Tesla Model 2 2026: $25K + Terafab chips + LG LFP batteries + Gigacasting full-stack rollout

Kind: captions Language: en Imagine walking into a car dealership and instead of seeing a gleaming vehicle on a pedestal, finding it completely disassembled. Every wire exposed, every weld visible, every internal part on display. As if someone had opened a Swiss watch and left it on the table for anyone to examine. That's exactly what Tesla did in showrooms in China and delivery centers in Luxembourg, displaying a completely disassembled Model Y to the public. It's not a common strategy in the automotive industry. Ford, GM, and Toyota never did anything like it. And it's no coincidence that Tesla chose to do things differently. What's most striking about this display isn't the visual spectacle, but what it reveals about how the car is built. At the heart of it all is a technology called giga casting, a process in which absurdly powerful pressure machines, nicknamed giga press, fuse liquid aluminum in gigantic molds to create the front and rear sections of the vehicle as a single solid piece. It sounds simple, but the impact is enormous. Traditional automakers use dozens, sometimes hundreds, of stamped and welded metal parts to form the same structure. Each weld is a potential point of failure. Each joint is an opportunity for rust, noise, or premature wear. With giga casting, these points of failure virtually disappear. The structure becomes more rigid, lighter, and simpler to assemble. Independent studies have already placed Tesla vehicles among the cheapest to maintain over time, and this technology is one of the main reasons for this. Fewer parts mean fewer things to break, fewer visits to the repair shop, and less money coming out of the owner's pocket. It's the kind of detail that doesn't appear in the sales brochure, but makes a real difference in the life of someone who uses the car every day. Anyone who follows the news in the electric vehicle sector has probably heard of the cybercab, Tesla's robot taxi that promises to circulate through cities without a driver. Since the project was announced, one of the promises that attracted the most attention was the idea that the vehicle would charge completely wirelessly, without a plug, without a cable, without any physical contact with the outlet. It was an ambitious vision, and also quite seductive, because it makes perfect sense in an autonomous car that needs to manage on its own in the city without anyone to connect a charger. But the reality of the launch, scheduled for later this year, brought a change that few expected. The first production cybercab models arrived with a physical charging port of the NACS type, the same standard already used in other Tesla vehicles. The integration was completely redesigned with a clean finish, a sealing gasket around it, a flexible structure in the housing, and without the motorized flap that exists in current models, probably to keep replacement costs low in case of damage. It's a functional, discreet, and well-resolved solution. But above all, it's a pragmatic decision. And understanding the reason behind it says a lot about how the company thinks. Wireless charging hasn't been abandoned. Tesla even obtained FCC approval for the ultra-wideband technology needed for this system to work. The problem isn't technical, it's infrastructural. There aren't enough inductive charging points in any city in the world right now. Launching a vehicle dependent on a technology the world isn't yet ready for would be like selling a phone that only works on a network that hasn't been built yet. It doesn't make sense, and Tesla clearly wasn't willing to hold back the launch because of that. The correct interpretation here is that the physical plug is today's solution, and wireless charging is the plan for tomorrow. The trend is for this functionality to arrive as a future hardware update as inductive charging infrastructure is installed in cities. It's a model the company has used before. Introduce the technical capability first and activate the feature later when the surrounding ecosystem is ready to support it. Those who buy or use a cybercab from the start won't be stuck with outdated technology, but they also won't have to wait for the future to get their car on the road. There's a detail in the cybercab that's almost never mentioned in reviews of the vehicle, but it says more about the philosophy behind the design than any technical specification. The height of the interior seats was deliberately calculated to match the standard dimensions of a wheelchair. It wasn't an engineering accident, nor a design coincidence. The cybercab's engineering manager himself publicly confirmed that this choice was intentional, and that the vehicle's famous butterfly doors were designed in conjunction with this specific height to create a functional transfer space between the wheelchair and the car seat. In practice, this means that a person in a wheelchair can position the equipment parallel to the cabin, open the butterfly doors that extend wide to the sides, and transfer between the chair and the vehicle seat without needing anyone's help. For a regular car with a driver, this lack of assistance would be merely a convenience. For an autonomous robo-taxi, where there are no humans inside the vehicle besides the passenger, it is an absolute requirement. A driverless vehicle that cannot be used independently by people with reduced mobility is not public transport. It is an exclusive service disguised as innovation. This concern goes beyond accessibility itself. It touches on an issue that is rarely discussed when talking about autonomous vehicles, which is who this technology really serves. It's easy to imagine a robo-taxi as something aimed at young urbanites with smartphones, but a fleet of autonomous vehicles circulating in a city needs to be available to the elderly, people with physical disabilities, to those who depend on public transport, and have no other option. If the vehicle wasn't designed for these people from the start, the technology solves the mobility problem for a small segment of the population and ignores those who need it most. Beyond this engineering choice, the cybercab arrived with a set of hardware features that reinforces the idea of a vehicle designed to operate without full-time human supervision. The external cameras are larger than those used in conventional models, and each one has an integrated high-pressure washing system to ensure visibility even in adverse conditions of rain, mud, or dust. There is also an internal camera facing the trunk, designed to identify items forgotten by passengers after each ride. These are simple solutions, but they demonstrate the level of detail applied to a vehicle that needs to operate continuously and autonomously. A few weeks ago, Elon Musk announced something that went almost unnoticed amidst so much other news, but which has the potential to change the cost structure of every vehicle Tesla will produce in the coming years. The project is called Terafab, and the idea is to build its own semiconductor factory in Austin, Texas. Days after the announcement, the first job openings were already available, including a position for a semiconductor infrastructure technical manager with proven experience in projects exceeding $100 million in capital. This is not a project on paper, it's an ongoing operation. To understand why this matters, it's necessary to know where the chips Tesla uses today come from. The AI 4 chip, responsible for processing the FSD autonomous driving system in all current models, is manufactured by Samsung. The next generation, the AI 5, is already designed and will be produced jointly by TSMC and Samsung, with mass production expected to begin in mid-2027. After that, Tesla has already signed an agreement with Samsung to manufacture the AI 6 chips in factories located in the United States. [music] Terafab is the next step after all this, the stage where the company ceases to depend on any external partner and begins to control the entire process internally. Building a chip factory is, without exaggeration, one of the most complex industrial tasks that exist. The tolerances involved in the semiconductor manufacturing process are so small that a vibration caused by a nearby highway can compromise an entire production run. The geographical location of the factory, the design of clean rooms, the control of temperature, humidity, and airborne particles, everything needs to be resolved before the first chip is manufactured. It's the kind of challenge that entire governments take decades to master. Tesla is trying to do this with the same speed it used to build its giga factories for batteries. The connection to the rest of the product line is direct. Musk has publicly identified chip supply as the biggest bottleneck [music] in scaling both the FSD and the Optimus humanoid robot. A car designed for the mass market with an aggressive price needs an autopilot chip that costs as little as possible to produce. Today, each chip that goes into a Tesla vehicle is a chip that is negotiated, awaited, and purchased with a profit margin from a third-party supplier. When Terafab is operational, this equation changes completely, and the savings generated in each unit produced will appear directly in the final cost of the vehicle. While the world debated tariffs and trade tensions between the United States and China, news emerged from an unusual place for the automotive sector, the Indo-Pacific Energy Security Summit. It was there that the US Department of the Interior confirmed a $3 billion agreement between Tesla and LG Energy Solution to build a prismatic LFP battery cell factory in Lansing, Michigan. Production is scheduled to begin in 2027 with an installed capacity of 50 gigawatt-hours per year from a single plant. For those unfamiliar with LFP chemistry, it's worth understanding what makes this type of battery so relevant to Tesla's plans. LFP stands for lithium iron phosphate, and it's a chemistry that trades the higher energy density of nickel cells for a number of practical advantages that make all the difference in large-scale applications. First, durability. LFP cells withstand more than 10,000 charge and discharge cycles without significant degradation, a number that leaves conventional nickel batteries far behind. Second, safety. The risk of spontaneous combustion, known as thermal runaway, is virtually nonexistent in this chemistry. Third, and perhaps most important for the mass market, cost. LFP cells do not use cobalt, one of the most expensive and ethically problematic materials in the entire electric vehicle supply chain. The Lansing plant will not be alone. It joins an LFP cell plant that Tesla already operates in Sparks, Nevada, and a lithium refinery already operating in Corpus Christi, Texas. When these three operations work together, what is formed is an end-to-end domestic supply chain from raw ore to ready-to-use cell without depending on any Chinese supplier at any stage of the process. [music] At a time when export restrictions, tariffs, and geopolitical instability have wreaked havoc on the global automotive sector. Having this chain within the country is not just a logistical advantage. It's strategic protection. The fuel cells produced in Lansing are initially destined for Megapack 3 systems, which are the industrial-scale energy storage units that Tesla supplies to power plants and grids. But the logic behind this investment goes beyond the energy market. A factory with a capacity of 50 gigawatt-hours per year produces far more than the Megapacks need, and LFP chemistry is exactly what makes sense for a competitively priced entry-level vehicle. It doesn't take much effort to see where this surplus capacity can be directed when demand for a more affordable car starts to grow. When Tesla launches a new version of the Megapack, it's not the kind of news that usually dominates tech headlines, but the Megapack 3 is different. And understanding what has changed helps to see a layer of the company's strategy that rarely appears in conversations about electric cars. The new energy storage system has been redesigned from scratch, and the most striking number is the 78% reduction in internal connections compared to the previous model. In a structure that operates continuously, exposed to variations in temperature, humidity, and the electrical demand of a distribution network, each connection is a potential point of failure. Reducing this number by almost 4/5 is not an incremental improvement. It's a reliability shift. The Megapack 3's thermal system has also been redesigned, and the chosen solution is curiously familiar to those who know Tesla vehicles. >> [music] >> The temperature control of the cells uses an improved version of the Model Y's heat pump, the same component that regulates the cabin and battery temperature in the company's best-selling SUV. Reusing a mature and extensively field-tested technology in a completely different product is a smart engineering decision. It reduces development time, lowers manufacturing costs, and leverages all the learning curve the company has already accumulated with millions of Model Y units on the road worldwide. On the installation side, Tesla is launching a platform called Megablock alongside the Megapack 3, which functions as a plug-and-play system for assembling industrial-scale energy storage plants. The impact on construction costs is significant. Savings reach $1 for every 2 and 1/2 dollars invested, representing a 40% reduction in the total deployment cost. And the time to put a 1 gigawatt-hour system into operation has dropped to just 20 working days. To give you an idea of what this represents, 1 gigawatt-hour is enough energy to power 400,000 homes. This volume of capacity being installed in less than a month is a number that didn't exist in the sector until now. The relevance of this to the electric vehicle supply chain goes beyond the energy market. The more Tesla manages to lower the cost and accelerate the installation of storage systems, the faster the renewable energy infrastructure grows, powering both Supercharger stations and residential power grids where cars are charged overnight. It's a self-reinforcing cycle. More cheap storage means more clean energy available, which reduces the operating cost of each electric vehicle in the fleet and makes the economic argument in favor of electrification even stronger for the end consumer. Looking at everything that has happened in recent weeks, there is a common thread running through every announcement, every engineering decision, and every agreement signed. The gigacasting exposed to the public was not an isolated marketing action. The chip factory in Austin is not a side project unconnected to the cars. The battery deal in Michigan doesn't exist solely to power stationary energy systems. Each piece of this puzzle was positioned with a specific destination in mind, and that destination has a name that Tesla hasn't officially pronounced yet, but that the entire industry already knows as Model 2, the approximately $25,000 car that could definitively change the electric vehicle market. The cost logic behind this vehicle depends on a combination of factors that need to work simultaneously. The unboxed manufacturing method, which builds different sections of the car in parallel before joining them, drastically reduces assembly time and the factory space required per unit produced. Gigacasting eliminates hundreds of welded parts and replaces them with single aluminum structures, cutting weight, line time, and potential for failure. LFP cells sourced from Michigan come in with a lower cost per kilowatt-hour than any alternative chemistry, are cobalt-free, have a long lifespan, and minimal thermal risk. And the internally manufactured chips, once Terafab is operational, remove the margin currently paid to external suppliers for each FSD unit installed. Alone, none of these elements solves the price problem. Together, they form the only known combination that makes an entry-level electric car profitable at scale. The timeline also starts to make will be operational in 2027. The AI 5 chip will begin mass production around the same time. The cybercab, which uses the same manufacturing base planned for the Model 2, is already in production at Gigafactory Texas in the second half of 2026. Each of these milestones was announced separately in different contexts with different justifications, but the dates overlap in a way that is difficult to attribute to coincidence. The company is building the conditions for a launch.