[U29] Pivot: rethinking AI + robotics
Elon Musk plans to commercialize AI-based-robots-as-a-service for space, transportation and labor
Dear Reader,
In this Update for Product | Strategy | Innovation I will cover my thoughts on AI, robotics and Wall Street’s immediate reaction to Elon Musk’s interest in humanoid robots with a 14% selloff in Tesla stock from its $938 close on Wednesday January 26 to an $804 open on Friday January 28. The trigger was not Tesla’s Q4 2021 results, but the conference call that followed the report. Elon Musk, officially Tesla’s Technoking, joined the call to provide an update on the company’s product roadmap.
Tesla’s Q4 2021 earnings beat expectations with almost 90% revenue growth year-over-year in the automotive business and 30% gross margins with pricing power in the battery electric vehicle (BEV) business. At first glance, it seemed earnings fell short, but this was due to one-time charge for payroll taxes associated with the CEO compensation plan. On the earnings call, Tesla reported 2 new gigafactories were on schedule and Giga Texas Model Y production would start with the new 4680 battery cells. Tesla reiterated guidance for 50% unit production growth in 2022 and stated this could be accomplished without any contribution from the 2 new factories. Production in the new Giga Berlin and Giga Texas facilities would thus be upside to their forecasted production growth in 2022.
Elon Musk said Tesla’s highest priority for new product development in 2022 is a humanoid robot code-named internally as Optimus.
Elon Musk also discussed just how significant the Full Self Driving (FSD) program will be on Tesla’s future profitability and the progress Tesla is making with the FSD beta testing program. Tesla reported they have 60,000 beta users in the program as of January 2022 with 150,000 customers using the Tesla Safety Score to qualify as either an FSD beta user or as part of the Tesla Insurance program. Before Elon Musk transitioned to the product roadmap portion of the call, he stated scaling production of the current BEVs was the highest priority for 2022. Adding new vehicle models adds complexity to the supply chain and with the current semiconductor chip supply shortage, any additional burden to supply chains would not be productive. This means scaled production of the Cybertruck and new Roadster are delayed until at least 2023 and the $25k entry-level BEV after those models.
But when Elon Musk discussed the product roadmap beyond ramping 4680 battery production for the Model Y and accelerated FSD development with the Project Dojo neural network training supercomputer starting around mid-2022, he did not discuss a product in the BEV pipeline. Elon Musk said Tesla’s highest priority for new product development in 2022 is a humanoid robot code-named internally as Optimus. This robot was first introduced in a deep dive presentation on Tesla’s artificial intelligence (AI) work at Tesla in 2021. Wall Street can model vehicle production and sales projections and even the margin improvement with FSD. But Wall Street doesn’t know how to model revenue projections for a humanoid robot. But that is one of Wall Street’s problems. Wall Street has trained public companies to feed it stock buy-backs and dividends because these just plug and play into validated financial models. Breakthrough innovation is often punished not rewarded.
I will cover the evolution of business strategy for Elon Musk in this update. The strategy does not limit other business models, but forms a recurring revenue model combining artificial intelligence (AI) and hardware robotics into revenue generating services versus selling products. I will cover 3 primary domains across 2 companies. The first company is the privately held SpaceX with a mature AI and robotics platform to launch satellites into low Earth orbit for both government and commercial customers. The second company is the public company Tesla with a mission to accelerate the transition to sustainable energy. AI and robotics are used to improve Tesla’s manufacturing operations, but end-user applications are less mature and are still in development.
There are 3 primary domains for AI-based-robots-as-a-service:
Space: launch satellites into low Earth orbit with reusable components to lower cost;
Transportation: deploy on-demand BEV robo-taxi fleets operating 20 hours per day, 7 days per week to accelerate the transition to sustainable energy; and
Labor: deploy humanoid robots operating 20 hours per day, 7 days per week to automate routine tasks/processes in industries where labor is in short supply like construction and manufacturing.
SpaceX is for Space
Space Exploration Technologies Corporation doing business as SpaceX was founded in 2002 by Elon Musk and is headquartered within the Greater Los Angeles metropolitan area in Hawthorne, California. Elon Musk is CEO and CTO. Gwynne Shotwell is President and COO. SpaceX manufactures and routinely launches Falcon 9 rockets to deliver satellites into low Earth orbit in 2 stages for government and commercial customers. SpaceX currently operates 4 launch facilities. The Space Launch Complex 40 (SLC-40) at Cape Canaveral, Florida; Space Launch Complex 4E (SLC-4E) at Vandenberg Space Force Base, California; Launch Complex 39A (LC-39A) at Kennedy Space Center, Florida; and the South Texas Launch Site near Brownsville, Texas.
Robo-Rockets-as-a-Service
The 1st stage of the Falcon 9 is a reusable fuel tank booster with 9 Merlin engines. The 2nd stage separates from the 1st stage to complete the transport with only 1 Merlin engine. The 1st stage booster returns to Earth about 8 minutes after liftoff to land on a designated landing pad for recovery and reuse on another mission. Falcon Heavy adds 2 additional 1st stage reusable fuel tank boosters with Merlin engines to the primary Falcon 9 fuel tank booster. The payload for Falcon Heavy is atop the 2nd stage rocket and can either be a Dragon transport vehicle to shuttle supplies to the International Space Station (ISS) or more recently Crew Dragon to transport astronauts to the ISS.
The video above shows the January 31, 2022 Falon-9 Launch from SLC-40 at Cape Canaveral starting around T - 00:01:40. The weather was clear and the southern flight path provides incredible views of the 2nd stage rocket separation from the 1st stage rocket at around T + 00:02:20. I was able to watch this launch live from Cocoa Beach just south of Cape Canaveral and could see the rocket launch and separation. It was incredible to watch. The 1st stage fuel tank booster lands at Landing Zone 1 in the video starting around T + 00:07:00 and included a time delayed sonic boom. I watched another Falcon 9 launch on February 3rd from the same location. The process just seems routine now. I would imagine a Falcon Heavy launch is even more impressive with even more rocket engines and 3 booster rockets returning for landings on Earth.
SpaceX is planning 52 launches in 2022 primarily for satellite deployments into low Earth orbit using the reliable Falcon 9 and Falcon Heavy systems. These space launch services generated an estimated $1.6 billion in revenue for SpaceX in 2021 using a fleet of rockets. SpaceX uses artificial intelligence (AI) to assist autopilot navigation for space travel. Simulations and a growing space data set power AI models. SpaceX has partnered with Microsoft to develop cloud services on MS Azure for space data.
Falcon Heavy, Starlink, Starship and Dragon are other SpaceX programs in various stages of development, but the Falcon 9 “AI-driven-space-robotics-as-a-service” program is the workhorse that generates sustainable revenue for the forward looking company. The other programs leverage and expand on the core AI and robotics systems.
Robo-Satellites-as-a-Service
Starlink in particular is a SpaceX program building a constellation of low Earth orbit satellites to provide broadband internet service worldwide including remote locations for $99 per month. Falcon 9 rockets deploy about 60 new Starlink satellites with each launch. Starlink provides a new stream of revenue beyond rocket launches. Starlink currently has about 1,700 satellites operating in low Earth orbit to provide broadband internet service to 145,000 customers across 25 countries. SpaceX has also partnered with Google Cloud to integrate Starlink ground stations directly into Google data centers to improve services for enterprise customers.
But the long term objective for SpaceX is to deploy 42,000 low Earth orbit satellites to improve coverage for worldwide services. Starlink leverages this growing fleet of low Earth orbit satellites to provide another service to generate recurring revenue at scale estimated at $30 billion annually for SpaceX. This revenue stream will support deep space exploration to establish yet another recurring revenue stream with ongoing commercial space transportation beyond Earth.
Deep space transportation would be done as a repeatable service for government and commercial customers with reusable rockets. The primary cost would be fuel if the cost to manufacture and operate the rocket is carried across many missions. Tesla is also developing a CO2 Rocket Fuel program to produce fuel for missions to destinations like Mars that would require fuel production on-site for return launches back to Earth. The program would also remove CO2 from Earth’s atmosphere for fuel produced here.
Tesla is for Transportation and Labor
Tesla was founded July 1, 2003 in San Carlos, CA. The original Tesla Roadster started production in 2008 with production of the Model S in 2012 and Model 3 in 2017 expanded the product portfolio beyond an electric sports car. The Model X and Model Y provided SUV options for both platforms. 2021 was a breakout year for Tesla with almost 73% automotive revenue growth compared to 2020, 29.3% automotive gross margins and the build out of 2 new gigafactories to increase vehicle production with 4 factories across 3 continents.
On a recent earnings call, Tesla CEO Elon Musk stated the primary objective for Tesla in 2022 is to scale vehicle production across all 4 of these factories. New model production for the Cybertruck and Roadster is delayed to reduce complexity with semiconductor chip shortages and ramping a new battery design. Just like the Falcon 9 rocket system is the workhorse for SpaceX, battery electric vehicle (BEV) production, sales & deliveries generate the majority of revenue for Tesla.
Tesla started installing a 2nd generation of hardware sensors and computer on all its new vehicles and pre-selling Enhanced Autopilot as a service in 2016 for an additional $5,000 on these vehicles. Full Self Driving (FSD) was offered for an additional $3,000. Tesla eventually offered basic Autopilot for highway driving on all new vehicles as part of the core features. FSD remained an added service priced at $8,000. 2019 brought the 3rd generation of hardware sensors, custom FSD microprocessor and FSD computer to improve its edge computing capabilities and performance with an in-vehicle neural network. Vehicles collect and upload data to the Tesla Cloud to improve the neural network with the growing Tesla BEV fleet. As the neural network improves with training, it is eventually deployed to the fleet to improve performance as an updated software version.
Robo-Navigation-as-a-Service
Tesla alpha tests FSD versions with employee drivers. But in October 2020, Tesla expanded use to about 1,000 FSD beta users that included some customers who had purchased FSD for their vehicles. Tesla increased FSD to $10,000 in 2021. In May 2021, the beta tester program expanded to about 2,000 employees and customers who had purchased FSD. In October 2021, Tesla further expanded the FSD beta tester program using a Tesla Safety Score to risk stratify interested customers to admit only the safest drivers in cohorts of about 1,000 new users.
Tesla said there were 11,700 FSD beta users in November 2021 and then 60,000 FSD beta users in January 2022. Tesla data report 12% of new vehicles purchased in 2021 bought FSD. Tesla also introduced a $200 monthly FSD subscription in 2021. There were also about 150,000 users of the Tesla Safety Score in January 2022. The Safety Score is used to qualify potential FSD beta users and also to set dynamic pricing for Tesla Insurance. I purchased FSD for my Tesla Model Y just like a “call option” to potentially return the full purchase price of the vehicle or even more when the robo-taxi option is available in the future.
AI is core to FSD with a neural-network processing realtime sensor data onboard to navigate and control the vehicle using computer vision. Some call Tesla vehicles a computer on wheels, but as FSD evolves and becomes more widely adopted for a growing percentage of miles driven, Tesla vehicles become more of a robot with wheels. Tesla even predicts the steering wheel and foot pedals will eventually be removed to reduce weight and cost for a dedicated, fully autonomous, BEV robo-taxi. Tesla’s technology stack to enable robo-taxis includes:
Vehicle sensors, FSD neural-network microprocessors, FSD computer;
Neural network model onboard dedicated to computer vision and vehicle control;
Project Dojo neural network training supercomputer to automate and accelerate training neural networks for specific use cases like computer vision;
Robo-taxi application to enable a BEV to operate as a robo-taxi;
(Future) Admin dashboard for operators to manage a fleet of robo-taxis routine tasks automated by the system; and
(Future) Mobile App to request a service from a robo-taxi like transportation, food delivery or courier.
Robo-Taxis-as-a-Service
FSD provides high margin software sales either with the purchase of Tesla vehicle, after vehicle purchase as a one time sale or with a monthly subscription. But Tesla’s long-term objective for FSD is the opportunity it provides to leverage AI and robotics to offer fully autonomous Transportation-as-a-Service (TaaS). This service would be similar to Uber and Lyft, but without a human driver in a BEV to lower the operating cost to $0.20 per mile or less. Operators like Tesla could price TaaS with BEV robo-taxis to consumers at about $0.25 per mile at scale.
When priced competitively and widely available, fully autonomous BEV TaaS will not only disrupt ride hailing services like Uber and Lyft, but the need for vehicle ownership with many individuals and families. A busy BEV robo-taxi operating 20 hours per day, 7 days per week could do the work of 4-5 internal combustion engine (ICE) vehicles used for only 3-5 hours per day. So Elon Musk can look at scaling a BEV robo-taxi fleet as the most strategic option available to execute Tesla’s mission of accelerating the transition to sustainable energy if each BEV produced as a robo-taxi displaces 4-5 ICE vehicles.
However, as Tesla continues to scale BEV unit production to build operating efficiencies, reduces battery costs and position Tesla for a lower priced BEV option, the financial models for a robo-taxi fleet improve dramatically.
Hertz reported an order of 100,000 Tesla Model 3 vehicles in 2021 for its rental fleet. And then Uber reported a deal with Hertz to offer Tesla vehicles to its Uber drivers in the near-term. This could evolve into Hertz or a Hertz/Uber partnership operating a Tesla fleet of robo-taxis. But a key strategic question for Tesla in the future will be do they need the short term vehicle sales and revenue share with a third party or should Tesla deploy its own capital to build and operate its own Tesla fleet of robo-taxis. The optimum strategy may be a hybrid of Tesla owners putting their own vehicles using FSD on-demand into a pool of available robo-taxis, a limited number of third parties operating fleets of dedicated Tesla robo-taxis in specific markets, and Tesla operating the majority of robo-taxis with its own Tesla BEVs over time when Tesla has available capital to do so. If Tesla continues to scale a profitable core BEV business, they will need opportunities to deploy capital into other profitable areas of the company.
The technology and regulatory hurdles for robo-taxis still require substantial progress. However, as Tesla continues to scale BEV unit production to build operating efficiencies, reduces battery costs and position Tesla for a lower priced BEV option, the financial models for a robo-taxi fleet improve dramatically. These enablers are likely to converge over the next 5-10 years. That is when Tesla may use production of a lower priced BEV to build fleets of dedicated robo-taxis to generate recurring service-based revenue just like SpaceX with its growing fleets of autonomous rockets and satellites.
Tesla Semi provides another large business service opportunity for logistics.
Within the next year or two, Tesla will probably stop selling FSD as a one-time fee and just offer FSD by monthly subscription to Tesla vehicle owners when they want to use FSD capabilities. Selling FSD for a one-time fee cannabalizes the recurring revenue potential from monthly subscriptions and Tesla’s own BEV robo-taxi fleets.
Tesla Semi provides another large business service opportunity for logistics. Tesla is selling Semis to large corporate accounts like PepsiCo and Walmart and truck operators who want to lower their total costs per mile. Autonomous navigation is a major high margin opportunity for all corporations operating fleets of commercial trucks with the growing shortage of truck drivers. But Tesla can also deploy fleets of its own autonomous Tesla Semis to move freight for end-users at a negotiated price per mile. Medium-sized companies might prefer to manage their own logistics between their distribution centers, but outsource it as a service from Tesla.
Another benefit of FSD’s scale with the growing number of beta users and daily miles driven is the opportunity the evolving technology provides to train a neural network for computer vision and control using sensor and camera data. When drivers override FSD these events help build the long tail of edge cases to improve performance. Project Dojo will introduce a new neural training supercomputer to accelerate training with even larger models in 2022. This evolving capability is funded with an active commercial FSD use case currently generating an estimated $1.2 billion in high margin revenue, but future use cases will benefit as Project Dojo performance improves. And that brings us to the next domain.
Robo-Labor-as-a-Service
One use case communicated by Elon Musk in 2021 for Tesla’s evolving computer vision and neural network expertise is a robot in human form to provide labor for basic routine tasks. These humanoid robots could work 20 or more hours per day allowing some time to recharge batteries on breaks. Current specifications assume the humanoid robot weighs 125 lbs with a height of 5’8” and can carry 45 lbs and deadlift 150 lbs. Tesla refers to this robot design internally as Optimus and Elon Musk recently stated he is the Optimus project lead. Elon anticipates early versions of the Tesla Bot would be used at Tesla and potentially SpaceX to perform routine tasks to improve and verify system components and validate the overall robot design for specific tasks like material handling.
The Tesla Bot technology will take a number of years to improve enough to take over a growing number of routine tasks. However, Amazon has already scaled advanced robotics to improve efficiencies stocking inventory and fulfilling orders to expand Amazon Prime for 2-day deliveries. Tesla can just repurpose the Tesla FSD microprocessor and computer assets plus sensors and cameras from the BEV supply chain to build early Tesla Bot prototypes. Mobility could be simplified with stationary tasks and with conventional wheels to start acquiring video to identify hazards and landmarks at specific sites for specific tasks. This would allow early humanoid prototypes to start with some basic autonomy in a pre-populated neural network.
What is the business case for Tesla Bot just within Tesla’s own operations? Tesla will reach 100,000 employees in 2022 to meet an annual vehicle production run rate of 2 million vehicles. Many of those employees work in manufacturing and distribution. Let’s assume at scale the Tesla Bot could do the work of 20% of Tesla employees or 20,000 workers using 2022 estimates. If these workers make $20 an hour, work 40 hours a week, are paid for all 52 weeks with paid time off and total salary includes 30% to cover benefits, then:
$20/hour * 40 hours/week * 52 weeks / 0.70 for benefits = $60,000 per year per worker
20,000 workers * $60,000 / worker = $1.2 billion per year is cost for this cohort of human labor.
The Tesla Bot materials are estimated to cost around $25,000 per robot.
The most productive workers are estimated to do 2x the work of the average for all workers. The Tesla Bot is assumed to match and sustain the productivity of the most productive worker or 2x the average for a cohort.
If the Tesla Bot worked 20 hours per day 7 days per week that would be 140 hours per week. Assuming the average worker completes 40 hours of work per week, the Tesla Bot could match at least 3x the average for the cohort.
So the overall efficiency of a Tesla Bot is 6x the average worker.
Manufacturing and warehouse workers are also at risk for about 24 workplace musculoskeletal disorder incidents per year per 100 workers based on Bureau of Labor Statistics estimates. That is another expensive line item cost associated with human labor doing routine tasks that would be saved with Tesla Bots who would not suffer from these injuries even with 6x the overall efficiency of the average worker. Note: These injury costs have been excluded from my current analysis.
Matching 20,000 workers with a fleet of humanoid robots would require 3,334 Tesla Bots at a cost of $83,350,000 plus the cost for assembly and electricity to operate robots for 1 year. Beyond year 1, the electricity and maintenance become the primary costs to operate the fleet of Tesla Bots. But even in year 1, the ROI is estimated to be 14x for the materials alone. If the cost to assemble and operate are less than $10,000 per robot for year 1, the ROI would be at least 10x. And assuming, $10,000 is the ongoing annual cost to operate and maintain each robot, the ROI would increase to 36x a year for the service life of the robot until replacement is required around year 5.
Using Tesla’s Q4 2021 operating margin of 14.7%, saving $12 billion is equivalent to generating another $81 billion in top line revenue or 1.7x Tesla’s 2021 automotive revenue for the entire year.
Now model what happens not in 2022, but in 2030 when Tesla vehicle production is estimated to reach 20 million units annually. If 100,000 employees are needed to produce 2 million vehicles, then 10x the human labor is needed to produce 10x the vehicles in 2030 assuming the increase in labor is proportional to unit production. Now imagine 10x the required Tesla Bots or 33,334 robots with material costs of $833,500,000 using current assumptions returning an ROI of 10x in 2030 to save $8,335,000,000 and an ROI of 36x per year in 3031 through 3034 on the all-in annual cost of $333,340,0000 for the Tesla Bot fleet to save the entire $12 billion a year in labor for 200,000 employees. Keep in mind those savings drop right to the bottom line to enhance after tax earnings.
Using Tesla’s Q4 2021 operating margin of 14.7%, saving $12 billion is equivalent to generating another $81 billion in top line revenue or 1.7x Tesla’s 2021 automotive revenue for the entire year. This would be equivalent to total production from 3.4 additional factories using Tesla’s 2021 production with only 2 factories and the sales, operations and logistics required to deliver all the vehicles produced to generate the revenue. Elon Musk as Tesla’s chief executive and Technoking is focused on ramping vehicle production and new factories to reach annual production of 20 million vehicles by 2030.
Given all the supply chain and vehicle product mix challenges and regulatory hurdles to open new factories, Elon probably balanced that perspective with the new product opportunity to design and build a fleet of humanoid robots. Tesla already has a lead to develop the Tesla Bot using existing technology and its innovation know how across both Tesla and SpaceX. Why wouldn’t you accelerate development of the Tesla Bot ahead of all other product development in 2022? Compared to autonomous rockets, satellites for internet communication, self-driving vehicles and robo-taxis, developing the Tesla Bot with an initial focus only on Tesla applications justifies its prioritization and business case assuming Tesla can scale production in 8 years to only 35,000 robots with adequate performance to do the work of 200,000 Tesla workers.
Tesla Bots could also play a significant role as SpaceX establishes deep space travel beyond Earth.
Tesla will also work with other companies to accelerate the adoption and supply of sustainable energy, scale Tesla Semis for logistics, other BEVs for business use and Tesla Bots to automate work for routine tasks. The business model could either be a monthly subscription for a fleet of Tesla Bots or a revenue share model on the human labor savings. Over time these saving would likely be reduced as robots replace more jobs. Robots would eventually need to operate under a paid license or pay some kind of tax as if they were labor to offset lost human labor wages and tax revenue. Longterm, vocational training would be able to retrain the displaced workforce to do higher skilled jobs, but revenue is required to fund these programs and underemployment. But the savings for Tesla and other companies to automate routine tasks with Tesla Bots will still be substantial with a significant ROI.
Tesla Bots could also play a significant role as SpaceX establishes deep space travel beyond Earth. Establishing any remote space exploration bases would require labor operating with life support in extreme conditions. These requirements do not have the same limitations for Tesla Bots designed to operate in extreme environments on the Moon, Mars and other destinations. Initial missions would need to transport materials and supplies, but Tesla Bots could reduce the need for human astronauts to do more routine tasks. Missions limited to Tesla Bots to unload materials and establish a basic living environment would de-risk missions including early human astronauts and Tesla Bots to construct a core space exploration base for the first groups of space explorers. Hostile environments on Earth like fires, explosive demolition, nuclear radiation events, chemical spills and other hazards could also use Tesla Bots to improve safety for human operators who deploy robots into these various tasks.
Conclusion
AI and robotics are converging fast. These technology platforms can combine computer vision, neural networks, hardware and mobility or flight to create repeatable and reliable autonomous use cases. Costs are dropping too as sensors and computing from scaled commercial technologies offer ready to use components. And neural networks for computer vision are rapidly advancing the capabilities for full autonomy to navigate and respond to an ever changing environment.
Elon Musk is in a unique position leading various companies, projects, products, operations, manufacturing, and services in various stages of development and commercialization. He is very involved in design reviews and product roadmap discussions with engineers at the various companies even as the CEO. In some cases, where uncertainty is high with new product development, Elon serves as project lead. He is CEO and CTO at SpaceX. This provides Elon a unique perspective to balance challenges in one area with opportunities to tackle these challenges with technologies from other areas or even different companies under his leadership. Tesla and SpaceX share material scientists to advance proprietary alloys used in both companies.
I also think Elon is gaining confidence in Tesla’s ability to realize its mission to accelerate the transition to sustainable energy.
Elon Musk clearly sees the overlap between Tesla’s growing expertise in neural network computing and training for computer vision and the opportunity to repurpose autonomous robotics technology from transportation to humanoid robots to automate routine tasks. This opportunity can be used with a significant ROI at scale to create the equivalent of multiple synthetic Tesla factories and operations generating the equivalent revenue to match the savings in labor costs. Knowing Elon he will take on the burden himself of retraining the displaced workforce in cities with his factories due to the local impact. Maybe he will call it the Optimus Vocational School to prepare workers for the trades requiring humans to build a civilization on Mars.
I also think Elon is gaining confidence in Tesla’s ability to realize its mission to accelerate the transition to sustainable energy. Tony Seba, independent of Elon Musk, has predicted an energy abundance within a decade just based on the business economics of deploying solar, wind, and other sustainable sources of energy with scaled battery storage. Tesla is piloting virtual power plants in cities where residential battery storage and solar energy can supply enough energy with peak demand to reduce the need for natural gas and coal-fired “peaker” plants to generate the required energy during surges in demand. This supply of energy from virtual power plants uses yet another Tesla AI-powered platform to automate energy arbitrage between residential customers and electric utility companies.
Tesla Bot presents a new “keyman risk” if Elon Musk is not able to launch and see several iterations of the product design through to reduce commercial uncertainty around a humanoid robot capable of replacing human workers doing routine tasks.
An abundance of scalable, sustainable energy and abundant labor with robots creates the opportunity to scale the global economy with an expanding GDP. This step function in global GDP is what can make commercializing space possible with an active space exploration base on the Moon, colonizing Mars and the ongoing supply needs between Earth and these deep space destinations. It’s almost like opening up commercial trade across the ocean with ships and then Europe colonizing the Americas. Growing commercial operations made the expansion to new territories possible. Not so great for the native populations of North America, but the Moon and Mars seem rather unpopulated for now.
I will also add three concerns.
Elon Musk could be focusing on Tesla Bot because of the ongoing supply chain and logistics challenges in 2022. So far Tesla has navigated the supply chain to produce vehicles and the logistics to deliver those vehicles worldwide better than any other vehicle manufacturer. But risks remain and it would be nice to have a shiny toy to divert attention if needed. I actually think this is a much smaller risk. I think Elon Musk has already determined a $3 trillion valuation is already “baked” for Tesla and the operations just need to grow into it. Elon is thinking about the next growth horizon beyond BEVs and energy for Tesla.
Tesla’s core operations to design and manufacture BEVs at scale; design and manufacture battery systems at scale for vehicles, residential energy storage and grid-scale energy solutions; and design and scale neural networks for multiple applications have found outstanding engineers, designers and operators to advance these projects and businesses. Elon still plays a role for new BEVs and manufacturing, but the Tesla Bot really requires his vision and ability to push innovation as a co-founder, significant owner and chief executive at Tesla. Therefore, Tesla Bot presents a new “keyman risk” if Elon Musk is not able to launch and see several iterations of the product design through to reduce commercial uncertainty around a humanoid robot capable of replacing human workers doing routine tasks.
Wall Street may become further disinterested in the Tesla Bot and punish TSLA the stock just like it did when Tesla was having issues scaling Model 3 production. Any hiccups with the supply chain, regulatory issues with FSD, missed BEV unit production numbers, etc. could cause a significant sell off in Tesla stock independent of improving financials and earnings just based on a change in sentiment. You could argue Tesla’s inclusion in the S&P 500 might help mitigate this risk, but FB sold off 26% in one day this week for similar reasons. TSLA is very volatile stock. If you want to play in the sandbox, know the risks.
We’ll see what the future holds, but an economy without constraints on sustainable energy and entry level labor can scale GDP proportional to the aggregate business opportunities and access to capital. All boats rise if Tesla can scale a commercially viable Tesla Bot. The narrative around Tesla just keeps getting better even with the wide swings in sentiment. And after watching 2 SpaceX Falcon 9 “space-robots-as-a-service” missions launch over the past week from Cape Canaveral to put satellites into low Earth orbit and reclaim their reusable stage 1 fuel tank boosters, the Tesla Bot just seems like another technical challenge for an Elon Musk team to conquer. If Tesla is successful, the business opportunity should drive all the demand they can handle.
Best,
Stephen
I’m long AMZN, GOOG and TSLA mentioned in this update. Nothing in this post is intended to serve as financial advice. Do your own research.