measure: more than 9.2 million vehicles delivered cumulatively, with production scale across three continents. On the economics, though, it is still a car company, and the margins of selling cars, however well run, fund the platform transition rather than justify a platform multiple. The foundation pays the bills; it does not justify the share price in itself.

FSD is where that starts to change. It is the first place the platform thesis shows up in actual revenue, with Tesla converting vehicle owners into software subscribers.

Active FSD Subscriptions reached 1.28 million in Q1, up 51% year-over-year, with 180,000 net additions in the quarter alone. The subscription shift driving that growth is structural.

On the Q4 2025 call, Taneja confirmed that nearly 70% of the 1.1 million FSD users at year-end were upfront purchasers, and that net additions would from then on come primarily via subscription. He was explicit that this shift "will impact automotive margins" in the short term, since subscription revenue is recognised monthly rather than as a single upfront payment.

The Q1 2026 call confirmed the shift in practice: Tesla removed the upfront purchase option in some markets during the quarter, and upfront purchases grew only 7% while subscriptions drove the bulk of the new additions. He also said subscriber churn is coming down and customers are driving more, both signs that existing subscribers are staying rather than leaving.

The move from upfront purchases to monthly subscriptions has not yet dented margins. Automotive gross margin excluding regulatory credits rose to 19.2% in Q1, up from 17.9% in Q4 2025, so the margin compression Taneja warned about still lies ahead rather than showing up in this quarter's numbers. As subscription revenue, recognised month by month, gradually replaces upfront purchase income, automotive gross margin should soften, and when it does it reads as the transition working rather than the business weakening.

Tesla also received partial regulatory approval for FSD in China during Q1, with broader approval hoped for by Q3 2026. In Europe, the path is unfolding differently from the one management described. The EU-wide approval Taneja guided for later in Q2 has not yet arrived; what has emerged instead is a country-by-country workaround.

The Netherlands granted FSD its type approval in April, and because individual member states can recognise another member's certification rather than wait for the European Commission, Lithuania, Estonia, Denmark, and Belgium have each approved FSD for use on their own roads since, five countries in roughly three months. Even without the EU-wide approval, member states keep signing off one by one: each market that opens grows the pool of vehicles where FSD can legally be offered.

Energy stands apart from the platform's other forward bets, already a working, profitable business rather than a promise. Tesla guides it to keep expanding, but unlike FSD and robotaxi its growth does not hinge on an autonomy breakthrough or regulatory approval: Megapack and Powerwall sell now, at gross margins that have ranged from roughly 29% to 39% over the past five quarters, comfortably above what the auto business earns. The growth ahead rests on adding manufacturing capacity and filling order books rather than on a technology still to be proven. What the energy segment lacks is predictability, because revenue arrives through large, project-based deliveries whose timing Tesla does not control quarter to quarter.

Q1 showed both sides of that. Deployments of 8.8 GWh were down 38% sequentially from Q4 2025's 14.2 GWh, a drop that reflects the timing of large, lumpy projects rather than any fall in underlying demand.

Taneja confirmed on the Q1 2026 call that full-year 2026 energy deployments are still expected to exceed the 46.7 GWh deployed in 2025, and described the order backlog as robust. That guidance sets a concrete bar for the rest of the year: with 8.8 GWh banked in Q1, the remaining three quarters have to deliver more than 37.9 GWh between them, an average above roughly 12.6 GWh each, to clear the 2025 total.

Q1's 8.8 GWh deployment came in well under that quarterly pace, so the full-year target leans on a second-half ramp that the new Houston capacity and Megapack 3 are meant to supply.

The manufacturing capacity behind that growth is still being built: the new Megafactory outside Houston, which will produce the next-generation Megapack 3, is guided to start production later this year.

Auto hardware, FSD, and energy all produce revenue that appears in the financial statements today. Robotaxi, however, is not yet scaled enough to show up in the financials. The service began in Austin in June 2025 and reached Dallas and Houston in April 2026, running on modified Model Y vehicles rather than dedicated Cybercab hardware. Waymo, by comparison, has run a driverless service since 2020 and now covers ten US cities on roughly 3,000 vehicles, the head start of several years of operation.

Tesla's bet is that a camera-and-software system built to work anywhere can scale faster than mapping cities one by one, which makes the speed of expansion, not the current city count, the number to watch, and that speed remains to be proven. Musk set expectations himself on the Q1 call: revenue "will not be super material this year." For valuation purposes, robotaxi is a 2027 story at the earliest.

Optimus sits even further out on the same timeline. Its first production line, at Fremont, is guided for late July or August 2026, which means that as of Q1 there is no operational line, no confirmed unit economics, and no revenue. Both robotaxi and Optimus are therefore long-duration optionality: the market is paying for them today based on what they could become, not what they currently earn. How much that optionality is worth is exactly what the analysts cannot agree on, and it is the single biggest reason their price targets run from $125 to $2,600.

If robotaxi and Optimus are years away from revenue, the place to look for management's conviction is capital investment. Tesla guided above $25 billion of capital expenditure for 2026, the largest commitment in its history, funding the infrastructure behind Cybercab volume production, the Fremont Optimus line, the Houston Megafactory, and Megapack 3. The size of that number signals that management intends to fund the transition, not just describe it.

Intentions, though, are tested in the cash flow statement: Q1's actual spend ran at a fraction of the guided pace, and whether the money follows the words is one of the cleanest tests the rest of the year will provide.

Tesla has repeatedly delivered outcomes that looked impossible a few years earlier: Model 3 production scaled when most analysts expected manufacturing failure, the Supercharger network grew into the world's largest EV charging infrastructure, and energy storage became a multi-billion-dollar segment. Investors who have watched that happen extend the benefit of the doubt to whatever Tesla promises next, and that trust is what lets the stock trade far above what the current business alone would justify.

The market is paying a platform price for a company whose reported revenue is still overwhelmingly automotive, on the strength of an improbable track record and the record capital management is now committing to the next set of ambitions. Each part of that case can be tested against evidence, and the first test is the simplest: whether the quarter holds up when you look past the headline figures.

So what did Q1 actually say?

WHAT THE Q1 2026 NUMBERS ACTUALLY SAY

Each major figure in the Q1 report, revenue, margins, profit, and cash flow, is checked in turn against what sits underneath it.

Start with revenue. Total revenue was $22.4 billion, up 16% year-over-year, and the growth rate flatters the result: the comparison base was soft, with Q1 2025 revenue having fallen 9% year-over-year, so part of this year's increase is recovery to prior levels rather than new ground.

One thing the revenue split cannot show is software progress.
FSD revenue is booked inside the automotive line rather than reported separately, so even if FSD grows into a substantial business, the reported mix would still read as a company that mostly sells cars.

Set the software question aside and the core car business shows a gap between revenue and volume. Deliveries were 358,023 vehicles, up just 6% year-over-year, while production ran to 408,386, up 13%. That gap has two consequences. The first is an inventory build: production outran deliveries by roughly 50,000 units, lifting global days of vehicle supply from 15 to 27, which sits awkwardly against the demand strength Taneja described and points to that strength being regional rather than broad. The second is on price: revenue up 16% on only 6% delivery growth means revenue per vehicle rose, which Tesla attributes partly to higher FSD sales and subscriptions.

That richer revenue per vehicle fed straight into margin. Auto gross margin was 21.1% on a GAAP basis and 19.2% excluding regulatory credits, the strongest ex-credits reading in eight quarters.

That breakout needs a second look, because Taneja attributed it to one-time benefits. On the call he named warranty write-downs of around $230 million and, in his words, "some relief on tariffs," which he did not quantify. Adjusting for those items places the underlying auto gross margin in the 17% to 18% range, roughly flat with Q4 2025. On the chart, the 19.2% reads as a clean breakout; strip out the one-time benefits and the underlying number sits back in the band the series has occupied for two years.

Energy gross margin was 39.5%, but more than $250 million of that came from one-time tariff recognitions from prior periods. Stripping those benefits places the normalised margin near 29%. Taneja's actual guidance was directional: on a normalised basis the company continues to expect energy margin compression from here, citing increasing competition and tariff impacts, with most battery cells procured from China.

The one-time tariff items pushed energy margin up this quarter; a separate accounting shift is about to push auto margin down. Tesla is moving FSD from one-time upfront purchases to monthly subscriptions, which spreads the same software revenue across many months instead of recognising it all in the quarter of purchase, so reported auto margin softens over the next several quarters.

The fall is an accounting effect, not a weakening business: one-time revenue is being replaced by recurring revenue, and recurring revenue is what the platform thesis needs. A softer margin print in Q2 and Q3 is what this transition looks like while it is underway.

GAAP net income was $477 million on $22.4 billion of revenue, a 2.1% net margin. Tesla, like most large companies, also reports an adjusted profit figure that strips out certain costs, and that version comes to $1.45 billion, three times the official number. Nearly all of the gap is one cost: stock-based compensation.

Stock-based compensation was $1.03 billion in Q1, up from $573 million a year ago, driven largely by the 2025 CEO Performance Award. That is more than Tesla's entire net income for the quarter. The adjusted profit figure treats this cost as if it does not exist, on the logic that no cash leaves the company. The dilution is borne by existing shareholders all the same: employees are paid in newly created shares, and every new share shrinks the slice of the company each existing share represents.

The choice between the two numbers matters less than it might seem. On the official figure, Tesla trades near 390 times trailing earnings; on its own adjusted figure, near 235 times. For scale, the S&P 500 trades around 25 times, so even the flattering number sits at roughly nine times the broad market and the official one near fifteen. A 2.1% net margin priced at those multiples is not a bet on the earnings Tesla reports today. It is a bet that a materially different and far more profitable company emerges, and the adjusted figure, which excludes a cost larger than the entire quarterly profit, is the weaker place to start that bet from.

Operating cash flow was $3.9 billion; after $2.5 billion of capital expenditure, free cash flow was $1.4 billion.

The headline FCF number, however, was aided by a $1.3 billion increase in accounts payable during the quarter, with days payable outstanding rising from 61 to 71 days. A meaningful portion of Q1's reported cash flow came from paying suppliers more slowly rather than from operations alone. A dollar of free cash flow from extending payables is lower-quality and reversible; a dollar from collecting core revenue is not. Tesla also issued $4.3 billion in gross new debt during the quarter, offset by $3.5 billion in repayments, for roughly $800 million in net new borrowing.

Taneja confirmed that the company expects negative free cash flow for the rest of 2026 as the capex ramp takes hold. The last time Tesla posted negative quarterly FCF was Q1 2024, at negative $2.5 billion, but that was a single-quarter event driven by an inventory build and a step-up in AI infrastructure spending. What Taneja is describing for 2026 is structurally different: multiple consecutive quarters of negative free cash flow driven by deliberate capital deployment, not a one-off swing.

Tesla is borrowing to fund the transition, not self-funding entirely from operations. The balance sheet is not under stress, but with negative free cash flow guided for the rest of the year, debt is likely to keep rising, and the cash flow profile will look worse before it looks better.

The quarter also carried a $2.0 billion equity investment in SpaceX, disclosed in a single line of the deck's cash commentary. The capital is not a passive stake: it anchors a partnership to build what Tesla calls the largest chip fab ever, meant to secure the compute supply its autonomy and robotics ambitions depend on, beyond what Tesla expects existing industry capacity to provide. The strategic logic is coherent, since the platform thesis runs on chips Tesla does not yet control.

The capital-allocation question is harder. SpaceX is controlled by Tesla's own CEO, and the commitment exceeds the quarter's entire free cash flow; Musk acknowledged the conflict on the call, noting the deal has to clear independent director reviews and balance two shareholder bases he sits atop. SpaceX has since listed publicly, so the stake is no longer held in a private company, but a related-party outlay of this size warranted more explanation than the single line it received.

That is the Q1 verification. The numbers hold up, but every line carries a caveat: revenue grew against a weak base, margins improved on one-time items, and cash flow was positive partly because Tesla paid its suppliers more slowly, with the rest of the year guided negative. Together they describe a car company spending heavily to become something else, while the stock is priced as if the transformation were already complete.

The next question is whether management can be trusted to deliver what they say comes next.

TIMELINE CREDIBILITY

That trust comes down to timing. The platform businesses are valued on when they arrive, not just whether, and the later they arrive the less they are worth today.

A robotaxi business generating billions in 2028 is worth far less in present value terms if it arrives in 2031 instead. For most companies, a two-year delay is a manageable disappointment. For Tesla, where the largest components of most analyst models are pre-revenue businesses valued on future timelines, a delay does not reduce value linearly. It compresses the discounted present value exponentially.

To illustrate: a business worth $100 billion in 2028, discounted at 12% to today, is worth roughly $80 billion. Push the same business to 2031 and the present value drops to roughly $57 billion, a reduction of nearly a third from timing alone, before any adjustment for the increased probability of failure that a longer timeline implies. That makes timeline credibility the single most important variable in any Tesla analysis. On that variable, Tesla has a long public record.

Musk acknowledged the pattern himself. In 2023, he called himself "the boy who cried FSD" (per Reuters). On the Q1 2026 call, asked when unsupervised FSD would reach consumer cars, he said: "I'm just guessing here, but probably in the fourth quarter."

Musk's pattern is late delivery, not non-delivery. Tesla eventually ships.

Model 3, Model Y, FSD supervised, the Austin robotaxi launch: all happened. All happened years late. The question for investors is not "will it happen?" but "when?", and the honest answer, by Musk's own admission, is that nobody knows.

The sum-of-the-parts models behind the most bullish price targets concentrate the largest share of their value in robotaxi revenue arriving on a defined timeline. If that timeline slips by two to three years, well within the historical range, the present value of that component falls sharply. Billions in assigned value can evaporate from a single delay that, based on Tesla's own track record, would not even be unusual.

Now apply the same logic to Optimus, the more extreme case: multiple analyst models assign it billions in value by 2030, with the first production line guided to start only in late July or August 2026. Everything has to land inside a four-year window: the line starts, the ramp works, the economics prove out, and commercial sales follow. A delay on the scale of Tesla's record, three to eight years, does not shrink the 2030 number; it pushes Optimus revenue past 2030 entirely. If Optimus follows the FSD pattern, even partially, those assumptions do not bend. They break. The effect is not a small reduction in that value but the removal of most of it: the billions analysts assign to Optimus today rest on revenue arriving around 2030, and once the start slips by several years, with the odds of failure rising over the longer wait, little of that value is left once it is discounted back to today.

The capex record is the most current test of that pattern, guidance pitched high and spending coming in low.

Tesla guided above $25 billion for 2026, but Q1 actual was $2.5 billion, an annualised pace of $10 billion. Hitting the full-year guidance requires averaging $7.5 billion per quarter for the remaining three.

There is recent precedent for that pace not holding. Taneja guided on the Q4 2024 call that 2025 capex would be "flat on a year-over-year basis," which against 2024's $11.3 billion implied roughly the same level for the year. That target was revised to above $10 billion by Q1 2025, and to $9 billion by year-end. Actual spend came in at $8.5 billion, below even the final reduced figure. At no point in the past eight quarters has Tesla spent meaningfully above $3.5 billion in a single quarter.

The capex number is simultaneously the strongest signal of transition commitment and the one with the weakest delivery track record. That combination makes it the single cleanest variable to monitor: the clearest statement of intent Tesla has made and the one it has most often missed, so each quarter's actual spend reads directly on whether the transition is being funded as promised.

Three specific production programmes now depend on that capex actually being spent: Cybercab volume production at Giga Texas, Optimus production at Fremont from late July or August 2026, and Megapack 3 at the Houston Megafactory. In 2025, the capex underspend had no visible consequences because no specific production milestone was linked to it. In 2026 that changes. If capex falls short again, it shows up in unit counts and launch timelines that the market can track quarter by quarter.

These launch dates, the spending pace, the city count, each resolve quarter by quarter rather than in a single verdict. The next section frames them as three possible futures for the platform thesis and the evidence that would tell them apart.

THREE POSSIBLE FUTURES

The next twelve months will produce evidence, and that evidence will fall into one of three recognisable states. These are not bull, base, and bear cases: they carry no probabilities and no price targets. They describe what the data could look like a year from now, so that each future can be recognised as the quarterly numbers arrive.

FSD has already begun to show the platform thesis in the numbers; in this future the other ambitions follow it, each turning into an operating business of its own.

FSD user growth remains strong as regulatory approvals expand beyond existing markets. Subscription adoption accelerates, creating a recurring revenue stream that becomes increasingly visible in the financial statements. Energy profit grows in absolute terms as deployment becomes more consistent quarter to quarter, with the Houston Megafactory contributing meaningful incremental capacity.

Robotaxi expands beyond its initial launch markets and demonstrates that demand exists outside tightly controlled pilots, with ride volumes and coverage growing steadily enough to measure. Optimus reaches internal deployment and begins performing economically useful work, with Tesla disclosing unit counts and early cost data. Capex delivery tracks closer to guidance than in 2025, and the factories behind the new product lines open within a quarter of schedule. The platform businesses start appearing in the numbers, not just in presentations. The debate shifts from whether the platform exists to how fast it is scaling.

Tesla continues making progress, but not enough to settle the debate.

FSD adoption grows, although more slowly than recent trends suggest, and the quarterly net addition numbers plateau rather than accelerate. Energy remains attractive but continues exhibiting significant quarter-to-quarter volatility, making it impossible to assign a confident run-rate. Robotaxi expands gradually but remains too small to contribute meaningful revenue, with material revenue staying, in Musk's guidance, perpetually a year or two away.

Optimus advances through development milestones without demonstrating clear commercial economics; management shares impressive demonstrations but no unit cost or internal ROI figures. Capex delivery lands somewhere between the $25 billion guide and the $8.5 billion pattern of 2025, enough to fund some programmes but not enough to hit all timelines. The result is a company that continues moving in the right direction while leaving investors with the same fundamental problem they face today: there is not enough evidence to confidently quantify the value of the platform businesses.

This is the most frustrating outcome for anyone trying to value the stock. Tesla remains too successful to dismiss and too uncertain to value with confidence.

The evidence increasingly fails to support the expectations embedded in the stock price.

FSD adoption slows materially as the early-adopter base becomes saturated. Regulatory expansion takes longer than expected or delivers weaker user growth than anticipated. Energy continues growing, but margins compress faster than the gradual normalisation Taneja guided, heading toward industry norms. Robotaxi remains geographically limited. Optimus remains pre-commercial with no visible path to attractive unit economics.

None of these outcomes imply Tesla becomes a bad company. The vehicle business could remain profitable. Energy could continue growing. Tesla could still be one of the most important industrial companies in the world. But the platform thesis would become increasingly difficult to defend, and it is the platform thesis, not the car business, that the price is built on. The low end of the analyst range, Wells Fargo's $125, is essentially this future priced by a bank: Tesla valued as a car company and little more.

Which future arrives will not be clear from any single quarter; it shows up in whether a specific set of dated commitments are met. The next section lists those forward commitments, each with the date it can first be checked.

MANAGEMENT'S FORWARD COMMITMENTS

Eight commitments, drawn from the Q4 2025 and Q1 2026 calls. Each entry states the commitment, who made it, and the earnings call or data release where it either shows up or does not. When each verification date arrives, you can compare what was said to what was delivered, and adjust your reading of the thesis accordingly.

FSD INTERNATIONAL EXPANSION

Taneja said Tesla was set up well for EU-wide FSD approval "later in Q2," gated by "how the regulators go about it." That can come two ways: a single EU-wide clearance through the European Commission's Technical Committee on Motor Vehicles, or country-by-country recognition of the Dutch approval. The EU-wide route has stalled: the Dutch RDW presented Tesla's case to the committee on 5 May, no vote was taken, and Nordic regulators objected over speeding and icy roads, Sweden most firmly, pushing approval past the Q2 target to the autumn session at the earliest.

Press reporting on the Commission's technical committee indicates a vote is not expected before October at the earliest, at least two quarters behind the "later in Q2" timeline Taneja guided. China has partial approval; Taneja said Tesla is "hoping" for broader approval by Q3. EU-wide and Chinese approval gate the single most important variable for FSD subscription growth: the size of the eligible fleet.

The national route has kept moving in the meantime. Five member states, the Netherlands, Lithuania, Estonia, Belgium, and Denmark, have approved FSD by recognising the Dutch certification. Denmark captures the split, raising concerns at the EU level while granting its own approval, which is why the map is filling in as a patchwork rather than through one bloc-wide vote. The patchwork matters because the domestic runway is short: New Street Research analyst Pierre Ferragu, among the more bullish on the stock, put FSD penetration of the North American installed base at 30 to 35%, heavily concentrated among Hardware 4 owners, and Taneja confirmed the framing. EU-wide or China approval would expand the eligible fleet by millions; the autumn committee vote and the Q2 and Q3 calls are the checkpoints.

UNSUPERVISED FSD TO CUSTOMER VEHICLES

Elluswamy guided a distilled V14 for Hardware 3 vehicles, branded V14 Lite, for end of June. It began rolling out to Hardware 3 early-access customers on 29 June 2026 as firmware 2026.20.5.1, inside the window management had guided, the first such build for the roughly four million Hardware 3 cars held on V12.6 since early 2025. It is a supervised-only build, since Musk conceded on the same call that Hardware 3 cars will never run unsupervised FSD, so V14 Lite closes a feature gap rather than the capability gap. On unsupervised FSD reaching customer cars, Musk said "probably in the fourth quarter," with a gradual geographic rollout, and noted that V14.3 already supports unsupervised operation in the robotaxi fleet, with no large-scale deployment until known software improvements are validated and released. The Q4 2026 call and any mid-quarter software releases are the verification points. Separately, Musk guided FSD V15, a "complete overhaul of the software architecture" running on AI4, for "hopefully by the end of this year, but certainly by early next year."

OPTIMUS PRODUCTION AT FREMONT

Musk guided start of production for "somewhere around the late July, August timeframe," converting the former Model S/X line. He was direct about the uncertainty: "I don't know what the production rate of Optimus will be this year. It is impossible to predict these things." He described the timeline as "insanely fast" by industry standards, four months from dismantling one production line to starting another. The Q3 2026 earnings call is where any unit count, however small, either appears or does not.

CYBERCAB AND SEMI

Production of both has begun, per Musk, with the caveat that "initial production of Cybercab and Semi will be very slow, but then ramping up, and going exponential towards the end of the year and certainly next year." No specific unit targets were given for either. On 30 June 2026 Tesla stated that engineering tests of the first production Cybercab had begun in Austin, a public-road milestone with no unit numbers attached. Engineering testing beginning on the final day of Q2 rules out any Cybercab volume in the Q2 figures. The first unit counts are unlikely before the Q3 call, and the exponential ramp Musk guided points to the Q4 2026 call and into Q1 2027 as the window where meaningful volume either appears or does not. An absence of unit numbers through Q1 2027 would indicate the ramp remains in its early phase.

MEGAPACK 3 AND THE HOUSTON MEGAFACTORY

Musk guided production start for "later this year." This commitment also appeared on the Q4 2025 call, where Taneja cited the launch of Megapack 3 as a driver of expected deployment growth. The Q3 or Q4 2026 call is the verification point. Megapack 3's production start is the supply-side event on this list: new capacity from Houston is what could lift quarterly deployments above the 8.8 to 14.2 GWh range of the past five quarters.

CAPEX ABOVE $25 BILLION

Taneja guided over $25 billion in capital expenditure for 2026, covering six factories, AI infrastructure, and the research semiconductor fab, revised upward from the "in excess of $20 billion" guided on the Q4 2025 call. Q1 delivered $2.5 billion; the remaining quarters must average triple that. Whether the figure moves up, holds, or is revised down on the Q2 call is itself a signal. Taneja also guided negative free cash flow for the remainder of 2026, following Q1's $1.4 billion positive result.

ENERGY DEPLOYMENT AND MARGINS

Taneja confirmed that full-year 2026 energy deployments are still expected to exceed 2025's 46.7 GWh, despite Q1's 8.8 GWh. The remaining three quarters need to average roughly 12.6 GWh, which is within the range the segment has demonstrated (Q4 2025 delivered 14.2 GWh), though it has never sustained that pace for three quarters running. Separately, Taneja guided normalised energy margin compression "from here" due to increasing competition and tariff impacts, with most battery cells procured from China. The Q1 reported margin of 39.5% included more than $250 million in one-time tariff recognitions; the normalised figure sits closer to 29%. The margin trend across Q2 and Q3, after the one-time benefits wash out, will show whether the compression Taneja guided is gradual or steep.

ROBOTAXI GEOGRAPHIC EXPANSION

Musk said Tesla would "expand on it to, like I said, probably a dozen states or more later this year," extending the robotaxi service well beyond its current three unsupervised metros. This is the most-watched variable for the platform thesis, since the eligible market for autonomous ride revenue scales with geography. The Q2 and Q3 calls, and any mid-year city announcements, are the checkpoints. The nearest hard marker was a Q4 2025 deck commitment of seven metros by the end of the first half, namely Dallas, Houston, Phoenix, Miami, Orlando, Tampa, and Las Vegas. The Q1 2026 deck quietly softened it, dropping the dates and relisting the five unlaunched cities as "preparations underway", and as of late June only Dallas and Houston had launched, leaving the target a clear miss.

Those are the open commitments. Here is what I am doing with them.

WHERE ALPHAFORGE STANDS

At $425.30, the stock does not offer a compelling entry.

AlphaForge rates Tesla NEUTRAL for existing shareholders; for new capital, a price below $360 begins to offer a meaningfully better risk-reward. In general, high valuations are associated with low expected future returns: at any given set of future cash flows, a lower entry price produces a higher return, so for a thesis that is multi-year by definition, the entry price is the single biggest factor a buyer actually controls. The $360 level is anchored to Tesla's own trading history, not a valuation output.

Tesla stock ran from $488.54 in December 2024 down to $214.25 in April 2025, a 56% drawdown in four months, then recovered to a high of $498.83 in December 2025 before settling at $425.30 on 2 July 2026. Against that record, a 15.4% move down to $360 is not an exceptional distance; it is the kind of swing Tesla delivers in most years without any change in the underlying thesis. The level is not a prediction that the stock will get there. It is a statement about where the risk-reward shifts meaningfully in favour of new capital.

Even at $360, the stock would trade near 330 times trailing GAAP earnings. The entry zone improves the future return; it does not make Tesla cheap on today's numbers.

Tesla is one of the most compelling long-term holds in the market. Just not at the $425.30 price.

I am not adding here. I am waiting for a pullback into the $360 zone, and I would accumulate more from there.

Long-term bullish on the company; disciplined on the entry.