This article is translated from Morgan Stanley's March 16 research report, "5 Key Takeaways from Fremont Plant Tour, "5 Key Takeaways from Fremont Plant Tour." The ideas in this article are all from Adam Jonas.
Original and Copyright: Adam Jonas@Morgan Stanley Translated by Gravel Villager
On Tuesday, March 15, we visited Tesla's Fremont, California, factory, test-drove the Model 3, Model Y, and Model S Plaid, and attended a meeting with IR. Our main takeaways are as follows.
A Tesla Fremont factory can be said to be "bustling" to say the least.
The plant was never previously designed to produce 450,000 units a year (the peak capacity of the plant was about 300,000 units before Tesla took over from Toyota), which was evident during the tour. Tesla doesn't shy away from the fact that the factory is designed to be inefficient, it has 4 assembly line buildings, one of which is a tent, and the car is assembled inside the tent. In addition, the packaging process for the battery cells is done at the Gigafactory in Nevada, which introduces a range of logistical inefficiencies.
Producing in California also involves higher costs, including: labor procurement, labor costs, and the logistical complexities associated with not being close to supply chains (and European customers). It is worth noting that the space in the factory has also been almost exhausted, and it provides little space for trucks to unload supplies in a reasonable location in the factory. In the factory's office building, there are 10,000 support staff involving human resources, accounting, engineering and other supply chain functions, creating an exciting hustle. Inside the factory, we saw forklifts honking their horns weaving through the traffic, and bustling traffic was busy rushing to produce the 3 and Y delivered in North America.
In Figure 1 below, we can see the difference in the layout of the Fremont and Shanghai factories. Compared to Fremont's vehicle assembly line design, Shanghai has plenty of room for trucks to unload supplies in each relevant area of the factory. Efficiency aside, we consider the Fremont plant to be an important learning "laboratory" for Tesla... Not only is it producing the world's most popular electric car, but it also provides experience for Tesla's future production expansion.
Tesla is showing strong pricing power
The comparable standard Model 3, which was about $36,000 a year ago, starts at about $47,000 today (up more than 30%). The Standard Endurance Model 3 saw the biggest price increase, but other models also rose $5-8 thousand. Tesla is raising prices in a strong position, but delivery wait times are still not decreasing. Depending on the model, the waiting time varies from 2-3 months to more than 6 months. If people buy a more expensive model with an FSD, Tesla starts prioritizing early delivery.
Tesla believes that this dynamic change in supply and demand will continue for at least the next 18 months. Based on the high-value products they offer customers, Tesla was able to use this time to further price. We have underestimated Tesla's pricing power and lead. We don't expect Tesla to need to maintain its market share by cutting prices until 2024/25, when the software can make a bigger contribution to the financial profit and loss statement. Still, we believe that Tesla's mission remains to "invest profits" in lower and lower prices to expand the potential market for entry-level electric vehicles in developed markets and further expand into ultra-low-priced emerging markets such as India.
The impact of raw materials has been perceived, but not as fast as the price increase. Raw materials don't keep Tesla up all night, and the chip problem is actually a bigger concern at the moment
Tesla has long-term contracts (not specifying the exact percentage), and the company is already feeling the pressure on the commodity, but this is already reflected in the results of the fourth quarter and should not have a sudden impact on profit margins (our opinion). In addition, since last September, and especially after further price increases this week, we have entered a cycle of rising prices, which will more positively offset any headwinds in the coming quarters.
Tesla isn't worried about nickel because 50 percent of their models now rely on LFP, and standard endurance models don't need nickel or cobalt. If nickel continues to rise in price, LFP can be raised to 60-70% of the overall new vehicles if necessary. Tesla stressed that the key factor in the near future is the supply of chips.
IV. Production and Profit Margin
Tesla reiterated that its sales growth guidelines, whether berlin and Austin are considered or not, are "easily over 50 percent." As the CFOs mentioned on the conference call, the key hurdle remains chips (not batteries or commodities at this time). Tesla has innovatively tweaked software to reduce reliance on specific chips, eliminating chips such as waist supports, and eliminating radar (FSD now relies entirely on 8 cameras and 12 sensors). Capacity ramp-ups in Berlin and Austin have also made some interesting contributions to the dynamics of profit margins. On the one hand, lower initial fixed cost absorption and low capital utilization will hurt profit margins. On the other hand, more efficient design and reduced reliance on Fremont are a positive factor.
Delivering vehicles in the country where the plant is located also brings cost advantages. Tesla also mentioned that model Y shares 75% of the common parts with Model З, and its manufacturing costs are similar. In addition, Shanghai's capacity, per unit of capital expenditure, is 65% lower than Fremont 's (but China's average selling price is also lower). The move to structural battery packs, as well as the move from 2170 to 4680, will also result in a 50% reduction in battery costs over the long term. Tesla has also reduced the assembly steps in Shanghai from 100 steps (the number of steps that other manufacturers usually have) to only 45 steps, and reduced the welding line from 1,000 robots in the Model 3 to 300 for the Model Y, while still seeking to improve manufacturing efficiency and reduce capital expenditures on the factory.
Tesla has learned from the initial over-automation of the Fremont Model 3, but is still writing its own code for their bots to make more seamless changes when needed and reduce reliance on bot vendors (other vendors won't write their own code). Tesla highlighted that their battery cells are currently sourced from the four leading battery companies in the rankings. In addition, its leading motors, frequency converters and battery management systems are key drivers of battery efficiency.
Tesla IR also mentioned that they believe that if we believe in the predictions of various manufacturers, if the electric vehicle share remains strong in 2025, reaching about 25%, the production capacity of the entire industry and all other related material orders will be in a serious shortage of raw materials supply in 2025. Tesla will be in a key position of dominance to capture meaningful demand for electric vehicles. We note that Tesla has reportedly briefly suspended operations at the Shanghai Gigafactory due to the COVID-19 pandemic. The company has not commented or confirmed at this time.
Five effects of FSD
While not disclosed, investors believe that the conversion rate in the United States (FSD) is about 15%, and in other regions it is in the lower single digits. Once Tesla publicly releases its latest urban FSD, the way it is accounted for will change. Compared to the current 60% FSD revenue recognition immediately, and the 40% revenue deferred method, Tesla will be allowed to recognize all revenue immediately. The impact on 500,000 vehicles in the U.S. (1/3 of projected 2022 deliveries) would be: a 15% conversion rate, based on 40% deferred revenue recognition at the FSD12,000 price, would roughly deliver approximately $360 million in incremental pretax (but note that cash flow remains unchanged) and potentially structurally increase margins by about 40 basis points (bps) in the future.
Tesla emphasizes that autonomous driving is an ARTIFICIAL intelligence problem and hopes to train its neural network over time based on the real-world driving data it accumulates. We note that while our Tesla target price includes ARPU value for the "web services" business and recurring revenue from "Tesla Mobility" (a shared car with human custody), we don't count any specific value for fully autonomous driving because we don't think the technology will be ready for large-scale deployment for many years (ten years or more).
Figure 1
Model 3 production line at the Fremont plant
Model 3 production line at the Shanghai plant
Note: Adam Jonas gives a TSLA price target of $1300.