Dolphin Research
2023.10.12 12:40

FSD Smart Driving: Unable to Support Tesla's Next Valuation Miracle

portai
I'm PortAI, I can summarize articles.

After the release of two research articles on Tesla's car-making issues, "Tesla: How Far is Musk's Trillion-Dollar Empire Dream?" and "The Lion King Meets the Wolf Pack: Can Tesla Protect Its Home?", some people expressed their protest: "If Tesla is just a new energy vehicle company, it may not even be worth $100 per share. If Tesla is just an electric car company, it's done for!"

The implication is that buying Tesla is not just buying its tangible car-making business at present, but also paying for the "potential stock" of imagination for the future. According to the plans of dreamer Musk, Tesla may eventually become an Iron Man in the green energy transformation: green energy generation, electric vehicle propulsion, cars as robots, robot-driven cars, intelligent car-sharing networks... creating a fully intelligent Tesla "mobile empire".

If you firmly believe in Musk's grand vision and buy into Tesla with this long-term thinking, then hold the stock for more than 10 years and don't pay attention to short-term fluctuations. Otherwise, it is simply using the rogue logic of "using long-term thinking to guide short-term trading".

Moreover, through the in-depth analysis of Tesla's car-making business in the previous two articles, it can be seen that apart from the macro-level impact of the US economy's resistance to high interest rates, Tesla does not have any fundamental advantages in car-making: a) 2023 is a small cycle for car-making, with only old cars being redesigned; b) competition is intensifying, leading to price reductions and the actual sales target being set at the lower end of the initial guidance - 1.8 million; c) the gross profit margin of car-making is landing, and it is basically following or even lower than the lower limit of the initial guidance.

But why has Tesla risen from 100 yuan/share at the beginning of the year to 260 yuan/share now? Is this kind of rise reasonable? Does Tesla have a reasonable valuation? In this article, Dolphin Research will discuss the true driving force of valuation in the process of this year's rise - "cognitive reassessment" - through the analysis of Tesla's valuation second curve: the product layout and business logic of its autonomous driving business.

I. Tesla: Two different levels of rise this year

To briefly review Tesla's rise since entering 2023, Dolphin Research divides it into two stages:

Valuation recovery in car-making: At the end of last year, market competition shifted from the supply side to the demand side, and funds had a "panic" perception of Tesla, falling below the conservative valuation of its car-making business. At that time, Dolphin Research clearly stated that the opportunity had come for Tesla fans (see the commentary "Tesla's Story Reshaped: The Moment of Testing Faith Has Arrived!"). After Tesla's stock price recovered to over $200, due to the pressure on the US macroeconomic expectations in the first half of the year, Tesla turned downward again. At this stage, the market valuation of Tesla is still mainly supported by changes in expectations for its car manufacturing business.

AI story boosts valuation: In May and June, a series of catalytic events occurred: a) The market began to realize that the US economy was not in recession, but rather resilient, and global car sales were not as pessimistic as imagined; b) ChatGPT+ NVIDIA sparked an AI boom; c) Tesla's charging stations in the US became a basic standard; d) Tesla's entire US lineup became eligible for a $7,500 subsidy. These events drove Tesla's stock price to nearly $300 per share.

During this process, it is worth noting that in late July, amid the liquidity contraction in the US stock market, Tesla did indeed fall back to over $200, which supported its valuation in the first half of the year. However, this time it rebounded again. From the latest stock price trend, it is evident that the "storytelling" businesses such as AI and charging stations, which have not been reflected in the financial reports, have raised Tesla's valuation bottom line. The market is gradually pricing Tesla's "AI story" into its valuation.

Ⅱ. Tesla forges an "iron wall" of AI autonomous driving.

The current stock price has already partially reflected Tesla's autonomous driving story. So, what is Tesla's business layout in this field, and how should we understand the commercial value and moat of this business?

Let's start with the conclusion: As an internet-oriented business model, Tesla has indeed achieved a high degree of commercial closed-loop in autonomous driving. The main uncertainties lie in the size of its commercial value and how much valuation it can project at present.

From the simple "one-time sale" of cars to the monetization of intelligent stock cars through software and traffic taxes, Tesla has almost single-handedly opened up the imagination space for monetizing car sales. However, the effectiveness of the business model that monetizes traffic and payments after acquiring customers through intelligent hardware is quite different for Apple and Xiaomi.

Taking into account the successful model of software and hardware integration in the smartphone business, a clear characteristic is: constructing hardware barriers around "intelligent demands" to achieve large-scale shipments of hardware.

In comparison to smartphones, where intelligence is the core barrier, Apple has achieved self-research and development in hardware chips, iOS systems, and more.

In terms of providing intelligent products in the automotive industry, excluding the heavily homogenized intelligent cockpit, the key lies in Tesla's layout in autonomous driving: data, computing power centers, chips, algorithms, all of which are self-owned, self-built, and self-developed, truly achieving full-stack self-research. Then let's take a look at the layout of the two product lines, hardware and software, that support Tesla's autonomous driving products, as well as the differences between Tesla and its competitors:

1) Low upfront hardware cost: Because Tesla relies mainly on cameras and a few low-cost sensors for its pure vision-based approach, the upfront hardware cost is generally very low. In fact, due to its advanced algorithms, even the resolution requirements for the cameras are not particularly high.

2) Smarter and faster software iteration: The level of "intelligence" required for the algorithms is very high, and the software iteration speed is fast. Especially after the self-developed FSD chip matures and the FSD Beta V9 version rewrites the perception algorithm, Tesla has been leading the way in the advancement of city NOA (Navigate on Autopilot) for more than two years.

3) High backend computing power and R&D investment: Because Tesla lacks sensors such as LiDAR and relies solely on 2D images and videos generated by cameras to convert them into 3D objects' distance, height, and speed, the algorithm's intelligence requirements are relatively high. In addition, Tesla has a large number of vehicle owners, resulting in a massive amount of data.

In terms of data cleaning, feeding, and model training, Tesla's autonomous driving path requires higher computing power. Therefore, compared to its competitors, Tesla is the only company that has built its own supercomputing center and has developed the D1 supercomputing chip and the DOJO supercomputer. In contrast, most of its competitors rely on cloud computing service providers for computing power.

On the other hand, domestic competitors in China can directly construct the 3D space of objects through the embedded LiDAR in the vehicle, which requires a lower level of algorithm intelligence (and thus lower demand for cloud-based supercomputing).

Through the above comparisons and analysis, we can see that Tesla's autonomous driving system, in order to achieve a closed-loop integration of hardware and software, actually has some similarities to cloud computing in terms of entry barriers, with a higher intensity of "sunk costs" in terms of capital and technology investment. Specifically:

a) Tesla's autonomous driving path relies heavily on advanced algorithms, and optimizing a massive number of algorithms requires a large amount of data feeding and training, resulting in a huge demand for supercomputing power and high sunk fixed costs. On the other hand, companies like Xiaopeng, Li Auto, NIO, and Huawei, which adopt a multi-sensor fusion mode with LiDAR embedded in the front-end of the vehicle, have a more variable cost structure.

b) In terms of economies of scale, one is a fixed cost, driven by commercialization and increased penetration rate, which will generate obvious internet effects and dilute fixed costs. The algorithm optimization is fast and advanced, making it easy to establish a true ecological closed loop through technological leadership, self-research, and self-construction.

On the other hand, the variable costs of competitors, even with more users, still require at least one lidar sensor per car, resulting in less dilution effect. Of course, if the lidar sensor path is established and installed in large quantities, it is only a scale advantage in the manufacturing industry.

Ⅲ. Can "impenetrable defenses" become "golden opportunities"?

From the analysis above, unlike Tesla's relatively weak competitive barriers in the "electrification" of automobiles, Tesla's ecological closed loop in the "smart driving" layout is very obvious.

But does the intelligent ecological closed loop necessarily translate into the core barrier of business? If the answer is affirmative, it means that you believe that in the future, when users buy cars, intelligent functions are the top priority demand.

Tesla's integrated software and hardware system is actually very similar to Apple's integrated software and hardware ecosystem in smartphones. For smartphones, we know that the evolution from feature phones to smartphones is based on the core barrier of "intelligent" experience.

For example, smartphones should not overheat due to intelligent functions, and should not generate a large amount of data junk (as seen in bloated Android phones), which affects the lifespan of the phone. Apple builds a software and hardware ecosystem around the intelligent experience, allowing it to not only enjoy the high premium of hardware, but also make a lot of money from software and traffic.

However, for cars, at present, most people in China still buy electric cars based on their energy-saving and fuel-efficient characteristics compared to fuel-powered cars. Intelligent features are more of an added bonus, rather than a "rigid demand" like buying a data plan or installing WeChat on a smartphone. And this is a key issue when estimating the product penetration rate and pricing ability in the future for intelligent layout monetization.

Perhaps many years later, the first rigid demand for buying a car will truly become the "smart driving" demand. This type of product architecture will clearly demonstrate the scale effect of the internet: high accumulation of intelligent mileage - precise algorithms - excellent smart driving experience - more people buying Tesla cars and using FSD - higher utilization of computing power capacity, forming a self-circulating software and hardware integrated barrier.

From the current perspective, we have not seen cars replicate the functional generalization of smartphones in the intelligent era. The concept of intelligent cars becoming the "fourth living space" (essentially increasing the internet logic of car user time) is still in the distant imagination stage. Also, due to this uncertainty and the penetration risk in different countries, Dolphin Research will consider the corresponding discount when constructing its business value judgment. Next, in order to understand its commercial value, Dolphin Research will first take a look at where Tesla's current commercialization of FSD stands, and share Dolphin Research's understanding of its business model and future direction.

1) Where is Tesla's current commercialization of FSD?

From the current charging form, the commercialization of intelligent vehicles, Tesla's main revenue streams are as shown in the figure below, with a very small proportion coming from connected cars. Overall, it is a typical internet software/SaaS business.

a) Smart driving business revenue, a drop in the bucket

This payment model will bring a large amount of deferred revenue, and the actual revenue level is reflected by the revenue, not the revenue recognized after deferral. Based on the calculation of the increase in deferred revenue and the amount recognized in Tesla's 2022 annual report, the total revenue from FSD, connected cars, free supercharging, and OTA related to Tesla in 2022 is $1.3 billion, which is still lower than the revenue from Tesla's sale of carbon credits ($1.8 billion in 2022).

According to the information shared by Tesla, the investment in the single smart driving computing power element, the DOJO project, will reach $1 billion by 2024. It is clear that from a financial perspective, Tesla's smart driving is currently in a state of bleeding, rather than the appearance of a gross profit margin of over 80%.

b) Will the inflection point come with the increase in smart driving mileage?

According to data from Twitter influencer Troy Teslik up to the third quarter of 2022, the FSD order rate did not increase all the way, but instead decreased as the number of cars sold increased:

Starting from the second quarter of 2018, it rose rapidly, with the global order rate increasing from 9.7% to a peak of 45.7% in the second quarter of 2019. However, by the third quarter of 2022, the global order rate had rapidly declined to only 7.4%.

A rough look shows that in addition to the previously mentioned price increase for FSD, there is a huge difference in regional penetration rates, with high-priced cars having high penetration rates and low-priced cars having low penetration rates.

In terms of regions: After the second quarter of 2019, North America continued to lead, followed closely by Europe. The Asia-Pacific region (mainly China) had an order rate of less than 1% due to legal and regulatory restrictions starting in 2021, but the delivery volume in China accounted for a high proportion of 35%. Based on the third quarter of 2022 order data, the FSD order rate in North America was 14.3%, Europe was 8.8%, and the Asia-Pacific region was only 0.4%.

In terms of vehicle models, the penetration rate of high-end models Model S/X with FSD is the highest, reaching 40%-50% in 2022, far higher than the 3%-4% for Model 3 and around 5%-7% for Model Y during the same period. However, the proportion of Model S/X in the delivery structure has decreased from 18.6% in the second quarter of 2019 to 5.4% in the third quarter of 2022. The decline in the delivery proportion of high-end models with high FSD adoption rates has led to a decline in overall penetration rate.

However, it seems that there has been a turning point in 2023: Tesla's cumulative autonomous driving mileage showed a non-linear increase at the beginning of 2023. The main reason for this factor is that in November 2022, Tesla pushed the FSD Beta V11 version to all FSD users in North America, which unified the highway and urban NOA functions and greatly expanded the usage scenarios.

Moreover, with the integration of City NOA and Highway NOA, the user penetration rate has also increased. From the accumulation of users, after the release of the V11 version, the number of testers reached 285,000 by the end of December 2022, and the number of FSD Beta testers reached 500,000 in May 2023 (with a penetration rate of approximately 25%+ relative to the existing vehicle stock). During this period, Tesla's sales in North America were only around 100,000+, so there is indeed a certain upward trend. Therefore, the market is starting to imagine the "turning point in penetration rate and the outbreak of intelligence".

  1. How should we view the long-term penetration rate and pricing prospects of FSD?

There seems to be a turning point for FSD in North America, but when estimating the commercial value of Tesla's FSD, we also need to consider the China region and the European region, where the sales distribution weight is relatively large. In China, which is different from the United States, there are two very obvious issues:

a. Tesla's domestic pricing is too high.

After the price reduction in the United States, Tesla's FSD buyout price is equivalent to 88,000 RMB, while the domestic price is 64,000 RMB, which only includes software usage fees. However, Dolphin Research estimates that Xpeng's actual software buyout fee for advanced intelligent driving services may be only 5,000 RMB. Xpeng's pricing for advanced intelligent driving services includes both software and hardware. In terms of version price difference, Xpeng's intelligent driving service is priced at 20,000 RMB, which includes additional hardware (two additional LiDARs + one additional Orin X chip, with a cost of approximately 13,000-15,000 RMB).

For example, Huawei's advanced ADAS software is priced at 18,000 RMB. Huawei's intelligent selection mode pricing is a typical separation of software and hardware, with hardware embedded and software charged separately. The hardware embedded cost is included in the car selling price, and the intelligent driving software is charged separately, with a promotional price of 18,000 RMB (original price 36,000 RMB). b. Different domestic competitive landscape

Although Tesla claims to have no competitors in the field of intelligent driving, in the Chinese market, domestic competitors in the field of intelligent driving may not have the same solid position, but the technological gap can still be seen with a telescope.

For example, in the case of Xiaopeng Motors, which is relatively advanced domestically, its urban NOA (Navigate on Autopilot) is about two years behind Tesla's layout. Moreover, Tesla has been more down-to-earth in pricing its intelligent driving features.

So does Tesla have a significant room for price reduction to narrow the gap with its competitors? Dolphin Research believes that in the short to medium term, the room for price reduction is not significant, mainly because the overall cost of Tesla's Full Self-Driving (FSD) is not low:

Currently, when most people are researching, they estimate the cost of Tesla's autonomous driving products to be around $1,000-1,500 (source: "The cost of building a Model 3 by Tesla is 160,000 RMB, and the cost of the ADAS system is 7,000 RMB." - former Tesla executive in charge of autonomous driving).

From the original quote, it can be seen that this BOM (Bill of Materials) cost mainly includes the cost embedded in the vehicle, without considering the higher computational power and corresponding software and hardware development costs required for Tesla's pure vision mode (supercomputing center, DOJO self-developed).

Therefore, comparing this cost with the cost of the domestic counterparts' laser radar + dual Orin chip solution of 20,000-25,000 RMB, it is obviously unfair, and Dolphin Research does not agree with the conclusion that Tesla's pure vision mode has a cost advantage.

Moreover, once we consider the full investment elements of the product, it is natural to understand why Tesla's autonomous driving services have repeatedly increased in price.

c. Regional penetration risk

Currently, on the software side, Tesla's Autopilot (free standard version) and EAP (Enhanced Autopilot, which adds functions such as automatic parking and smart summoning on top of Autopilot) can be used in China.

However, for scenarios such as city driving and highway NOA that are relatively in demand, the corresponding top-of-the-line FSD (Full Self-Driving) version is currently only available in North America and has not been configured in Chinese models. There are several key thresholds for Tesla's autonomous driving business to land in China:

Data collection qualifications: As a foreign company with millions of Tesla owners in China, Tesla's shadow mode is highly likely to be deemed as mapping activities. If the BEV algorithm is deemed to involve mapping procedures, it needs to cooperate with domestic mapping companies that have Class A mapping qualifications.

Data threshold: To ensure data and information security, Tesla is prohibited from transmitting collected data from China back to the United States. Therefore, it needs to establish data centers in China (Tesla has already built a data center at its Shanghai Gigafactory in 2021).

Establishing a supercomputing center: In order to efficiently process and train models with data stored locally, Tesla wants to set up a computing center in China to specifically serve the Chinese market. Option one is to introduce the Dojo supercomputing platform, and option two is to outsource server graphics cards, such as the NVIDIA A100.

Expanding local R&D team: Tesla needs to establish a local closed-loop data team. Improving algorithms relies on analyzing specific scenarios unique to certain regions. At the same time, data needs to be closed-looped within China, so Tesla may need to form a local closed-loop data team to support algorithm adaptation and continuous optimization for niche scenarios.

According to market information, Tesla is currently forming a local operations team of about 20 people to promote FSD in the Chinese market. Meanwhile, Tesla is also attempting to establish a data annotation team in China, with a scale of about hundreds of people, to prepare for training FSD algorithms.

Although there seems to be a solution, the underlying business of autonomous driving is sensitive data monetization. Coupled with the pricing gap compared to competitors, Dolphin Research tends to believe that its penetration rate in China may also be difficult to achieve a high level in the long term.

Based on the above risk considerations, Dolphin Research has separated Tesla's expected FSD penetration rate for existing vehicles in different regions. Overall, China's penetration rate is barely in the double digits, Europe is expected to reach about 30% after maturity, and the US market can achieve a penetration rate of 50% in the long term.

In terms of pricing, Dolphin Research did not directly extrapolate the price of $199 and 12,000 linearly as in the market model. Dolphin Research made reasonable extrapolations based on the purchasing power of different markets, with China being the lowest and Europe and the US being roughly equal, as shown in the graph.

If we take Dolphin Research's coverage of similar internet businesses as a reference, for consumer-facing paid internet services, apart from essential payment for data usage, which can achieve 100% penetration, non-essential products have a maximum paid penetration rate of about 50% for global streaming service Spotify, and a penetration rate of 25%-30% for long-form video in China. Dolphin Research believes this is a relatively reasonable state.

Based on this estimation, Tesla's FSD business revenue in 2027 would be $15.6 billion. Considering the high gross profit and high growth potential of the long-term business, applying a 15x price-to-sales ratio and discounting it back to 2023 at a 10% discount rate, the corresponding valuation is $170 billion.

3) The 3P model for autonomous driving may be an inevitable choice

As a paid internet business model, during the calculation process, Dolphin Research found that if FSD only penetrates Tesla's existing vehicles, with such heavy investment in capital, assets, and technology, Tesla may only have around 20 million active devices by the end of 2030 (the user base is too small to do consumer business and the advertising business for monetizing B2B traffic). After adding a penetration rate, the number of true paying users would be reduced to a few million. In contrast, Apple has over 2 billion active devices worldwide, with over 1 billion high-value iPhone active devices. With high traffic and high user quality, Apple is able to monetize its traffic from flow to payment.

With millions of users and high capital and technology investment, facing only 20 million potential customers, it is difficult to lower the average customer price on one hand, and on the other hand, the commercial scenarios for high-tech products are indeed limited.

Dolphin Research reasonably speculates that in the long-term business path, Tesla's heavy investment model in smart driving services may need to expand beyond internal customers like Tesla and adopt a 3P business model, similar to Huawei's open platform model. Recently, Musk also expressed the possibility of licensing Tesla's FSD to external car brands during the earnings report meeting.

This model has been followed by predecessors as well. Microsoft's Office business, after the decline in market share of Windows in the entire mobile/computer operating system market, began to use the SaaS subscription model, stepping out of the Windows ecosystem and being compatible with Mac/iOS/Android systems, experiencing a second spring for the Office business.

Therefore, in Dolphin Research's valuation of FSD, the valuation space of the 3P business is also taken into account. However, the basic judgment behind this assumption is that "intelligence" does not constitute an ecological barrier for the mutual promotion of automotive software and hardware ecosystems, as the first consideration for users when buying a car is still the energy efficiency, not the advanced driving intelligence.

In terms of core valuation judgment, in Tesla's original retail pricing, due to third-party automakers controlling traffic and acting as channel value, Dolphin Research has halved the selling price of FSD business for other car brands. Due to the uncertainty of the business, the penetration rate is estimated to be in the low single digits. The actual business evolution is as follows.

Based on this calculation, the revenue of Tesla's FSD 3P model in 2027 is 4.4 billion, giving it a 15x PS ratio, discounted back to the end of 2023 at a discount rate of 10%, corresponding to a valuation close to 50 billion US dollars.

Fourth, it is difficult for FSD to replicate Tesla.

Of course, in addition to the FSD business, Tesla may have more commercial monetization opportunities in its pursuit of the AI summit. For example, the FSD business targeting B2B customers (taxi companies, trucks, logistics vehicles) with higher certainty in the long term, as well as the "mutation eggs" along the path of intelligence, such as intelligent robots, open imagination for 3P derived charging piles, and the potential cloud business of supercomputing centers in the future... Due to the abundance of opportunities and the distance to realization, Dolphin Research does not make any estimates in this AI story.

This time, the core is to estimate the valuation of FSD in the B2C business. Overall, the valuation of Tesla's smart driving 1P business is approximately 170 billion US dollars, and the valuation of the 3P business is approximately 50 billion US dollars. In total, Tesla's overall valuation for FSD targeting passenger car owners is 220 billion US dollars (with a single share price of 50-70 US dollars). In terms of valuation, it is reasonable for Tesla to fall within the valuation range of around $150-250 billion, which is in the same range as vertical SaaS products such as Salesforce and Adobe.

However, compared to Dolphin Research's neutral and slightly optimistic valuation of $650 billion and a per-share value of $210 for Tesla's automotive business, this level of valuation contribution is far from being able to recreate Tesla.

In addition, Dolphin Research estimates the value of Tesla's energy business to be approximately $73 billion (due to the homogeneous business model and low barriers to entry in the energy sector, it does not possess significant scarcity value. The main value lies in its current high growth potential. Therefore, Dolphin Research will not go into further detail here. If you are interested, you can communicate privately with Dolphin Research).

Therefore, the combined valuation of Tesla's three business segments with relatively high visibility is $940 billion, with a per-share value of approximately $293. The distribution of the core business value is as follows:

However, it should be noted that in this valuation, the valuation of Tesla's manufacturing business (at $200 per share) and energy business (at $20+ per share) provides a relatively certain valuation that can provide a safety margin for Tesla. On the other hand, the valuation of FSD is essentially in a "pie in the sky" state. When this part of the valuation is included, it becomes a bubble state of valuation sentiment, increasing the risk. Therefore, after this update, Dolphin Research believes that the range of Tesla's overall business distribution should be between $220-300 per share.

For Dolphin Research's previous articles, please refer to:

In-depth Analysis

Earnings Report Analysis/Conference Call Minutes July 20, 2023 Earnings Report Analysis: "Tesla: Only True Fans Dare to Embrace the Trillion-Dollar Dream"

July 20, 2023 Conference Call: "Tesla Minutes: Gross Margin Lost, Tesla May Continue to Lower Prices"

April 20, 2023 Earnings Report Analysis: "Tesla: Promising Plans, Challenging Execution"

April 20, 2023 Earnings Report Conference Call: "Tesla: Confidently Selling Cars at Zero Profit, Harvesting with Autonomous Driving"

January 26, 2023 Earnings Report Analysis: "Tesla's Story Reshaped, Testing the Faithful"

January 26, 2023 Conference Call: "Tesla Minutes: 'No Competitor for Telescope-like Autonomous Driving, Second Tesla May Be in China'"

October 20, 2022 Earnings Report Analysis: "Critical Question: How to Maintain Profitability When Demand is Insufficient?"

October 20, 2022 Conference Call: "Minutes: 'Internal Combustion Engine Cars Will Perish, No Production Cuts at Any Time'"

July 21, 2022 Earnings Report Analysis: "Without the Shanghai Factory's Lifeline, What Can Tesla Rely On?"

Earnings Report Review on July 27, 2021: Tesla: The Best Keeps Getting Better

Telephone Meeting Summary for Tesla's Q1 2021 Earnings Call on April 27, 2021: Tesla's Q1 2021 Earnings Call Summary

Earnings Report Review on April 27, 2021: After Tesla's Unsurprising and Unalarming Q1 Report, What Can We Expect?

Risk Disclosure and Statement for this Article: Dolphin Research Disclaimer and General Disclosure

特斯拉持续下跌，信仰崩塌或是机会？
进击的特斯拉
特斯拉
深度

The copyright of this article belongs to the original author/organization.

The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.