Terafab Revealed: Musk Ditches TSMC & Bets $25B on Intel 14A to Dominate AI

Kind: captions Language: en 22nd, 2026, during the Q1 earnings call, Elon Musk dropped a bombshell. Tesla has chosen Intel 14A, not TSMC, for the $25 billion Terafab project. Intel's stock immediately jumped 3.6% in after-hours trading. Why is Musk betting the future of his chips on Intel, a company that just reported a net loss of $3.7 billion in Q1? TSMC is still the king of the foundry industry and is still manufacturing the AI5 chip for Tesla. And yet, specifically for Terafab, Musk went in the opposite direction. What are the five real reasons behind this audacious move? Let's dive right in. >> [music] >> Elon Musk has announced a decision that sent shockwaves through the entire semiconductor industry. Tesla will use Intel's 14A chip manufacturing process, equivalent to a 1.4 nanometer feature size, for the Terafab project in Austin, Texas. And here is the important point. This is Intel's first major external customer for this technology. The market reacted instantly. Intel's stock rose 3.6% in after-hours trading. HSBC immediately upgraded its recommendation from hold to buy, raising its price target from $50 to $95, nearly doubling in a single night. So, what is Terafab? It is a $25 billion joint venture between Tesla, SpaceX, xAI, and now Intel. The ultimate goal sounds almost crazy. Produce 1 terawatt of computing power per year, twice the total computing capacity of the entire United States today. But here is the most interesting part. Just 1 week earlier, on April 15th, Tesla had completed the tapeout of its AI5 chip at both TSMC and Samsung. Meaning TSMC is still Tesla's primary chip manufacturing partner. So, why, specifically for Terafab, the most important and most expensive project in Tesla's history, did Musk turn to Intel? The answer has five reasons. And each one touches on problems you may never have thought about before. The first, and perhaps the most fundamental reason, does not lie in technology. It lies in the map of the world. Over 90% of the world's high-end chips are currently manufactured in Taiwan and South Korea. TSMC, the largest chip manufacturer on the planet, is headquartered in Taiwan. Even though they have built a fab in Arizona, even though the US has poured billions of dollars through the CHIPS Act, the technological heart of TSMC still beats in Taiwan. And this is the major problem. Tesla now is not just an electric vehicle company. Tesla is building Optimus, a humanoid robot. It is building Cybercab, a robo-taxi. It is building supercomputer clusters for FSD. SpaceX is building Starlink. xAI is building data centers in space. Every single one of those products needs chips, millions, billions of chips. If tensions in the Taiwan Strait escalate, even just one incident, Tesla's entire chip supply chain could freeze for months. Every AI ambition, every promise made to investors, could vanish in just a matter of weeks. Intel is a completely different story. Intel has factories in Ohio, Arizona, and Oregon, all on American soil. No worrying about transoceanic shipping, no worrying about geopolitical conflict, no worrying about sudden tariffs. So, this is not just about business. This is survival insurance. Musk himself said this during the Terafab presentation. Either we build Terafab or we don't have chips. And we need chips, so we're building Terafab. The statement sounds simple, but behind it lies an entire life or death strategy. And to build on American soil, the most natural partner is Intel. Second, and this is where the story gets truly dramatic. TSMC currently sits at the top of the world. Who are their VIP customers? Apple, Nvidia, AMD, Qualcomm, Broadcom. Every name on that list is a giant. And every company is competing for every wafer, every 2-nanometer production line. TSMC has committed over $165 billion in investment through 2028. Sounds enormous, right? And yet, it is still not enough. The AI hunger is devouring every unit of high-end chip manufacturing capacity on this planet. So, if you are Tesla, where do you stand in TSMC's priority queue? The straightforward answer, not number one. Apple orders hundreds of millions of A-series chips every year. Nvidia is consuming every bit of 2-nanometer capacity for AI GPUs. Tesla, no matter how large, is still the third, fourth, or even fifth customer in line. Now, look at Intel. The story reverses entirely. In Q1 2026, Intel reported a net loss of $3.7 billion on revenue of $13.6 billion. A loss, a heavy one. CEO Lip Bu Tan publicly declared, "If Intel cannot find a major external customer for its foundry business, the company will have to exit that business entirely." This is a life or death situation. Intel needs Tesla more than Tesla needs Intel. And what does that mean for Musk? It means Tesla will be customer number one, not customer number 10. It means Tesla gets priority access to the best production lines, the best engineers, the most flexible contract terms. It means when a technical problem arises, Intel's entire team will throw everything they have at solving it. Because if Tesla walks, Intel collapses. At TSMC, Musk has to stand in line behind Tim Cook. At Intel, Musk is Tim Cook. Third, at this point, you may be wondering a very reasonable question. Does Intel's 14A technology actually perform better than TSMC's 2 nanometers? Or is this just a something is better than nothing deal? The answer is very interesting, and it shows that Musk is playing a far longer game than most people think. TSMC's 2 nanometers is currently the gold standard of the industry. It already manufactures the AI5 for Tesla, mature, stable, efficient. This is the pinnacle of technology for to 2026. So, what is Intel 14A? It is a next-generation process node at 1.4 nanometers, approximately 15 to 20% more efficient than its predecessor, the 18A process. To put it simply, if 2 nanometers is today's Ferrari supercar, then 14A is the Ferrari 3 years from now. Not yet released, not yet fully refined, but definitively faster and more efficient. And here is the point that left many analysts speechless. Intel 14A is not expected to be truly ready until 2028 to 2029. So, why would Musk choose a technology not yet on the market instead of the already mature 2 nanometers? The Terafab timeline. Let's do the math together. Terafab just broke ground this past April. Building a high-end fab typically takes 2 to 3 years. Add the time for validation and production ramp-up, and Terafab will realistically be running at full capacity around 2028 to 2029. Exactly when 14A reaches maturity. Musk said this himself on the earnings call. By the time Terafab scales up, 14A will be ready for primetime, ready for mass production. It sounds simple, but this is an extraordinarily sophisticated strategic calculation. In other words, Musk is not choosing the best technology of today. He is choosing the best technology of tomorrow, timed precisely for when Terafab needs it most. This is the thinking of someone who is not building for next year. He is building for the entire decade ahead. Fourth, this part, to be honest, is the most dramatic in the entire story. Before coming to Intel, Musk knocked on Samsung's door first. And what happened next, no one anticipated. The Terafab team approached Samsung Electronics asking Samsung to help Tesla build its own dedicated chip factory. On paper, this was a completely reasonable proposal. Samsung has decades of fab experience, has 2 nanometer gate-all-around technology, and already has a factory in Taylor, Texas. Every condition seemed to align. But Samsung said no. They countered with a very cleverly worded alternative. Instead of helping Tesla build its own fab, Samsung invited Tesla to come and use their existing capacity in Taylor. On the surface, this was a polite offer. But beneath the surface, it was a straightforward refusal. Why did Samsung refuse? The business logic is simple to the point of being ruthless. If Samsung teaches Tesla how to build a fab, one day Tesla will be able to manufacture its own chips without needing Samsung anymore. They were being asked, in effect, to raise their own future competitor. No clear-headed CEO would ever agree to that. TSMC operates by the same principle. TSMC is the king of foundries precisely because they guard their know-how like buried treasure. They sell chips, not expertise. You can be a VIP customer of TSMC, but you will never be allowed to look inside their blood room. Then Intel entered the story. And Intel's position was completely different. Intel had nothing to lose. The foundry business was hemorrhaging billions of dollars. The stock had been in decline for years. So when Musk called, Intel did not ask, "What can we protect?" They asked, "What can we do to help?" Not only that, according to reports out of South Korea, Intel has been actively recruiting engineers with Samsung experience to support the Terafab project. They are throwing everything they have into this bet, including going after talent from the very competitor that turned Tesla away. That moment, when Samsung closed the door and Intel opened its arms, was the moment Musk knew he had found the right partner. Samsung was afraid of becoming a casualty. Intel accepted the role of genuine partner. That is the difference that decided everything. Fifth, and perhaps the most compelling reason. This is not just a routine business deal. These are two ships rescuing each other in open water. Look at the numbers after the announcement. Intel's stock rose 3.6% in after-hours trading. HSBC upgraded from hold to buy, raising its price target from $50 to $95, a nearly 90% increase. Over the past 30 days, INTC has risen 51.6%. Year-to-date in 2026, it is up nearly 70%. And over the full prior year, Intel's stock has gained more than 210%. This is not the market's reaction to an ordinary deal. This is the market's reaction to a resurrection. Investors are betting that Intel, after years of struggle, has finally found its way out. But Tesla benefits just as much, perhaps even more. With Terafab and Intel 14A, what does Tesla have in its hands? Complete control over the chip supply chain for 10 million Optimus robots per year by 2027. A bypass around the supply bottleneck the entire industry is forecasting for 2028 to 2029. Geopolitical insurance for Musk's entire empire. And most importantly, the right to co-design chips, meaning chips architecturally optimized for Tesla's own neural networks, not the one-size-fits-all approach that comes with buying off the shelf from TSMC. Analyst Jay Goldberg of Seaport Research summed it up precisely. Having the customer matters more than the timing. And Barclays wrote in their report, "Chips will be Tesla's next growth pillar, not vehicles, not batteries, not Optimus as a standalone product, silicon." However, in the interest of fairness and credibility, the counterargument must be presented because not everyone believes in this bet. Expert John Petty issued a warning worth serious consideration. Across the entire modern semiconductor era, not a single newcomer has ever successfully built a world-class foundry from scratch. Amazon tried. Google tried. Both retreated after confronting the terrifying complexity of this industry. And there are troubling numbers. Tesla's CapEx for 2026 exceeds $25 billion. Meanwhile, GAAP net margin in Q1 stood at a razor-thin 2.1%. Tesla produced 50,000 more vehicles than it sold last quarter, a sign that demand is plateauing. So the existential question here is, will the market have enough patience to let Musk burn cash over the next 3 to 4 years waiting for Terafab to bear fruit? After examining all five reasons, I want to share my personal perspective directly. Musk's decision to go with Intel 14A is not an emotional one. It is a move calculated across at least three distinct layers. In the short term, Intel needs customers, so it is ready to offer the best possible terms. In the medium term, the 14A technology aligns perfectly with Terafab's construction timeline. And in the long term, this is a pivot to rebuild the entire chip industry order with Tesla as a new player standing on equal footing with Apple and Nvidia. But make no mistake, this is still a gamble. Intel has a history of delays with its 18A process. Musk's own track record includes projects with significant deadline overruns. The Cybertruck was 2 years late. Unsupervised FSD has been delayed by more than 6 years. Nothing guarantees Terafab will go according to plan. However, there is one very important distinction worth noting. When Musk places a bet, he tends to put his entire fortune behind it. The Gigafactory for batteries, a resounding success. Starlink for satellite internet, currently reshaping the entire telecom industry. Reusable Falcon 9 already rewrote the rules of aerospace. Terafab could become the fourth successful project on that list. Or it could become Silicon Valley's most expensive case study in failure. You do not know. Even Musk cannot be 100% certain. >> [clears throat] >> But one thing is clear. Over the next 3 years, the semiconductor industry will witness one of the largest battles in the history of technology. And Tesla has already chosen its side. million people. That is the number of workers worldwide that McKinsey projects will be forced to change occupations entirely before 2030. Not because of an economic crisis, but because of robots. This coming July, Tesla is launching Optimus V3. 10,000 components, hands with 22 degrees of freedom, capable of working 12 continuous hours at a cost of 51 cents per hour. Meanwhile, a single American worker costs $46 per hour. Could this be the moment that changes labor permanently? And Tesla has set a target of 1 million robots in the very first year. What does that number actually mean? Let's dive right in. Before getting into the analysis, I want everyone to have a clear and solid grasp of the broader picture that is unfolding because this story has far more layers to it than what you see on the surface. On April 22nd, 2026, during the Q1 earnings call, Elon Musk officially confirmed what the entire technology industry had been waiting for. Optimus V3 will begin production at the Fremont factory in late July or early August of this year. Not a test run, not a prototype, real production on a real assembly line. And in order to make that happen, Tesla had to make a decision that many in the industry are calling the boldest in the company's history. The entire Model S and Model X production line at Fremont, a line that had been running for 14 years, was completely dismantled. Every machine, every wiring system, every welding robot, all torn down to make room for Optimus. The time to do all of that? 4 months. And the target for this new production line is 1 million robots per year. To help everyone get a clearer picture, here is the timeline to keep in mind. May 2026, the Model S and Model X officially cease production after 14 years. July through August 2026, Optimus V3 launches and production begins. By the end of 2026, Tesla is targeting 50,000 to 100,000 robots. And by 2027, the Giga Texas factory will be operating at a capacity of 10 million robots per year. Looking at that timeline, one question becomes impossible not to ask. What exactly is Tesla betting on? When many people first heard the news that Tesla was discontinuing the Model S and Model X, they assumed it was a sign the company was weakening. But the reality is precisely the opposite. And when you look at the numbers, everything becomes much clearer. The Model S and Model X currently account for only about 30,000 vehicles per year, while the Fremont line has a capacity of up to 100,000 vehicles. That means 70% of its capacity is being wasted every single day. Meanwhile, that same production line, if converted to Optimus, could generate far greater revenue. This is not Tesla abandoning automobiles. This is Tesla shifting resources away from what is slowing down toward what is exploding. But why is Optimus so compelling? The answer lies in a set of calculations that, the first time I read them, I genuinely had to stop for a moment before I could believe what I was reading. An American manufacturing worker costs an average of $46 per hour, factoring in wages, insurance, paid time off, and worker protections. A human being works approximately 2,080 hours per year. And every 2 to 3 years, an average of 39% of manufacturing workers leave their jobs, forcing companies to spend additional money recruiting and training replacements. When all of those figures are factored in, the actual cost of a single worker position can reach $180,000 to $220,000 per year. Optimus V3 operates at a cost of 51 cents per hour. It works 8,760 hours per year, four times that of a human. No vacation days, no strikes, no sick days. And Tesla calculates that each Optimus robot can save $57,550 per year compared to hiring a human worker for the same job. If Tesla deploys 50,000 robots by the end of 2026, the total savings figure comes to nearly $3 billion per year in labor costs alone. And that is only the first year. This is why Elon Musk has said that Optimus could become a more important product than Tesla's entire automotive division combined. And this is why he is willing to retire two iconic models tied to the company's history, because the economic logic is too clear to ignore. But numbers that look good on paper are one thing. Whether Optimus V3 can actually deliver on what Tesla is promising is a more important question. And that is exactly where I am going right now. This is the part I think will surprise most viewers the most, because V3 is not simply a minor upgrade. It is a product built from the ground up with a completely different design philosophy. Optimus V2, the version many people have seen in demo videos, had basic hands, moved fairly rigidly, and was used primarily for demonstrations. It was impressive to watch, but not yet capable of doing real work in an industrial manufacturing environment. Tesla knew that. And they spent all of 2025 redesigning V3 with a single goal, real work, mass production, and a design that cannot be easily replicated. V3's hands have 22 degrees of freedom, nearly equivalent to a human hand. The way Tesla achieved this is technically interesting. They relocated the entire heavy actuator system up into the forearm, then used an artificial tendon system to control each individual finger remotely. The result is a hand that is lighter, more agile, and capable of gripping gently enough not to crack thin plastic components, while also strong enough to handle heavy metal parts. Built into it are tactile sensors. The robot can detect the amount of force being applied at each finger and adjust in real time. This is something that many more expensive industrial robots still have not been able to do. V3's vision system features eight integrated cameras working in conjunction with an AI architecture Tesla calls FSDV15, which the company refers to as physical AI. And this is the point I find most interesting from a design philosophy standpoint. Tesla sees Optimus V3 as a Tesla vehicle. But instead of wheels, it has two legs. Instead of a steering wheel, it has two arms. All of the machine learning data accumulated from millions of miles of Tesla vehicles driving in real-world conditions is being translated into the robot's ability to process and interact with the physical world. This is an advantage that no other robotics company has, because no one else has millions of vehicles out there collecting real-world data every single day. And there is one strategic detail that few people have noticed. Tesla spent more than a year filming actual workers doing their jobs on the floors of the Fremont and Giga Texas factories. Every motion of picking up a component, every gesture in an assembly sequence, every way of handling an unexpected situation on the line, all of it was recorded to train V3. The robot learns by observing humans, the same way a new employee learns from someone more experienced. The only difference is that the robot never forgets, never gets tired, and immediately shares what it has learned with the entire rest of the robot fleet. But perhaps the most strategically impressive decision Tesla made was not in hardware or software. It was in how they handled information. Tesla deliberately kept the entire V3 design secret all the way up until the moment production began. The reason was entirely practical. During the V1 and V2 launches, competitors had analyzed every frame of released footage, measuring joint angles, calculating component ratios, and attempting to reverse engineer the designs. This time, Tesla completely reversed that process. By the time competitors saw V3 for the first time, Tesla was already in production. The time for any response was nearly zero. This is the strategy of a company that knows it is in the lead and wants to hold that gap for as long as possible. Currently, more than 1,000 Optimus Gen 3 robots are operating inside Tesla's Fremont and Giga Texas factories. They're performing three primary categories of tasks, moving 4,680 battery cells between assembly stations, sorting and delivering components on demand, and conducting quality inspections using AI camera systems. Each shift runs 12 hours continuously without breaks. I want to be straightforward about something here, because I believe honesty matters more than only saying things that sound impressive. Those tasks are still relatively simple compared to the full scope of a manufacturing process. The robots are not yet capable of handling steps that require complex judgment or flexible responses to unexpected situations. Tesla acknowledges this as well. They will start with simpler skills and build upward from there. But the reality is that even those simpler tasks, when executed by thousands of robots running 24 hours a day, already produce an enormous economic impact. And that leads directly into the biggest question this video set out to answer. The McKinsey Global Institute, one of the most respected economic research organizations in the world, estimates that automation could eliminate between 400 and 800 million jobs globally by 2030. That figure is equivalent to 375 million people, approximately 14% of the entire global workforce being forced to transition into completely different occupations. Goldman Sachs projects the humanoid robotics market will reach $38 billion by 2035, a six-fold increase compared to estimates made just a few years earlier. These numbers are not science fiction. They are economic forecasts produced by the world's largest financial institutions. Organizations that have no reason to either exaggerate or downplay what the data shows. But, and this is something I think is important to say, history has seen similar concerns before and the outcomes were never as straightforward as people assumed. When ATMs arrived in the 1970s and 1980s, the world feared that bank tellers would disappear. What actually happened was the complete opposite. The number of bank tellers grew from 300,000 to 600,000 over the following 40 years. Because ATMs made it cheaper to open branches and more branches needed more people to serve customers. But the story does not end there. >> [clears throat] >> After 2010, when mobile banking, digital wallets, and AI genuinely matured, the number of bank teller positions began to actually decline. And there has been no sign of that stopping. Automation does not kill jobs immediately. It creates an extended transitional period that can last for decades and then the real far-reaching impact finally arrives. The real question is not whether robots will replace humans. The real question is when that wave arrives, will we have enough time to adapt? I want to look honestly at what robots still cannot do because fairness demands it. And because this is what creates genuine value for viewers rather than simply presenting one side. No robot currently operates at the speed and reliability of a conventional automotive assembly line. Traditional fixed industrial robots, the kind that have been in factories for 30 years, are still faster, more precise, and more reliable than humanoid robots for fixed high repetition assembly tasks. Optimus V3 is currently performing well in tasks involving movement, sorting, and material transport. But it cannot yet fully replace a skilled worker in operations requiring advanced skill and flexible judgment. And this is something Elon Musk himself acknowledged openly during the Q1 2026 earnings call. He said that initial output would be very slow and that predicting specific numbers was not possible because this is an entirely new supply chain, entirely new technology with nothing existing to reference. Even the person who created it does not know exactly how much it will produce in the first year. That does not mean Tesla is failing. It means that for the first time in history, a company is attempting to mass produce humanoid robots from zero. No roadmap, no precedent, no pre-existing supply chain, only data collected day by day, pressure from investors, and a July deadline closing in fast. And in that context, 4 months to completely dismantle and rebuild a factory is something Musk himself called insane. Not to boast, but because he understands better than anyone the level of risk this decision carries. But Tesla is not the only one running in this race. And this is a part I think needs to be addressed because the full picture is far more interesting than looking at Tesla alone. Boston Dynamics recently unveiled the production version of Atlas at CES in early 2026. And the entire production allocation for this year sold out on the day of the announcement. Units are being delivered to Hyundai and Google DeepMind. Their factory target is 30,000 robots per year by 2028. Atlas has 56 degrees of freedom, more than the Optimus V3. It can lift up to 50 kg. And it has one very practical feature that most people overlook. Atlas can swap its own battery without human assistance, maintaining continuous operation without stopping the production line. But here is what creates the fundamental difference between the two companies. Atlas is not targeting the mass market. Atlas is targeting only high-end industrial factories at an estimated price that could exceed $150,000 per robot, five to seven times higher than Optimus V3's target price. Tesla wants a $20,000 to $30,000 robot for everyone, every factory, every industry. Boston Dynamics wants the best robot for customers willing to pay the highest price. Two different philosophies, but both pointing toward a future where the question is no longer whether robots will enter factories, but how many robots, doing what jobs, and who controls them. And that is precisely what makes July a milestone worth watching. Not only for those who follow technology, but for anyone who goes to work every day and finds themselves wondering what the future of their job is going to look like. July 2026 will be here very soon. And when Optimus V3 officially launches, it will be one of those milestones that people are still talking about many years later. The question is not whether it will happen. The question is whether you understand it well enough to make the right decisions for yourself. That is why Tech Revolution exists, to update news and deliver analysis in the most accessible way possible. For Tesla Semi saves $40,000 per year per truck compared to diesel. Real-world figures from fleets that have driven 13.5 million miles. PepsiCo is expanding to 50 units. Walmart just received their trucks. DHL has placed numerous new orders for 2026. Meanwhile, diesel in California just hit $7 per gallon, an all-time historic record. So, why is this $40,000 figure enough to force billion-dollar corporations to abandon diesel right now? Today, let's explore every number in the Tesla Semi versus Diesel Showdown. Let's dive right in. First, we need to understand a foundational truth. The heavy-duty trucking industry in the United States consumes approximately 45 billion gallons of diesel every year. That is an enormous number, meaning hundreds of billions of dollars flowing into fossil fuels. But in 2026, for the first time in history, an electric truck has proven that it can break that cycle. Not with promises, but with real-world numbers on the road. And the battle begins with the thing that matters most to every business, money. An average eight diesel truck gets about 6.5 miles per gallon. That sounds decent enough until you look at the bill. In California right now, diesel is at $7 per gallon, the highest ever recorded in history. That works out to between 48 cents and 80 cents per mile in fuel costs alone. On a 400-mile trip, that is $190 to $320 gone every day, every truck. And diesel prices are something no one can control. They dance with global oil prices, with wars, with OPEC decisions, meaning businesses virtually cannot know in advance what their operating costs next quarter will be. Now, look at the Tesla Semi. In a real-world test in Texas with commercial electricity at 9.12 cents per kilowatt-hour, the Tesla Semi ran at 1.64 kilowatt-hours per mile. That comes out to just 15 cents per mile compared to 48 cents for diesel, a nearly 70% reduction in fuel costs immediately. No negotiation needed, no waiting needed. But the more important number is right here. Electricity prices are far more stable than diesel. Businesses can sign long-term electricity contracts and know exactly what their costs are each month. No geopolitical volatility, no OPEC, no surprises. And if any fleet installs solar panels at their own depot, charging costs drop even further. Put the two numbers side by side, 15 cents versus 70 cents per mile. Over 100,000 miles per year, that is a difference of $55,000 in fuel costs alone for one truck. Multiply that by 50 trucks, 100 trucks, that number becomes millions of dollars. That is why Pepsico and Walmart are racing to convert. Most of the answer lies right in this simple calculation. But fuel is only the tip of the iceberg. Because there is a cost that silently eats into the profits of every fleet operator that many people outside the industry do not see. A class 8 diesel truck is an extremely complex machine. Turbochargers, multi-speed transmissions, exhaust treatment systems, high-pressure fuel pumps, diesel exhaust fluid, diesel particulate filters. Each component is a point that can fail at any time. Average maintenance costs fall in the range of 15 cents to 20 cents per mile. And every time a truck goes into the garage, it is not running. Not running means not making money. Average uptime for diesel fleets typically reaches only about 85% meaning out of every 100 days, 15 days the truck sits waiting for repairs. There is one real-world detail that is very noteworthy. A fleet operator once reported that they needed five mechanics working continuously just to keep 40 diesel trucks running. That is an enormous labor cost that few people think about when talking about trucks. The Tesla Semi is completely different. Electric motors are by nature far simpler than diesel. No turbochargers, no complex transmissions, no oil changes, no DEF, no DPF. The regenerative braking system significantly reduces brake pad wear. According to industry estimates, maintenance drops by 30 to 40% over the vehicle's lifetime with costs of only about 7 cents to 10 cents per mile. And here is the most impressive number. 95% uptime. Out of every 100 days, the truck runs 95 days compared to 85 days for diesel. Those 10 days of difference multiplied by revenue per day, multiplied by the number of trucks in a fleet, that is a sum of money that no fleet manager can ignore. That same fleet when switching to electric only needed one mechanic instead of five. Fewer parts, fewer breakdowns, fewer repairs, less cost. Simple logic, but the impact is very large. Of course, at this point some will say that cheaper operating costs are true, but the purchase price of a Tesla Semi is much higher than diesel. So, is it really worth it? The Tesla Semi long range is priced at approximately $290,000. The standard range approximately $260,000. Meanwhile, a new class 8 diesel truck costs around $175,000. Looking at the upfront purchase price, diesel wins clearly. A difference of $85,000 to $115,000. The US government is actively supporting the transition. The federal tax credit for zero-emission trucks goes up to $40,000 per vehicle. In California, the HVIP program provides subsidies of up to $120,000 per truck. Combined, incentives can bring the effective price of a Tesla Semi down to approximately $130,000 to $150,000. Lower than a brand new diesel truck. And this is not a program on paper. HVIP data shows that approximately 892 vouchers have been requested or issued for the Tesla Semi accounting for over 80% of all class 8 battery electric vouchers. To receive a voucher, a fleet must have a real reservation, real money, a real commitment, not just signing up for fun. Now, when you add up fuel costs, maintenance, and uptime over a lifetime of 1 million miles, the math tilts decidedly toward the Tesla Semi. Take a specific example from Pepsico. On the 425-mile Frito-Lay route, the Tesla Semi saves over $200 per trip compared to diesel. That sounds small, but multiply it over a lifetime of 1 million miles, approximately 7 years of operation, and each Tesla Semi saves an estimated approximately $480,000. Nearly half a million dollars for a single truck. Multiply $480,000 by 100 trucks in Pepsico's fleet, that is $48 million. With 1,000 trucks, that is nearly half a billion dollars. That is why the CFOs of these corporations do not need anyone to convince them. The numbers say it all. However, to be fair, not every fleet benefits equally. Electricity costs vary by region. 9 cents per kilowatt-hour in Texas is very different from 20 cents per kilowatt-hour in some other states. Fleets running fixed short routes will save the most. While fleets running long-haul interstate routes across multiple states still face the charging infrastructure problem. The $40,000 per year figure is real, but it depends on the specific conditions of each fleet. This is an important perspective for us to maintain fairness in the analysis. So, the numbers on paper are clear. But the real question is, on the road, in the rain, hauling heavy loads, running long distances, does the Tesla Semi actually perform that well? 13.5 million real-world miles will answer that question. Pepsico is the name that cannot be skipped when talking about the Tesla Semi. They were the first customer to receive trucks in December 2022, a time when nearly the entire trucking industry was still skeptical about whether an electric truck could haul heavy cargo over long distances. Three years later, the answer is clear. Pepsico did not just keep their original fleet. They expanded to 50 Tesla Semis in California. From five trial units to 50 commercially deployed trucks. That is not the action of a company testing for fun. That is the action of a corporation that has seen real savings and decided to bet bigger. And there is one detail that I consider very noteworthy. According to Elon Musk, Pepsico drivers, after driving the Tesla Semi, did not want to go back to diesel. You have to understand that professional truck drivers are the group most resistant to new technology. They have driven diesel their entire lives, accustomed to every engine sound, every gear shift. When these very people say they do not want to go back, that is a signal stronger than any test on paper. DHL, one of the largest logistics corporations in the world, received their first Tesla Semi in December 2025 in Livermore, California. The truck currently runs about 100 miles per day, only needs to charge once per week, achieves a range of up to 500 miles at full load, and reduces approximately 50 tons of CO2 per year. DHL currently operates more than 150 class 8 battery electric vehicles across the United States, and has placed additional Tesla Semi orders for delivery in the second half of 2026 with plans to expand to Ohio and Pennsylvania. The target by 2030 is for 66% of DHL's last-mile fleet to be zero emission. Then, there is Road One Intermodal Logistics, a name few people know, but whose story says a great deal. Road One is not a startup chasing trends. They have been Tesla's aluminum shipping partner since 2012, 13 years of continuous collaboration. When a company that has worked with you for 13 years decides to buy your product with their own money, that is not marketing. That is a pure business decision based on real-world data. Road One's Tesla Semi runs the Oakland to Fremont route hauling aluminum averaging 38,000 lb, achieving an efficiency of 1.9 kWh per mile, accelerating to 60 mph in 20 seconds even at full load. And Road One has already planned to expand from one to 10 units. ArcBest ABF Freight also completed a 3-week pilot covering 4,494 miles at 1.55 kWh per mile, far exceeding Tesla's original promise of under 2 kWh. DHL achieved 1.72 kWh per mile on a 390 mile route with a total weight of 75,000 pounds. In total, several hundred Tesla semis are operating on the road having accumulated 13.5 million miles. A single truck has surpassed 440,000 miles and the entire fleet maintains 95% uptime. This is no longer a prototype on display. These are commercial vehicles running daily hauling real freight on real roads. But if the truck is that good, then why are there only a few hundred? Why not a few thousand? The answer lies in production capacity. Until early 2026, Tesla did not have a large-scale semi production facility. Every truck before that was either pre-production or a pilot build. But that has changed. Tesla's dedicated semi factory in Sparks, Nevada officially opened in early 2026. Spanning 1.7 million square feet located right next to Gigafactory Nevada where 4680 battery cells are produced. This location is not a coincidence. When the vehicle factory sits right next to the battery factory, the entire supply chain is shortened to nearly zero. This was the exact bottleneck that had delayed the semi program for years and Tesla has finally solved it. Volume production began in March with a target of mass production before the end of June. In the Q1 2026 earnings call held on April 22nd, just 2 days ago, Tesla confirmed semi production remains on schedule. Though it will be slow initially due to the new supply chain, but the company expects exponential growth by year-end. The long-term target is 50,000 trucks per year and at full ramp, this factory alone could generate potential revenue of over 13 billion dollars per year for Tesla. However, this needs to be said directly. Between the target of 50,000 trucks per year and reality is a large gap. Estimates suggest that in the second half of 2026, Tesla may deliver approximately 5,000 to 10,000 units, far more modest than the ambition. Elon Musk is famous for setting overly optimistic timelines. The Tesla Semi was announced in 2017 with a promise of production starting in 2019. But mass production did not truly begin until 2026, nearly 9 years behind schedule. That is a fact that needs to be acknowledged. But it also needs to be fairly noted that the truck has proven itself on the road. The factory is open and this time there is concrete evidence, not just promises. Now to a question that many people are certainly thinking. Electric trucks are good. The savings are clear. But when the battery runs out, where do you charge in the middle of the road? This is the biggest counter-argument the diesel side always uses and honestly, until recently, they had a point. Diesel stations are present on every corner of America. An advantage accumulated over more than 100 years. But Tesla is building the answer. And the pace of deployment is accelerating rapidly. Tesla just updated its Megacharger map with 64 new locations across 15 states bringing the total to 66 sites. Texas leads with 19. California follows with 17. The target is to complete 37 stations in and 46 stations before the start of 2027. Each Megacharger provides power of up to 1.2 megawatts, dozens of times more than a standard Supercharger. At that speed, the Tesla Semi can charge enough range for several hundred miles in about 30 minutes, exactly the mandatory rest time for drivers under federal law. And here is the strategic move that I consider the smartest. Tesla has signed a partnership with Pilot Travel Centers, the operator of more than 800 truck stops across North America. The first station is expected to open in summer 2026. Instead of building a network from zero, Tesla is leveraging existing infrastructure from the industry's largest partner where there are already parking lots, amenities, and restaurants for drivers. Each station will have four to eight stalls integrated with Megapacks to buffer peak load. In addition, approximately 20 dedicated Tesla Semi service centers are being deployed across the United States. Looking more broadly, this is something that few competitors can replicate. Tesla is not just selling trucks. They are building an entire ecosystem consisting of vehicles, batteries, charging stations, maintenance services, and fleet management software. Volvo does not build fueling stations. Daimler does not manufacture batteries. ExxonMobil does not design trucks. But Tesla does all of it. That is the competitive advantage that no company in the traditional trucking industry can recreate in the short term. It must be frankly acknowledged that 66 Megacharger stations are still far too few compared to tens of thousands of diesel stations across the United States. Current charging infrastructure is mainly concentrated in California, Texas, and some coastal states. Inland states are still virtually blank. This means the Tesla Semi is currently best suited for fixed regional routes, not yet a comprehensive solution for long-haul cross-country operations. That is a real-world limitation that any fleet manager needs to factor in. At this point, we have gone through the entire showdown. On fuel, Tesla Semi wins with a 70% difference. On maintenance, it wins with 95% uptime. On total cost of ownership, it saves nearly half a million dollars per truck over its lifetime. On real-world evidence, 13.5 million miles from Pepsico, DHL, Road One, ArcBest. On infrastructure, it is being built fast but remains the biggest challenge. And there is one final card that Elon Musk is still holding. Something that could turn the Tesla Semi from saving $40,000 per year into saving $140,000 per year. Autonomous driving. Full self-driving. Right in the Q1 2026 earnings call, Musk confirmed that the unsupervised robo-taxi service has launched in Dallas and Houston with not a single accident or injury recorded. FSD version V14.3 launched in April with a new advanced architecture. If self-driving technology is perfected on passenger cars and applied to the Semi, the driver cost of approximately $1 per mile will disappear. Over 100,000 miles per year, that is $100,000 in additional savings. Combined with $40,000 in fuel and maintenance savings totaling $140,000 per truck per year. The truck runs 24/7. No rest, no wages, no health insurance. If that happens, the Tesla Semi will not just be a transport vehicle. It will become a mobile money-printing machine on the highway. But it must be emphasized that autonomous trucking is still the future, not the present. Regulatory, technical, and social barriers remain very large. Do not base today's purchasing decision on tomorrow's self-driving promise. Base it on the numbers that have been proven. $40,000 in savings per year. 95% uptime. 13.5 million real-world miles. That is already enough. So we have together gone through the entire showdown between Tesla Semi and diesel. From fuel costs, maintenance, total cost of ownership, to real-world evidence from Pepsico, DHL, Walmart, and the Megacharger infrastructure. Diesel is not dead today. But the $40,000 in savings per year, 95% uptime, and 13.5 million real-world miles have clearly shown the direction of the trucking industry. Dakota Shearer, a 30-year driver, hauled 11 tons through a mountain pass in a Tesla Semi and said, "The cargo feels like it doesn't exist." Pepsi drivers, after the pilot program, voluntarily refused to go back to diesel. Tom Sterba declared it straight out, "I want to retire in this truck." What is it about this vehicle that made even the most hard to convince people in the transportation industry change completely? In April 2026, the Tesla Semi entered mass production. 800 kilowatts of power, fuel costs 70% lower, 95% uptime after 13.5 million real-world miles. What does diesel have left to compete with? Let's dive right in. Diesel trucks are not outdated technology. They have been the backbone of the American economy for the past 100 years. 70% of all goods in the US, from food and medicine to electronic components, all move through these diesel trucks. And the drivers who operate them? They are not the type of people easily persuaded by a grand launch event or a polished advertisement. They live by their truck. If the truck breaks down on the road, they lose money. If the truck burns more fuel than expected, they lose money on the run. For them, every decision must be based on reality, not promises. And yet Tesla dared to step in here. In 2017, Elon Musk stood on stage and announced the Tesla Semi. The trucking industry's reaction at that time? Complete skepticism. Many experts stated plainly that electric batteries could never replace diesel in heavy-duty freight hauling because the battery weight was too great, charging time was too long, and charging infrastructure was nearly non-existent. And Tesla delayed. No deliveries in 2019, not in 2020, not in 2021. It was not until late 2022 that Tesla managed to deliver the first four units to Pepsi after five consecutive years of delays. At that point, the market looked at the Tesla Semi as a promise not yet kept. But that was exactly when the real story began. And that story changed everything. The number that makes every diesel driver stop. I'll put out one number, and I want you to truly feel the weight of it. A fully loaded diesel truck, weighing around 37 tons, needs more than 60 seconds to accelerate to highway speed. The Tesla Semi does it in 20 seconds with that same 37-ton weight. Why does this number matter? Because in reality, acceleration rate directly affects safety when merging onto the highway, vehicle control on grades, and the mental fatigue of the driver after many consecutive hours of driving. A vehicle with slow response demands that the driver calculate and stay tense without stop. The Tesla Semi uses three independent electric motors on the rear axle, producing 800 kilowatts, equivalent to roughly 1,072 horsepower. Not 1,072 horsepower in the sense of a nice number on paper, but instant electric torque from the very first second the accelerator is pressed, with no need to wait for the engine to build RPM the way diesel does. But this is what truly surprised me when researching this vehicle. The Tesla Semi's drag coefficient is 0.36. The Bugatti Chiron, one of the most expensive and sophisticated supercars on the planet, priced at over $3 million, has a drag coefficient of 0.38. The Tesla Semi, a 40-ton truck, cuts through air better than a Bugatti. Franz von Holzhausen, Tesla's chief designer, confirmed this was not accidental. This was a deliberate design objective from day one. The Semi's entire shape was optimized to reduce drag, cut energy consumption, and extend operating range. The result? Energy cost of just 17 cents per mile, while diesel costs between 50 cents and 70 cents for the same distance. What does diesel have left to counter that number? From 2022 to 2025, Tesla made no noise. They quietly placed a few hundred Semi units into real-world pilot programs with Pepsi, Saia Freight, and several other trucking companies. And those trucks ran. A total of 13.5 million miles were accumulated during the pilot phase. A single truck surpassed the 440,000 mile mark, equivalent to circling the earth nearly 18 times. Uptime reached 95%. 80% of incidents were resolved within 24 hours, half of those in under 1 hour. This is no longer theory. This is real data from daily commercial operations. And in March 2026, Tesla began volume production at a dedicated facility right next to Gigafactory Nevada, 158,000 square meters. In April 2026, mass production officially launched. The target is 50,000 trucks per year at full capacity with Tigress Financial Partners projecting deliveries of 5,000 to 15,000 units in 2026. After nearly a decade, the Tesla Semi is no longer a promise. It is rolling off the production line every day. Now, I want to talk about what I consider the biggest change this vehicle brings. Not the numbers, not the speed, but the experience of the person sitting inside it. Throughout 100 years of commercial trucking history, drivers have always sat offset to one side, left or right, depending on the market. This creates an inherent blind spot on the opposite side. When backing into a tight dock, drivers have to estimate, climb out to check, then climb back in. This is a procedure repeated dozens of times per day, every week, every year, throughout an entire career. The Tesla Semi breaks completely with that convention. The driver sits at the center of the cab, a position that has never appeared on any class 8 truck in history. On both sides are screens displaying a panoramic view around the vehicle, covering both the tractor and the trailer behind. No blind spots, no cli