Tesla Inc.’s stock is downgraded by UBS Group, on concerns that the US electric carmaker’s shares have risen “too much, too soon” on optimism over its artificial intelligence plans.
Increasingly difficult to justify valuation
UBS downgrades TSLA from Neutral to Sell. TSLA is more than just an auto company, and there are some positive developments (e.g. Energy, FSD) that add additional support. This is increasingly important as expectations for the core Auto business deteriorate. TSLA has always had a premium attached to it for other, future, growth initiatives. Properly valuing that optionality is difficult. This premium has widened of late, we believe, on AI enthusiasm. After going through the different businesses we can more substantially value, at current levels, we are still left with a >$500bn “stub” for that future growth. Even if we give that “stub” a 5-year time horizon, that implies a 5-year future value of $1T. And this is just to justify current levels; one would need to see an even larger opportunity to justify a Buy rating. While TSLA is investing heavily in AI and the tech is making progress, investment is costly, pace of improvement may slow and the payoff is long dated. If market enthusiasm for AI diminishes, this may impact TSLA's multiple. Given the lack of visibility and the risk that growth opportunities materialize on a longer time horizon (or not at all), with the stock at 86x NTM P/E, downgrade to Sell.
When ex-auto contribution to price nears the highs, good opportunity to sell
Our valuation attribution analysis shows that the market has (fairly consistently) valued TSLA’s core auto business between $60-$90/share. The “other attribution” has been volatile but past 2 year average is ~$140/share. With the recent rally, it’s now nearly ~$175/share. Fig. 3 shows when other becomes larger contributor to price, stock starts to trend down. O ur SOTP view values auto at ~$57, Energy, which has shown recent strong growth (and higher margin) is worth ~$18. We estimate FSD/robo-taxi could be ~$18 (see UBS Evidence Lab FSD survey inside), but that's only ~$93 of the more easily identifiable value, implying a premium/future option value that is ~61% of today’s price.
Where could we be wrong? 1) TSLA price disconnecting from fundamentals has occurred in the past and can persist for a while. 2) 2Q24 results may be above expectations, causing positive revisions to 25/26F helping to sustain momentum. 3) Investor tepidness to be negative into “robo-taxi” day, and/or there is a true upside surprise at the event. In our view, we believe TSLA is making good technological progress, but the challenge is hard and may face operational challenges (regulatory, etc.). More meaningful than robo-taxi is showing the new vehicle, as that could change 25/26F numbers. But consensus already considers higher units, and we believe the vehicle could pressure auto margins.
Valuation: PT to $ 197 from $147 based on 55x P/E (45x prior)
55x is at the high end of the past 2-year range and is supported by our SOTP analysis.
Pivotal Questions
Q: Can TSLA get to >5mm v ehicle deliveries by 2030?
No. We currently forecast 2030 units at ~3.9mm, 19% below consensus. More mid-term, we are also, on average, ~11% below consensus on 2025-27 deliveries. Our view is informed by more tepid demand for EVs (and demand saturation for the current model lineup) in the US and more competitive markets in Europe and especially China. New and refreshed models are needed for higher levels of demand, but this is already considered in our forecast. Further, in the mid-term (~2027), achieving >3mm units for delivery when max capacity may be ~3mm could be challenging.
Q: Can Energy grow at a ~30% CAGR through 2030?
Potentially. TSLA Energy has grown impressively of late, and importantly, this is currently a high margin business for the company. However, we caution that energy storage deployments can be lumpy. Looking ahead, the opportunity for stationary storage is large, so a ~30% CAGR through 2030 to get to ~$42bn in 2030 revenue is possible. However, more capacity investment would be needed, the timing of which isn't clear, and the law of large numbers may come into play as we move through the decade. We model a ~22% CAGR through 2030E.
UBS VIEW
We are Sell rated on TSLA. TSLA is more than just an auto company, and there are some positive developments in other areas such as Energy and FSD. Further, TSLA has always traded with a premium attached to it for other, future growth initiatives. Properly valuing that optionality is difficult. However, at current levels, we believe that unidentifiable premium is too significant. Given the lack of visibility and the risk that these growth opportunities materialize on a longer time horizon (or don’t materialize at all), we rate the stock Sell.
EVIDENCE
We created a proprietary valuation attribution analysis to better understand TSLA valuation. We took a detailed look at the addressable market for the upcoming lower-cost vehicle. We utilize the UBS Evidence Lab Global EV Survey to support our view on demographics in the US and their willingness/ability to use an EV. We also looked at various UBS Evidence Lab datasets on Tesla's perception in China. We also utilized a UBS Evidence Lab survey to help us better understand the FSD opportunity.
WHAT´S PRICED IN?
Tesla has many factors driving its price action, but at current levels and all else equal, we believe 2030 auto units deliveries would need to be ~6.9mm units vs. UBS forecast of ~3.9mm and consensus of ~4.8mm.
Company Description
Tesla, Inc. manufactures and sells fully electric vehicles and energy generation and storage systems, and related services. The company's mission is to accelerate the world’s transition to sustainable energy. T he company is also working on automated driving technology and Optimus, a robotic humanoid.
More than an auto company, but how much more?
TSLA is often labeled “not an auto company” or “more than an auto company”. We are more inclined to agree with the latter considering the vast majority of today’s revenue/profits are related to the manufacturing and selling of vehicles. This makes valuing TSLA stock challenging at times. We are not here to say that TSLA should be valued as an auto company as there are other initiatives. Some of these initiatives may have payoffs way out in the future. TSLA stock price been sensitive to future businesses, opportunities and TAMs (i.e. the potential of what TSLA could become). Some premium for this potential is fair, but we do believe investors should be cognizant of how much the auto business and other more readily valuable businesses contributes to the stock price to help ground the valuation and select entry and exit points.
UBS' TSLA valuation attribution analysis
Below, we took a look at the historical contribution of TSLA’s automotive business to TSLA’s total market value. We specifically focus on auto only for this analysis as for a large period of time, this has been, and remains, the company's primary driver. A uto also has the richest data for historical consensus expectations. For this analysis, we took consensus 2030 auto delivery expectations (at that time) and consistently applied a $40k ASP, a normalized Auto EBIT margin of 15% and 20x forward P/E multiple. Yes, some will quibble with our assumptions, particularly the multiple considering legacy OEMs trade at single-digit P/Es and Toyota, as a best in class OEM, historically may trade closer to~10x. However, we’d a) argue that TSLA's auto business still has more growth potential ahead vs. a more mature Toyota and b) we believe we are being somewhat generous to auto for this analysis. We then discount that nominal auto value at the cost of equity (at that time) to that period. We then compare the implied auto contribution to the total market capitalization of TSLA at that time.
A few things from this analysis surprised us:
The market value attributable to auto was more consistent than we expected. It has generally been between $200-$300bn since 2021 or ~$60-$90 in per share terms. This surprised us because consensus 2030 deliveries used to be >7mm and now stand >30% lower at ~4.8mm units. Mitigating the impact of this negative revision is the discounting (i.e. we are now closer to 2030). Bears may argue even current 2030 expectations are too high, which may be warranted (UBS is at ~3.9mm). However, we felt using consensus expectations to figure out market expectations was more appropriate.
How much of TSLA's valuation is attributable to “other” initiatives? Over the past 4 years, on average only ~30% of TSLA’s stock price is attributable to core auto, again using our assumptions from above (less if one were to assume a lower margin or multiple). Over the past two years, this has been closer to ~36%. At TSLA’s current price, only 28% of TSLA’s current value is attributable to core auto, an 8pp swing since the end of June and at the lows over the past 2 years.
While valuation disconnections from more tangible fundamentals have occurred in TSLA’s history, and it can persist for extended periods of times, we believe that when these hopes get elevated, it is not a good time to initiate or add to positions. Rather, it’s a time to sell. Over the past two years, when the percent of the stock price that our analysis shows is attributable to auto dips below the 30% average, the stock has entered downward trend channels (see 6/22 and 6/23).
What about TSLA’s other endeavors?
With our attribution analysis showing less than 30% of the current stock price is attributable to core auto, we asked ourselves, did something change? While the recent 2Q24 auto deliveries of ~444k was better than expectations, it was still down -5% y/y and was likely supported by some promotional activity. Further, in our view, quarterly deliveries coming in 20-25k (~5%) better doesn’t warrant a 15% rally since the delivery print (and a ~22% rally over the past 10 days). So, is there reason to be more optimistic on some of TSLA’s other initiatives?
Tesla Energy
One area that the market may have gotten more excited about, that also came out with the deliveries release, is Tesla Energy, which reported 9.4 GWh of energy storage product deployment in the quarter (+157% y/y and +132% q/q). At their annual meeting, Elon Musk also indicated the company was tracking towards 200-300% y/y growth in energy storage. This is important since Energy is currently higher gross margin than Auto (Energy was 24.6% in 1Q24), so this could help stem negative revisions from the pricing/promotion actions taken to stimulate auto demand. While we are more positive on Energy and raise our Energy sales forecasts, we caution that Energy sales tend to be lumpy.
Tesla currently has their Megafactory in Lathrop, California, which has an annual capacity to build 10,000 Megapacks or ~40 GWh of storage. During 1Q24 earnings, TSLA announced that they had commissioned their second general assembly line as they look to ramp Lathrop to full capacity. TSLA is also opening their Megafactory in Shanghai which will add an additional 40 GWh of energy storage capacity. We expect this facility to be operational in 1H25. We now forecast TSLA will deploy ~27 GWh of storage this year, and ramp to ~37 GWh next year.
AI
Tesla is investing heavily in AI. Earlier this year, Musk announced that TSLA had installed and commissioned 35k Nvidia H100 GPUs and are aiming for 85k GPUs by the end of the year. Musk has also talked about spending ~$10bn on combined training and inference AI this year, the latter being primarily in the car. Tesla is also is developing their own silicon for their Dojo supercomputer, though in January 2024, Musk stated Dojo was “a long shot worth taking because the payoff is potentially very high. But it's not something that is a high probability."
The fruits of AI investment at Tesla would be related to Full Self Driving (FSD), which is their current advanced driving assistant, their robo-taxi initiative (planned robo-taxi day, now reportedly postponed to October), and Optimus, their humanoid robot initiative.
We believe TSLA stock price has gotten caught up in the AI trade/phenomenon. While we don’t doubt Tesla is making very good progress on such initiatives, aside from FSD (we were impressed by recent improvement), the other initiatives are purely R&D. Large TAMs for sure, and we have more confidence in Tesla being able to manifest the power of AI in the physical world than we do for other companies. But, these are complex issues, that will take a long time to play out and success is not a certainty. As AI enthusiasm fades, we believe this can impact TSLA's multiple.
Full Self Driving (FSD)
Of large debate is Tesla’s ability to charge, on a monthly recurring basis, for their advanced driver assistance system: FSD (supervised). Earlier this year, Tesla started rolling out the new version of FSD based on an end-to-end AI learning system. Tesla offered a 1-month free trial to all US Tesla vehicles with the proper hardware. They then lowered the monthly subscription cost to $99/month from $199/month.
We conducted a survey on 306 Tesla owners, ran over June 10-28th, to gauge customer reception and perception to FSD. Key points:
A caveat, we believe our survey skews high to consumers that currently have (pay for) FSD vs. the reality of the marketplace. T his in part may be driven by ~53% of our respondents indicated that a Model S/X was their primary Tesla vs. S/X making up only ~6% of Tesla's 2023 US sales. S till, we are able to gather some insights from looking only at 3/Y owners and, in particular, respondents who don't currently own/subscribe to FSD to see what future "take rates" might be (though admittedly this is a smaller sample size).
Overall, we would say consumers are satisfied with FSD functionality (62% very satisfied) and feel safe using it (64% completely safe).
Of those that are not currently paying for FSD (so either free trial is over or still on free trial) 17% indicated they were likely to pay for FSD over the next 6 months, while 33% said somewhat likely.
However, our respondents did over-index to Model S/X vs. 3/Y when in reality, 3/Y is a significantly larger pool (and given the price point of the vehicle, points to a different willingness to pay, in our view). So, when we look only at respondents who said their main vehicle was a 3 or Y, very likely to pay for FSD over next 6 months dropped to 9% and Net Likely vs. Unlikely is only +3% (Figure 15).
At the current $99/month level, 43% of Tesla customers surveyed (that indicated they are likely to continue to pay or will be likely to pay for FSD) indicated they would be willing to pay around this price ($76-$100/month). This dropped to 35% when only looking at Model 3/Y users. To us, this suggests that significantly higher adoption levels may need further price reductions.
The willingness (and ability) to pay for FSD matters to us as we think about future, low cost Tesla models. These new lower cost models will have a less expensive monthly vehicle payment. But, that also means that adding an additional $99/month to that payment for FSD will be a significant increase.
In China, while Tesla is making a push to be able to make FSD available, we have doubts over Tesla's ability to charge for FSD given many competitors already offer an autonomous solution without charging a monthly fee (or some even an upfront fee). Rather, we view FSD in China as almost a necessity to remain competitive with domestic Chinese offerings that have similar functionality.
Robo-taxi
We believe robo-taxis are a difficult technological endeavor and the business model also may face regulatory hurdles as well as questions around consumer adoption. Thus, we believe a meaningful r obo-taxi operation in the US is further out (not this decade). Tesla reportedly/recently pushed the planned 8/8 robo-taxi day to October. A subtle reminder that Tesla has had difficulty meeting timelines.
On the technology side, TSLA's move to an end-to-end (GenAI) approach has allowed for meaningful improvement in the FSD product. But the FSD product is a very good advanced driving assistant system (L2+). To move to a robo-taxi solution, the march of 9s must continue to be able to operate in all conditions within the operational design domain (ODD). We believe the end-to-end approach, on it's own, will eventually start to show slowing improvement and diminishing returns. Further, the cost for this improvement can be high. More, unique data is needed to continue the system improvement. TSLA does have access to a lot of data from their fleet, and they employ simulation. But, we believe that, similar to what is seen in GenAI with large language models (LLMs), progress can slow, and each additional improvement to the model can become increasingly more expensive to take (obtaining that new data to get the improvement and the compute to process it).
The other challenge with an end-to-end approach could be a regulatory one. There will inevitably be accidents and with an end-to-end approach without guard rails, it is difficult to understand why the system made a decision it did, thus difficult to correct for the problem (as rare as the specific event might be). We recently held an expert call on Gen AI for autonomous driving that highlighted the issue regulators may have with the lack of traceability and transparency that comes with an end-to-end approach that TSLA currently utilizes. Further, in certain geographies,TSLA still needs to test their robo-taxi with a driver before being permitted to test autonomously and then to offer driverless rides (see Waymo and Cruise).
Optimus
While we are excited about the progress Tesla is making on their humanoid robot, we believe this opportunity is much further out. At their annual shareholder meeting, Elon Musk helped frame the long-term opportunity. H e mentioned there could be, one day, 10bn humanoid robots, so if TSLA gets 10% share, that is 100mm Optimus units a year, and if sold for $20k, suggests a $2T revenue opportunity, $1T profit opportunity and a $20T market value opportunity (20x). However, we have difficulty underwriting significant (or specific) value to this opportunity today given the many uncertain variables, the probability of success and the unknown time frame.
Sum-of-the-parts
Below we take a look at a sum-of-the-parts (SOTP) for TSLA to support our P/E based valuation. We also show an upside and downside scenario.
Auto. We forecast 2030 units at 3.9mm units and ~$146bn in auto revenue. Applying a 15% “normalized” EBIT margin, applying a 20x P/E multiple and discounting that back to a year from now (mid-2025E) at ~13% cost of equity gives us $57/share value. 20x is above traditional OEMs in the mid-single-digits P/E and Toyota at ~10x but warranted given Tesla would still have more room for growth and the "normalized" Auto EBIT margins for TSLA (especially relative to legacy OEMs) could be higher.
Energy. We forecast 113 GWh storage deployment and ~$28bn in Energy revenue (22% 2024-2030E CAGR). Applying a 20% “normalized” EBIT margin, applying a near-market multiple of 25x P/E multiple and discounting that back to a year from now (mid-2025E) gives us $18/share value.
FSD. We forecast ~$1.5bn in 2030 revenue, which equates to ~$0.33 in EPS, and apply a software like 50x multiple, which discounted back to a year from now (mid-2025E) gives us a $9/share value.
Robo-taxi. According to CNBC, Waymo, which actually gives driverless rides/operates a robo-taxi service today, latest valuation (2020) was $30bn. GM’s Cruise similarly once had a $30bn valuation, though we believe a current mark could be lower. In our base case, we apply the latest Waymo valuation of $30bn or $9/share. Our upside case more than triples the Waymo valuation to ~$100bn (we'd also note that Uber has a ~$144bn market value). Our downside case assumes no value. We note that Baidu has recently been in the news as their Apollo Go robotaxi autonomous ride-hailing platform has been gaining attention and is piloting in areas such as Wuhan. However, we note that the entire Baidu market cap is ~$35bn. T he stock has rallied of late (~14%) on some of this robotaxi optimism, but that rally only contributed ~$4bn to the market cap.
The TSLA Premium (AI, Optimus, other initiatives). These are difficult to value given they are, at best, R&D today. Here, we are relying on the "collective wisdom of markets". Going back to our earlier TSLA valuation attribution analysis, on average over the past 2-years, the market has assigned ~$490bn of value (~$141/share) to ex-auto initiatives. Between Energy, FSD and robo-taxi, we only identified $36/share of that, leaving the TSLA premium at $105/share. Note, that assuming the 13% cost of equity, the 5-year future value of that premium is $672bn and the 10-year future value is $1.2T. Given that many of these initiatives are likely to take time to materialize, we believe this seems more than fair. In our upside case, we use the more recent (high) level from our valuation attribution analysis while our downside case assumes no premium, which is closer to the minimum we've seen vs identifiable valuations.
VALUATION
We value TSLA shares using a 55x P/E multiple and apply that to our current 3Q25-2Q26 TSLA EPS forecast. This yields our new $197 price target. 55x (45x prior) is at the high-end of the NTM P/E valuation range over the past 2-years but is supported by our SOTP work above.
Pivotal Questions
Q: Can TSLA get to >5mm vehicle deliveries by 2030?
UBS VIEW
No. We currently forecast 2030 units at ~3.9mm, 19% below consensus. More mid-term, we are also, on average, ~11% below consensus on 2025-27 deliveries. Our view is informed by more tepid demand for EVs (and demand saturation for current model lineup) in the US and more competitive markets in Europe and especially China. New and refreshed models are needed for higher levels of demand, but this is already considered in our forecast. Further, in the mid-term (~2027), achieving >3mm units for delivery when max capacity may be ~3mm could be challenging.
EVIDENCE
We took a detailed look at the addressable market for the upcoming lower-cost vehicle. We utilize the UBS Evidence Lab Global EV Survey to support our view on demographics in the US and their willingness/ability to use an EV. We also looked at various UBS Evidence Lab datasets on Tesla's perception in China.
WHAT´S PRICED IN?
At current levels, all else equal, we believe 2030 deliveries would need to be ~6.9mm units vs. UBS forecast of ~3.9mm and consensus of ~4.8mm.
How impactful will the 'Model 2.5' be?
Tesla’s eventful robo-taxi day may get most of the fanfare, focusing on AI and a robo-taxi service. However, in our view, a robo-taxi and other AI initiatives are still much further out. That is why we are focused on the “new” vehicle that Tesla plans to introduce, which we dub the 'Model 2.5'. To us, this new vehicle could be the only data point to come out of the event (now expected in Oct) that can materially get investors to revise 2025/26 estimates.
Recall, Tesla scrapped the Model 2, which was expected to use a brand new unboxed manufacturing process to significantly lower the cost to produce the vehicle and offer a “$25k EV”. In their 1Q24 earnings release, Tesla indicated they plan to launch new vehicles, including more affordable models that utilize aspects of that next generation platform as well as aspects of their existing platform and will be able to be produced on the same manufacturing lines as Tesla’s current vehicle line-up.
We expect the Model 2.5 to look meaningfully different from the Model 3/Y silhouette as to differentiate the vehicle from potential Model 3/Y buyers. We also expect the starting price to be materially lower, potentially down to $25k (though we’d expect few variants to actually transact at that price).
What is the opportunity for the Model 2.5? In the US, the small standard car category (featuring the likes of the Toyota Corolla and Honda Civic) has accounted for ~7% of total sales over the past 5 years, or ~1mm units, on average. The Toyota Corolla is the current segment leader at ~23% share.
When we look at the Model 3/Y and their segments, small premium II car and small premium II SUV, respectively, these Tesla models have been able to garner 40-50% share of those segments over time. If Tesla were able to garner similar share in the small standard car category, that would yield a ~450k annual sales opportunity for the Model 2.5 in the US. However, appealing to a lower tier consumer may yield additional challenges for that consumer to go electric. For instance, access to at home charging may be more limited for lower income buyers.
In Europe and China, smaller entry vehicles may appeal to those consumers relative to the US. I n 2023, European sales of small entry level cars was ~1.1mm or ~6% of sales. However, Eastern & Central Europe made up about ~600k of that figure. W e believe these regions are likely to lag Western Europe on electrification. The Model 2.5 is also likely to face more competition in Europe as legacy OEMs look to electrify their small, affordable vehicle offerings. Further, smaller EVs from China are likely to enter the European market.
In 2023, the Chinese affordable entry car market was ~1.7mm units, or ~6% of sales. However, t he Model 2.5 is likely to face intense competition as well, as even today, competitors for a low cost BEVs in China exist.
To further show the competitive challenge Tesla has in China, we looked to UBS Evidence Lab and China 360, who recently published a brand perception survey which surveyed 1,800 Chinese customers. In regards to Autos, the survey highlighted that the top 5 factors of importance for Chinese customers are safety, trustworthiness, high quality, value for money and versatility. T he survey highlighted that Chinese customers view Tesla's brand image as one more of p remiumness, trendiness and innovation. On the top 5 factors, Tesla only screened well on high quality.
Guided by capacity
While the market opportunity may be there globally, we must always keep in mind that production and deliveries must be guided by capacity. We currently forecast TSLA 2024 production at 1.78mm and deliveries at 1.69mm. As of 1Q24, Tesla’s stated installed capacity is 2.35mm units. Tesla has indicated that the move to utilize the same manufacturing lines as their current vehicle lineup would enable them to fully utilize their expected maximum capacity of close to 3mm units. We believe this means utilizing more of the space at their Austin and Berlin facilities. We note Tesla recently received approval to expand capacity in Berlin to increase manufacturing capacity to ~1mm units. Europe imported ~170k TSLA vehicles from China last year, but in light of potential tariffs on EVs made in China, this could make manufacturing more vehicles in Europe more attractive. TSLA produced ~200k vehicles in Germany last year according to S&P Mobility. We believe Tesla may eventually explore expanded capacity at Shanghai, but for now, we believe that facility is operating at max capacity.
Tesla indicated that they will utilize that ~3mm of capacity before investing in new manufacturing lines. We would assume this means new manufacturing facilities as well. Consensus is currently looking for ~3.1mm unit deliveries in 2027, which likely means production has to be higher than that. Further, operating at 100% utilization is very difficult. So, we see risk to current unit expectations over the coming years.
New models important as current demand stagnates
The reason why the new model is so critical to TSLA growth is that demand looks more challenging in Tesla’s major regions of the US, China and Europe.
US - Is Tesla demand saturated?
In the US, YTD BEV penetration is 7.7%, up from 7.6% in 2023. TSLA's current share of the BEV market in the US is ~51%, which has come down from ~56% in 2023.
In the US, even recent levels of demand for TSLA vehicles seems to be driven by pricing actions. This is not just price cuts but now also financing actions. In May, TSLA lowered the APR for the Model Y to 0.99% (from a market rate of closer to 6.39%) and lowered the monthly lease payment for the Model 3 to $299 from $329.
Europe - demand lower y/y
In Europe, BEV penetration YTD is 12.1% (fairly flat y/y), and TSLA has ~17% share, which is down from ~19% last year while penetration has remained relatively flat. This is driven primarily by falling demand in Germany, France and Norway. We believe BEV penetration has stalled, in part because there is no regulatory need for OEMs to push higher BEV mix this year. That should change next year, which could mean we see BEV penetration in Europe increase, but TSLA share decrease further.
China - A very competitive market
In China, YTD NEV retail penetration is ~40%, TSLA's share of the NEV market being ~7%, with TSLA losing 100bps of share vs 2023. China remains a very competitive market with a multitude of new entrants and new models being brought to production a lot faster than in other parts of the world.
Can the Model 2.5 deliver on margin expectations?
This remains a key and open question, but the math seems difficult. For instance, if we assume that a Model Y sells for ~$45k and has 20% gross margins, then the COGS/unit is ~$36k. The Model 2 was expected to start at a $25k price point but was also supposed to use an unboxed manufacturing concept that was expected to reduce COGS by ~50% vs. Model 3/Y or call it ~$20k of COGS for the original Model 2 plan. If we assume the plan was for the Model 2 to have had similar gross margins to the Model Y, Model 2 COGS would need to be ~$20k. Part of this reduction is from a smaller battery, reduced content (i.e. no glass roof) as well as better integration and controller reduction, etc. However, we believe roughly 1/3rd was due to the new unboxed manufacturing process. This means Tesla would need to find an additional $5k in cost to get the targeted gross margins of the original Model 2 plan. If we assume Tesla were able to get half of the savings from the originally planned unboxed manufacturing benefit, then, that would imply a gross margin of 10%. Of course, pricing could be higher (and effective pricing will likely be higher from differing trims). However, the higher the price of the Model 2.5 goes and the closer it gets to Model 3/Y pricing, the differentiation and cannibalization problem gets larger.
Risk to the current share price is currently at 1:1.8 to the downside
UPSIDE ($338): Our upside case uses a 79x PE on our 3Q25-2Q26 EPS estimate of $4.29. This is supported by our SOTP analysis that attributes $106 to Auto, $30 to Energy, $13 to FSD, $29 to Robo-taxi and $160 as a Tesla premium for all other initiatives. This implies a value per share of $338.
BASE ($197): Our valuation uses 55x PE on our 3Q25-2Q26 EPS estimate of $3.57. This is supported by our SOTP analysis that attributes $57 to Auto, $18 to Energy, $9 to FSD, $9 to Robo-taxi and a $105 TSLA premium for other initiatives in line with market implied average over past 2-years. This gets us to our $197 PT.
DOWNSIDE ($67): Our downside case uses 23x PE on our 3Q25-2Q26 EPS estimate of $2.86. This is supported by our SOTP analysis that attributes $48 to Auto, $14 to Energy, $5 to FSD. In our downside we do not attribute value to Robo-taxi or include a TSLA premium in our valuation, which is closer to the minimum we've seen vs identifiable value. This implies a value per share of $67.