Tesorb Signal: Tesla, SpaceX, and the Musk Companies

The next-generation Tesla Roadster was unveiled as a prototype in November 2017 with production promised for 2020. Nine years and at least eight broken deadlines later, the car still doesn’t exist. We walk through the full timeline of delays, the escalating performance claims, and what the Roadster saga reveals about Tesla’s pattern of promising big and delivering late.

That was almost nine years ago. And the car still doesn’t exist.

This is Tesorb Signal and I’m Lena Ruiz. Today we’re talking about the Tesla Roadster, the most delayed product in Tesla history, and what it tells us about how Musk makes promises.

Let’s start with what made the original Roadster so important. Back in 2008, Tesla was a tiny startup and the Roadster was its first car. Built on a Lotus chassis with a lithium-ion battery pack, it proved that electric vehicles could be fast, beautiful, and fun to drive. Tesla sold about twenty-four hundred of them. That car changed the entire auto industry.

So when Musk unveiled a next-generation Roadster in 2017, it felt like a full-circle moment. The specs were outrageous. Zero to sixty in 1.9 seconds, which would have been the fastest production car ever. A 200 kilowatt-hour battery. Six hundred twenty miles of range.

And a base price of $200,000, with a limited Founders Series at $250,000.

Tesla started taking deposits immediately. $50,000 for a standard reservation. A quarter million for one of the first thousand cars. People paid.

Then the delays started. Musk said 2020. Then he pushed it to 2021. Then to 2022 and 2023. Then in February 2024, he promised a production version unveil by the end of that year with deliveries starting in early 2025. That didn’t happen either.

By November 2025, at a shareholder meeting, Musk admitted production wouldn’t begin until 2027 or 2028. He set a demo date for April 1, 2026. And he openly acknowledged that April Fools’ Day gave him, quote, “deniability because I can say I was just kidding.”

That’s the CEO of a publicly traded company joking about the credibility of his own product timeline. For a car that people put down real money on nearly a decade ago.

But this is where the story shifts. In March 2026, Musk changed his language from “demo” to “unveil” and said the reveal would happen in late April. He called it “a banger next-level.” Tesla filed new trademark applications with an updated vehicle silhouette. They filed patents for a single-piece composite seat. Signs of actual progress.

Then on April 22, during the Q1 2026 earnings call, Musk pushed it again. Late April became “maybe in a month or so.” That’s the eighth time the timeline has moved. In nine years.

And during all those delays, the performance claims kept getting more extreme. The original promise was 1.9 seconds zero to sixty. Now Musk says it will do zero to sixty in under one second. He’s talked about SpaceX cold gas thrusters that would let the car briefly hover. Literally. Hovering.

At some point, the Roadster stopped being a car in development and became a hype machine. The claims get bigger every time the delivery date slips further away.

Meanwhile, competitors shipped real products. Rimac, a Croatian automaker, and Lotus both launched their own electric hypercars. Chinese automakers like BYD and Xiaomi have achieved performance records with production electric vehicles. The world didn’t wait for Tesla.

But does any of this matter beyond the Roadster itself?

I think it does. Because the Roadster isn’t the only product where Tesla has made enormous promises and asked people to be patient. The Cybertruck was announced in 2019 for production in 2021. Volume production didn’t start until years later, and the final product was very different from what was promised.

The Cybercab was unveiled in 2024 and is just now entering early production. Full Self-Driving, which we talked about in episode six, has been “coming next year” for nearly a decade.

The pattern is consistent. Announce something incredible. Collect deposits or build hype. Push the timeline. Escalate the promises to keep attention. Eventually deliver something, but later and different than what was originally described.

For the Roadster specifically, there’s a human cost. Thousands of people put down $50,000 to $250,000 starting in 2017. Those deposits have been sitting with Tesla for almost nine years. Some of those depositors may have passed away. Others may have simply moved on. Tesla hasn’t published how many reservations remain active.

Put it another way. If you’d put that deposit money toward a house in 2017, you’d have been living in it for years by now. These people are still waiting for a car.

So what’s the real takeaway here? If the Roadster unveil actually happens next month, it will come nine years after the prototype and at least seven years behind schedule. And it will still just be an unveil, not a delivery. Production is targeted for 2027 or 2028. The first depositors could be waiting a full decade for their car.

The original Roadster proved that electric cars could be exciting. This one is proving something else entirely. It’s proving that there’s no promise too big to make and no deadline too firm to break. As long as the hype keeps people looking forward instead of looking back.

That’s the signal. I’m Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk’s companies, visit our website at tesorb.com.

Got a tip or feedback? Send it to [email protected], we’d love to hear from you. This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

Scripts are researched and drafted in collaboration with AI assistants and reviewed before publishing. Corrections are posted to the episode page.

Tesla quietly disclosed a $2 billion acquisition of an unnamed AI hardware company, buried in a single sentence in its Q1 2026 10-Q filing. The deal was never mentioned on the earnings call. We break down what the filing actually says, what the deal structure reveals about the target, and why Tesla’s AI spending now dwarfs its car profits.

This is Tesorb Signal and I’m Lena Ruiz. Today we’re unpacking what might be Tesla’s most important acquisition ever, and why they didn’t want anyone to notice.

Here’s what happened. On April twenty-second, Tesla released its first quarter 2026 earnings. Revenue up sixteen percent. Earnings beat expectations. Elon Musk talked about robotaxis, the Cybercab, and the Roadster. The usual.

But two days later, when reporters started reading Tesla’s 10-Q filing line by line, they found something buried in Note fourteen. The very last note. Under a section called Subsequent Events.

One sentence. Here’s what it said. In April 2026, Tesla entered into an agreement to acquire an AI hardware company for up to two billion dollars in Tesla common stock and equity awards.

That’s it. No company name. No description of what the company does. No explanation of how it fits into Tesla’s strategy. Just a two billion dollar deal, tucked behind the financial statements, behind the legal disclosures, behind everything else.

Now, the deal structure tells us a few things. Only about two hundred million dollars is guaranteed upfront. The remaining one point eight billion is tied to performance milestones and something the filing calls “successful deployment of the company’s technology.”

That language is important. Successful deployment means this technology hasn’t been deployed yet. Tesla isn’t buying a proven product. They’re buying a team and a bet on something that doesn’t fully exist at scale.

And they’re paying entirely in stock. Not cash. Tesla is sitting on nearly forty-five billion dollars in cash and short-term investments. They could have written a check. Instead, they chose to pay in shares, which means existing shareholders get diluted if the milestones are hit.

So what could this company be? Nobody knows for sure. But the timing gives us clues.

On April fifteenth, Tesla completed the tape-out of its AI5 chip. That’s the next generation of Tesla’s custom AI processor. A few weeks earlier, Intel officially signed on as the manufacturing partner for Terafab, Tesla’s twenty-five billion dollar semiconductor factory in Austin.

Tesla is building an end-to-end AI chip supply chain. Design. Manufacturing. Packaging. This mystery acquisition could fill any gap in that chain. It could be a chip design firm. A packaging company. An AI accelerator startup with intellectual property Tesla needs for Terafab.

And here’s the bigger picture. This two billion dollar deal isn’t happening in isolation. In the same quarter, Tesla invested two billion dollars in SpaceX common stock. That’s the entity that absorbed xAI in the merger earlier this year. Tesla also raised its capex guidance from twenty billion to twenty-five billion for 2026. Most of that increase is going to AI.

Add it up. That’s potentially four billion dollars in AI investments and acquisitions in a single quarter, on top of twenty-five billion in planned capital spending. All while the core car business delivered just four hundred seventy-seven million dollars in net income.

Let me say that again. Tesla made four hundred seventy-seven million in profit and committed four billion to AI bets. The AI spending is more than eight times the profit.

Now, here’s the part that raises questions. Tesla didn’t mention this acquisition on the earnings call. They talked about the SpaceX investment. They talked about Terafab. They talked about the AI5 chip. But a two billion dollar acquisition of an unnamed company? Silence.

For context, Tesla’s previous largest acquisition was about ninety-six million dollars back in 2019. That included Grohmann Engineering for factory automation and parts of Maxwell Technologies for battery tech. This new deal is more than twenty times that size.

There are a few possible explanations for the silence. The deal may not be fully finalized. There could be competitive reasons to keep the target company quiet. Or Tesla might be trying to minimize attention on the dilutive impact of issuing up to two billion dollars in new stock.

But here’s what we know for sure. Tesla is making a massive bet that its future isn’t cars. It’s AI hardware. Custom chips. Semiconductor manufacturing. The infrastructure that powers autonomous driving, humanoid robots, and the data centers that train the neural networks behind all of it.

If you’ve been following this podcast, you’ll remember our very first episode on Terafab. That was the announcement. This acquisition might be part of the execution.

So what does this tell us? Tesla just made the largest acquisition in its history. Twenty times bigger than anything they’ve done before. And they tried to hide it in a footnote. That tells you two things. First, Tesla believes its future depends on owning every piece of the AI hardware stack. And second, they’re not entirely sure they want investors asking questions about it yet.

That’s the signal. I’m Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk’s companies, visit our website at tesorb.com.

Got a tip or feedback? Send it to [email protected], we’d love to hear from you. This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

SpaceX just landed a rocket for the six hundredth time. The next day, Blue Origin lost a customer satellite on only its third New Glenn flight. We break down what SpaceX’s launch dominance looks like in numbers, what went wrong for Blue Origin, and why the gap between first and second place in orbital launch keeps getting wider.

Tesorb Signal — Episode 007 Transcript

SpaceX at 600 Landings: Why Nobody Can Catch Up

Six hundred. That's how many times a SpaceX rocket has landed itself after delivering a payload to orbit. Six hundred successful touchdowns on drone ships and landing pads since 2015. The day that happened, last Friday, nobody even blinked. It was just another launch.

The next day, Blue Origin tried to reuse one of its own rockets for the first time. It landed the booster. And then the upper stage failed and put a customer's satellite into a useless orbit.

Two launches. One day apart. One company on its six hundredth landing. The other losing a payload on its third flight. That gap tells a story.

This is Tesorb Signal and I'm Lena Ruiz. Today we're looking at SpaceX's growing launch monopoly, what Blue Origin's bad weekend means for the industry, and why the gap between first and second place keeps getting wider.

Let's start with what six hundred landings actually means. SpaceX launched 167 orbital missions in 2025. That's more than half of every rocket launched on Earth that year. They put up 85% of all satellites. Their closest Western competitor, Rocket Lab, managed 18 launches. The gap isn't narrowing. It's accelerating.

And it's not just volume. It's cost. A Falcon 9 launch runs about $67 million. But because SpaceX reuses the first stage, the most expensive part of the rocket, their actual cost per flight is dramatically lower. One booster that flew recently was on its 22nd mission. That's the same rocket, flying again and again, while competitors are still throwing theirs away.

So far in 2026, SpaceX has launched over 40 Falcon 9 missions and put more than 1,000 Starlink satellites into orbit. They completed the GPS III constellation for the Space Force. They're averaging a launch roughly every three days. And they're doing it so reliably that it barely makes the news anymore.

Now let's talk about what happened at Blue Origin. On April 19, the company launched its third New Glenn rocket. The booster, nicknamed "Never Tell Me the Odds," had flown once before in November. It landed perfectly on a drone ship in the Atlantic. That part worked.

But the upper stage didn't. One of its two engines failed to produce enough thrust to reach the target orbit. The payload, an AST SpaceMobile satellite called BlueBird 7, ended up too low to operate. It'll burn up in the atmosphere. The satellite is a total loss.

The FAA grounded New Glenn until Blue Origin completes an investigation. That's a big deal. Blue Origin had plans to launch up to 12 more missions this year, including a prototype moon lander and Amazon's first batch of internet satellites to compete with Starlink. All of that is now on hold.

Here's the context that makes this interesting. Blue Origin has been developing New Glenn for over a decade. This was only its third flight. And it was the first time the company carried a paying customer's satellite. Losing that payload on your first real commercial mission is exactly the kind of setback that shakes customer confidence.

SpaceX had setbacks too. In 2015, a Falcon 9 blew up during flight and destroyed a cargo ship headed to the space station. In 2016, one exploded on the launch pad and took out a satellite for Facebook. But SpaceX recovered fast. They investigated, they fixed the problem, and they kept flying. The question for Blue Origin is whether they can do the same.

And that connects to something bigger. This isn't just about two rocket companies. It's about whether anyone can build a credible alternative to SpaceX in orbital launch. Right now, the answer is: not yet.

United Launch Alliance has Vulcan, which flew successfully but hasn't hit high-volume operations. Rocket Lab is developing its medium-lift Neutron rocket. Europe's Ariane 6 is flying but can't match Falcon 9 on price. And China has its own ecosystem but doesn't compete for Western commercial contracts.

The Space Force, NASA, and commercial satellite operators all need a healthy launch market with more than one reliable provider. National security depends on it. But building a rocket that works is hard. Building one that works reliably, lands itself, and costs a fraction of the competition is something only one company has figured out.

SpaceX didn't get to 600 landings overnight. The first one was in December 2015, and people thought it was a miracle. Now they land boosters like it's routine. Because for them, it is. That's the real moat. Not the technology. The thousands of reps.

So here's what to watch. If Blue Origin's investigation is quick and clean, they could be flying again in weeks. If it drags, the ripple effects hit Amazon's Starlink competitor, NASA's moon landing timeline, and Blue Origin's credibility with future customers. Meanwhile, SpaceX will keep launching. 601. 602. Every few days, like clockwork.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com.

Got a tip or feedback? Send it to [email protected], we'd love to hear from you. This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Got a tip or feedback? Email [email protected]

On April 10, the Netherlands became the first country in Europe to approve Tesla’s Full Self-Driving system for public roads. Eighteen months of testing. 1.6 million kilometers driven. 13,000 ride-alongs. Over 400 compliance requirements. We break down what’s different about FSD in Europe, why one small country could unlock the continent by summer, and what it means for Tesla’s global autonomy strategy.

Tesla FSD Lands in Europe: What's Different and What's Next

Tesorb Signal — Episode 6

April 24, 2026

On April tenth, a Dutch government agency did something no European regulator has ever done. It approved Tesla's Full Self-Driving system for use on public roads. Eighteen months of testing. One point six million kilometers driven. Thirteen thousand ride-alongs. Over four hundred compliance requirements. And just like that, Europe's door cracked open.

This is the Tesorb Signal and I'm Lena Ruiz. Today we're breaking down what Tesla's first European FSD approval actually means, how the European version differs from what American drivers use, and why one small country could unlock an entire continent by summer.

Let's start with what happened. On April tenth, twenty twenty-six, the RDW, that's the Netherlands' national vehicle authority, granted Tesla a European type approval for FSD Supervised. That makes the Netherlands the first country in Europe to officially authorize Tesla's driver assistance system for customer use on public roads.

This wasn't a rubber stamp. The RDW spent over eighteen months testing the system. Tesla logged more than one point six million kilometers on European roads. They ran thirteen thousand customer ride-alongs and over forty-five hundred track test scenarios. The documentation alone covered more than four hundred separate compliance requirements under a United Nations regulation called R-171.

And here's how fast Tesla moved once the green light came. Within one day of approval, Tesla pushed the FSD software to Dutch customer vehicles through an over-the-air update.

The infrastructure was staged and ready. They were just waiting for the paperwork.

Now here's where it gets interesting. European FSD is not the same system American drivers are using. The RDW was explicit about this. Their official statement says the US and European software versions are, quote, not comparable one-to-one.

So what's different? A few key things. In the US, Tesla self-certifies its vehicles and regulators step in after the fact. In Europe, Tesla had to get pre-market approval before a single customer could use the system. That's a fundamentally different bar to clear.

The driving behavior is also more conservative. The European build errs on the side of caution. There's no aggressive driving mode. When the system isn't sure about a speed limit, it slows down rather than guessing.

The driver monitoring is stricter too. Your hands don't need to be on the steering wheel, but they need to be ready to grab it immediately. Sensors track your eyes and attention constantly. If you look away too long, the system warns you. If you keep ignoring it, FSD disables itself.

And this is important. The RDW was very clear that these vehicles are not self-driving. Their exact words: FSD Supervised can take over many driving tasks, but the vehicles are not autonomous or self-driving. This is a Level 2 driver assistance system. The driver is legally responsible at all times.

For context, this puts Tesla in the same legal category as BMW's hands-free highway driving or Ford's BlueCruise. Both of those systems already have European approvals. But Tesla's system is significantly more capable.

It handles city streets. Intersections. Unprotected turns. Roundabouts.

That distinction matters because it's actually what got Tesla through the door. Europe has extremely strict rules around autonomous vehicles. Calling your system autonomous invites a whole different level of regulatory scrutiny. By positioning FSD as a very advanced driver assistance system, Tesla sidestepped the hardest questions entirely. Smart regulatory strategy.

So what happens next? This is the part that could change everything. The Netherlands isn't the destination. It's the template.

Under European Union rules, every member state can now recognize the Dutch approval through a process called mutual recognition. That means Germany, France, Spain, and the rest don't need to repeat the full eighteen-month review. They can adopt the RDW's decision nationally.

Think of it like a driver's license from one EU country. One authority does the testing, and the rest honor the result.

Some countries are already moving. Belgium has signaled strong support for a fast-track adoption. Spain has been running its own testing program with thirty Tesla vehicles since November twenty twenty-five, logging eighty thousand kilometers with zero reported incidents.

Tesla has published footage of FSD testing on the streets of Rome. Even Norway, which tends to be cautious, is engaging seriously and asking detailed technical questions about how the system handles Nordic conditions.

The big date to watch is May fifth. That's when the TCMV, the EU's Technical Committee on Motor Vehicles, meets to hear the RDW present its approval. The European Commission recently extended the time slot for that presentation from twenty minutes to a full hour. That tells you member states have serious questions and want real answers.

If a majority vote follows, either at that meeting or at the next session on June thirtieth, FSD Supervised could become available across the entire European Union automatically. Tesla's projection is EU-wide access by summer twenty twenty-six.

And there's a business angle here that connects to the bigger picture. Elon Musk's compensation package includes a milestone requiring ten million active FSD subscriptions for his stock awards to vest. Tesla hit one million subscriptions at the end of twenty twenty-five. That's a meaningful start, but it's a long way from ten million.

Opening Europe at ninety-nine euros per month is a direct path to closing that gap. The EU has roughly one point five million Teslas on the road. Even a modest take rate would move the needle significantly.

So here's what this really changes. This isn't just about one country. What the Netherlands did on April tenth was build a regulatory bridge. Tesla spent eighteen months and over a million kilometers proving to one European authority that FSD Supervised is safe enough for public roads.

Now every other EU country can walk across that bridge without starting from scratch. If the May fifth meeting goes well, we could see FSD rolling out across Europe by the end of summer.

One approval. Twenty-seven countries, and a whole continent waiting.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com. Got a tip or feedback? Send it to [email protected], we'd love to hear from you. This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Got a tip or feedback? Email [email protected]

Wall Street expected Tesla to burn $1.5 billion last quarter. Instead, the company generated $1.5 billion. A $3 billion swing nobody saw coming. But Tesla told investors not to get comfortable.

On the earnings call, the CFO reframed the entire company. Full self driving is now the product. The vehicle is just the delivery mechanism. 1.28 million people are paying for FSD subscriptions, up 51% year over year. Then Musk dropped a bombshell: Hardware 3 vehicles will never get unsupervised autonomy.

Tesla raised capex guidance to over $25 billion. Model S and X production lines are shutting down to make room for Optimus robot assembly. A $3 billion chip fab called Terafab is going up at Giga Texas. The Cybercab is entering pilot production. Every major bet Tesla is making is on AI and autonomy, not cars.

Wall Street expected Tesla to lose one and a half billion dollars last quarter. Instead, the company made one and a half billion. That's a three billion dollar swing that nobody predicted. And then management got on the call, and said, don't get used to it.

This is Tesorb Signal and I'm Lena Ruiz. Tesla just reported first quarter 2026 earnings and the numbers surprised everyone. But the numbers aren't the real story. Today we're looking at three things. The financial surprise that flipped the narrative. A dramatic shift in what Tesla says it actually is. And a twenty-five billion dollar spending plan that tells you where this company is headed.

Let's start with the money. Tesla brought in twenty two point four billion in revenue, up sixteen percent from a year ago. But the number that really mattered was cash flow. That's basically how much actual cash the business generates after paying all its bills.

Wall Street expected Tesla to burn through one and a half billion dollars. Instead, Tesla generated one and a half billion. A nearly three billion dollar swing between what analysts predicted and what actually happened.

Margins improved too. But here's the asterisk. About two hundred and thirty million of that came from one-time items, things like warranty adjustments and tariff benefits that won't repeat. So the quarter was genuinely good, just not quite as good as the headline suggests.

And there's a cloud. Tesla delivered three hundred fifty eight thousand vehicles. Up six percent, but below what analysts expected. And the company built fifty thousand more cars than it sold. That gap matters because it suggests demand might not be keeping up with production.

Now here's where the quarter gets really interesting, and where the one thing I want you to take away from this episode comes into focus. The CFO used a phrase on the earnings call that reframes everything. He said Tesla now treats full self driving as the product and the vehicle as just the delivery mechanism.

Think about that for a second. It's like Apple saying the iPhone isn't really a phone. It's a platform for the App Store. The hardware is just the box the real product comes in.

The numbers back it up. One point two eight million people are now paying for full self driving subscriptions, up fifty one percent from last year. Fewer people are canceling. Drivers are using it for longer stretches. The software is becoming the reason people buy the car.

But then came the bombshell. Musk said hardware three vehicles, which is every Tesla built before twenty twenty-four, will never get unsupervised full self driving. Never. He said the older chip has one eighth the processing power of the new hardware four. It's a hard ceiling that affects millions of existing owners.

Tesla's solution? Discounted trade-ins and plans for small retrofit centers in major cities.

Meanwhile, Tesla expanded its robotaxi service to Dallas and Houston. If you listened to our episode three, you know the reality behind these expansion announcements. When we dug into the numbers, we found as few as four to twelve truly driverless cars operating at any given moment across all cities. The maps are expanding faster than the fleets.

Musk acknowledged this himself on the call. He told a story about a dozen robotaxis getting stuck in a left turn lane in Austin because a Waymo had crashed into a parked bus. The Teslas were too cautious to go around. He also mentioned cars getting caught in loops, circling the same block over and over because of construction. The cars aren't unsafe. They're just too careful.

Now let's talk about where Tesla is putting its money, because this is where the transformation from car company to technology company gets real. Tesla raised its capex, its capital spending budget, to over twenty-five billion dollars for twenty twenty-six. Up from twenty billion just months ago. And the CFO warned that the company will likely burn cash for the rest of the year.

Most of that money is going toward things that have nothing to do with cars. The Model S and X production lines at Fremont are shutting down next month. In their place, Tesla is building assembly lines for Optimus, its humanoid robot. Musk called it the biggest product ever.

Think of it like a factory metamorphosis. The same building that used to make luxury sedans will now make robots that walk. That's not diversification. That's transformation.

There's also a three billion dollar chip fab going up at Giga Texas, called Terafab. If you listened to episode one, that's the twenty-five billion dollar chip factory we covered in detail. And the Cybercab, Tesla's purpose-built robotaxi with no steering wheel, is entering pilot production. Every major investment Tesla is making right now is a bet on autonomy and AI, not on selling more cars.

So here's what this quarter really tells us. The earnings were strong. The cash flow surprised everyone. But Tesla is using this moment of strength to fund a complete identity shift. The question isn't whether Tesla can keep selling cars. It's whether a company born in the auto industry can remake itself into an AI and robotics company while still moving a million and a half vehicles a year. That's the bet. And Tesla is all in.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com.

Got a tip or feedback? Send it to [email protected]. We'd love to hear from you. Of course, this podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Got a tip or feedback? Email [email protected]

SpaceX filed for an IPO targeting a 1.75 trillion dollar valuation and a 75 billion dollar raise. That would shatter every record in stock market history. We break down what Starlink’s revenue actually supports, why Musk is giving retail investors 30 percent of the shares, the xAI merger, the new 60 billion dollar Cursor deal, and the risks nobody in the hype cycle wants to talk about.

Tesorb Signal — Episode 004 Transcript

SpaceX IPO: The Biggest Debut in History, or the Most Expensive?

One point seven five trillion dollars. That's the number SpaceX just put on itself. If this IPO goes through at that valuation, it would be the largest public offering in history. Not by a little. By a lot.

The previous record holder was Saudi Aramco, which raised twenty-nine billion dollars in 2019. SpaceX is targeting seventy-five billion. Nearly three times the record.

This is Tesorb Signal and I'm Lena Ruiz. Today we're breaking down the SpaceX IPO. What the numbers actually say, what investors are really buying, and why this might be the most consequential stock market event of the decade.

Here's what we know so far. SpaceX filed its confidential draft registration with the SEC on April first. The public version of that filing, the S-1, is expected in late April. The roadshow kicks off the week of June eighth, and the company is targeting a Nasdaq listing sometime that month.

Twenty-one banks are involved. Morgan Stanley, Goldman Sachs, JP Morgan, and eighteen others. Wall Street has a name for the project. They're calling it Project Apex.

Now let's talk about what makes this IPO so unusual. SpaceX is reportedly planning to allocate thirty percent of IPO shares to retail investors. That's everyday people, not hedge funds or pension funds. The standard for most IPOs is five to ten percent.

This is classic Musk. He's cultivated a massive retail following through Tesla, and now he's giving those same investors a front-row seat. There's even a dedicated retail investor event planned for June eleventh, with fifteen hundred attendees from seven countries.

Elon Musk is not just launching an IPO. He's throwing a show.

But here's the question nobody in the hype cycle wants to ask. What exactly are investors buying at one point seven five trillion dollars?

Let's look at the revenue. Starlink generated roughly ten billion dollars in 2025. The entire company brought in somewhere between fifteen and eighteen billion, depending on whose numbers you trust.

Reuters says SpaceX made an eight billion dollar profit. But a report from The Information, a tech publication, says it lost five billion. That's a thirteen billion dollar gap, and until the S-1 drops, nobody outside the company can settle it.

Even using the most generous numbers, a one point seven five trillion dollar valuation means investors are paying roughly a hundred times revenue. For context, when Meta went public in 2012, it was valued at about twenty times revenue. Nvidia at the peak of the AI boom topped out around forty times. SpaceX would be nearly three times more expensive than Nvidia at its most expensive.

So why would anyone pay that? Three reasons.

First, Starlink. This is the real business. Ten million subscribers across a hundred and fifty countries. Revenue projected to hit twenty billion this year. Starlink is adding customers faster than any telecom network in history.

And it's profitable. Estimated margins north of sixty percent. If Starlink were its own company, it would already be one of the most valuable internet service providers on the planet.

Second, launch dominance. SpaceX launched a hundred and sixty-five missions in 2025. That's eighty-five percent of all U.S. orbital launches. It just hit its six hundredth Falcon booster landing last week. No other company is remotely close to this cadence.

And the cost advantage is enormous. SpaceX has brought launch costs down from fifteen thousand dollars per kilogram in 2008 to under a thousand today.

And then there's Starship. The V3 version just completed a full-duration static fire test at Starbase. Flight twelve, the first V3 flight, is targeting May. If it works, it could carry over a hundred tons to orbit. That's three times what the current version can manage. A successful test before the June roadshow would be the single biggest catalyst for investor confidence.

Now here's where it gets complicated. In February, SpaceX merged with xAI, Musk's artificial intelligence company. That merger brought in the social media platform X, the Grok chatbot, and something Musk is betting the whole thesis on. Orbital data centers.

The idea is wild but logically consistent. AI training requires enormous amounts of power. Terrestrial data centers are hitting limits on energy, cooling, and local opposition. Musk's pitch is that solar-powered data centers in orbit could eventually be cheaper than building them on Earth. SpaceX filed with the FCC for up to one million satellites to make it happen.

And just yesterday, SpaceX made another move. It struck a deal with Cursor, the AI coding startup, for the right to acquire the company for sixty billion dollars later this year. Or, if it prefers, it can pay ten billion for collaborative work instead. Cursor's valuation was two and a half billion just eighteen months ago.

The message is clear. Musk isn't just adding AI as a side story to the IPO. He's building an AI conglomerate. Rockets, satellites, internet, social media, a chatbot, a supercomputer, and now one of the hottest developer tools on the planet. All under one roof, all heading into a single public offering.

But let's be clear about what's proven and what isn't. There are no orbital data centers operating at scale today. Cursor is a real product with real users, but sixty billion dollars is an extraordinary price. And xAI is currently burning about a billion dollars a month. The AI pieces add a visionary narrative to the IPO, but they also add a very large cash furnace.

There's one more thing investors need to understand. Musk isn't giving up control. The IPO uses a dual-class share structure. He holds about forty-two percent of the equity but controls roughly seventy-nine percent of the votes.

That means public shareholders will own pieces of the company, but they won't steer it. SpaceX will go public the same way it's been run as a private company. Musk decides.

And that brings us to the biggest risk that has nothing to do with rockets. Market conditions. The Nasdaq has been volatile. Oil prices are spiking. Geopolitical uncertainty is running high. Even a great company can have a bad IPO if the market isn't cooperating.

History tells us that mega-IPOs almost always underperform the broader market over the first year. Only one of the five largest IPOs in history, Meta, has beaten the S&P 500 long term.

So where does that leave us? SpaceX is a real business with real revenue and genuine competitive advantages. Starlink alone might justify a massive valuation. But at a hundred times revenue, investors aren't buying what SpaceX is today. They're buying what Musk says it will become. Orbital data centers. Mars colonization. A hundred gigawatts of AI compute beamed from space.

That's either the investment opportunity of the century, or the most expensive bet in stock market history. Probably both.

And speaking of big numbers, Tesla reports earnings later today. Revenue, margins, Robotaxi timelines, and what Elon says about getting back to full-time CEO. We'll have the full breakdown in our next episode.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com.

Got a tip or feedback? Send it to [email protected] — we'd love to hear from you. This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Got a tip or feedback? Email [email protected]

Tesla just expanded its robotaxi service to Dallas and Houston, making four cities total. The stock surged 12%. But behind the headlines, the numbers tell a more complicated story. We break down the fleet size gap, the safety data, the pricing war with Waymo, and what it all means for the trillion-dollar race to dominate autonomous rides.

Tesorb Signal — Episode 003 Transcript

Tesla Robotaxi: Four Cities, But How Many Cars?

Four cities. That's the headline. Tesla's robotaxi service just launched in Dallas and Houston, joining Austin and San Francisco. Four cities sounds like momentum. It sounds like scale. But here's the thing. When you zoom into those maps, the actual service areas are tiny. We're talking twenty-five square miles in Houston. Only a few neighborhoods around Highland Park in Dallas. And, just a handful of cars to cover them.

This is Tesorb Signal and I'm Lena Ruiz. Today we're asking a simple question. Is Tesla actually scaling its robotaxi business, or is it scaling the map?

Let's start with what just happened. On April eighteenth, Tesla opened robotaxi rides in Dallas and Houston. Fully unsupervised. No safety driver in the front seat. You download the Tesla app, tap the robotaxi tab, enter a destination inside the geofence, and a Model Y shows up to get you. Rides run from six a.m. to two a.m. Introductory pricing is three twenty-five base fare plus a dollar per mile. That's significantly cheaper than Uber or Lyft in the same cities.

The timing was not an accident. Tesla reports first quarter earnings tomorrow. Having a "we just expanded to four cities" talking point on the earnings call is exactly the kind of momentum signal that moves the stock. And it worked. Shares jumped twelve percent in the four trading days leading up to the launch.

Now, here's where it gets interesting. Let's talk about what Tesla isn't telling us. They didn't disclose how many vehicles are operating in Dallas or Houston. They didn't say whether every ride is truly unsupervised or if some cars still have monitors. And the geofences are small. In Houston, the service covers the Jersey Village and Willowbrook area northwest of downtown. That's it.

For context, Tesla's Austin operation launched in June of last year. It took nearly a year for that geofence to grow from twenty square miles to two hundred and forty-five. And even now, independent tracking suggests only four to twelve truly unsupervised vehicles are on the road at any given time in Austin. Out of a fleet of roughly eighty cars.

So four cities on the map, but potentially fewer than twenty driverless cars actually operating at any moment across all of them.

Let's switch gears and talk about the elephant in the room. Waymo.

Waymo has been operating in both Dallas and Houston since February. Two months before Tesla showed up. And the scale difference is staggering. Waymo runs about twenty-five hundred robotaxis across ten U.S. cities. It's completing five hundred thousand paid rides every week. All fully driverless. No safety monitor. No chase car. And it's targeting one million weekly rides by the end of this year.

Tesla's total paid ride count through November was about seven hundred thousand. Cumulative. Waymo does more than that in two weeks.

And then there's the safety data, which is hard to ignore. Tesla's Austin fleet has reported fourteen crashes to federal regulators since launching last June. That works out to roughly one crash every fifty-seven thousand miles. Tesla's own safety report says the average human driver has a minor collision every two hundred and twenty-nine thousand miles. By Tesla's own math, its robotaxis crash about four times more often than humans. And most of those miles were driven with a trained safety monitor in the car.

Waymo, by comparison, has logged over a hundred and twenty-seven million fully driverless miles. Independent research shows Waymo reduces serious injury crashes by ninety-one percent compared to human drivers.

There's also a transparency gap. Waymo, Zoox, Aurora, and other companies provide detailed descriptions of every crash in their federal filings. Tesla redacts everything, marking the incident narratives as confidential business information. We literally cannot verify whether Tesla's system was at fault in any of these crashes. That matters when you're asking people to get in a car with no driver.

Now, before this starts sounding one-sided, let's talk about why some very smart investors still believe Tesla wins this race long-term.

First, pricing. Tesla is dramatically cheaper. In San Francisco, the average Tesla robotaxi ride costs about eight dollars. Waymo charges roughly double that. Uber and Lyft are even more. Tesla is doing what Uber did in its early days. Buy the market with unsustainable pricing, then figure out profitability at scale.

Second, the data flywheel. Tesla has millions of customer vehicles on the road collecting driving data every single day. Waymo has twenty-five hundred cars. Tesla's argument is that this massive data advantage will let its AI improve faster than anyone else's over time. That's a real argument. Whether it's playing out in the crash data yet is a different question.

Third, the Cybercab. Tesla's purpose-built two-seat robotaxi is expected to enter production later this year. It has no steering wheel, no pedals. It's designed from the ground up to be cheaper to build and operate than retrofitting a Model Y. If Tesla can manufacture Cybercabs at scale, the economics change dramatically.

Morgan Stanley estimates the U.S. robotaxi market alone could be worth over a trillion dollars. Even at twenty-five percent penetration, that's a two hundred and fifty billion dollar opportunity. The prize is enormous. The question is whether Tesla can close the gap between the map and the math.

So what does all this add up to. Tesla is expanding its robotaxi footprint fast. Four cities today. Seven planned by the end of June. But footprint isn't the same as scale. Right now, Tesla has map dots. Waymo has a transportation network. The real test isn't how many cities Tesla can pin on a map. It's how many cars it can put on the road, how safe those cars are, and whether riders trust them enough to keep coming back.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com. Got a tip or feedback? Send it to [email protected] — we'd love to hear from you. Of course, this podcast wouldn't be possible without the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Got a tip or feedback? Email [email protected]

Tesla’s Optimus humanoid robot is now working on factory floors and posing for photos at the Boston Marathon. But the real Optimus story isn’t about what the robot can do. It’s about Tesla’s plan to build a million of them a year. We break down Gen 3’s capabilities, how it stacks up against competitors like Boston Dynamics, Figure AI, Agility Robotics, and Unitree, and why manufacturing scale may matter more than robot demos.

This weekend, a Tesla robot stood on Boylston Street in Boston, posing for selfies with marathon runners. A humanoid robot. In public. No velvet ropes. No controlled demo. Just a robot hanging out at the finish line of one of the most famous races on Earth.

And nobody freaked out.

This is Tesorb Signal and I'm Lena Ruiz. Today we're talking about Optimus, Tesla's humanoid robot. Not just what it can do, but why the real story isn't about the robot at all. It's about the factory.

Let's start with what's happening right now. Tesla has Optimus robots working inside its factories. Not in a lab. On the production floor. Sorting battery cells, handling parts, doing repetitive work that's perfect for a machine that doesn't get tired.

The hands are the headline. Gen 3 has twenty-two degrees of freedom per hand, with fifty actuators total. That's four and a half times more than Gen 2, in just two years. Musk has called the hand "the majority of the engineering difficulty of the entire robot." Harder than Cybertruck and harder than the Model X.

By early this year, Tesla said it had cracked the hardest problems. Human-level dexterity. Real-world AI integration. And critically, volume production scalability.

That last one matters most.

Here's where the story shifts. Tesla isn't the only player in humanoid robots. Not even close.

Boston Dynamics unveiled its all-electric Atlas at CES, the largest Consumer Electronics Show in the world. It won Best Robot from CNET's panel of over forty tech journalists. Figure AI has raised over a billion dollars. Its Figure 02 ran ten-hour shifts at a BMW plant in South Carolina, helping build over thirty thousand cars before being retired. Now Figure 03 is pivoting to the home market. Agility Robotics has Digit, already deployed in warehouses. And China's Unitree is selling a humanoid for under fourteen thousand dollars.

The competition is real. Some of these robots are incredibly impressive.

But here's what none of them can do. None of them can be manufactured at car scale and car prices.

Boston Dynamics Atlas is expected to cost around a hundred and fifty thousand dollars. Agility's Digit is roughly two hundred and fifty thousand. Figure AI hasn't disclosed pricing, but estimates range from thirty to a hundred and fifty thousand.

Tesla is targeting twenty to thirty thousand dollars per unit. At scale.

And they're backing it up. Tesla is converting its Fremont factory, the one that used to build Model S and Model X, into an Optimus production line. The target? One million robots per year. They've already stopped making the S and X to make room.

The entire humanoid robotics industry has deployed maybe a few thousand robots total. Tesla wants to build a million a year.

Now, Tesla's secret weapon here isn't some AI breakthrough. It's the same thing that made Tesla dominant in electric vehicles. Manufacturing.

Tesla knows how to build complex electromechanical systems at massive scale. It has the supply chains, the battery expertise, and the AI training infrastructure from Full Self-Driving. The Optimus robot runs on the same neural network as Tesla's self-driving system. Every mile driven by a Tesla on the road generates data that helps train the robot. No other humanoid company has that flywheel.

So let's be honest about what still has to go right.

The gap between a robot sorting batteries in a factory and a robot doing useful work in your home is enormous. Factory tasks are structured and predictable. A home is chaos. Uneven floors. Kids' toys. A cat that knocks things off the counter. That's a different AI challenge entirely.

Timelines have slipped too. Tesla aimed to show the Gen 3 production model in Q1 of this year. That deadline passed. Musk said it's walking around but needs finishing touches. Low-volume production is now expected this summer, with a real ramp in 2027.

And the skeptics have a point. Roboticist Rodney Brooks, co-founder of iRobot, has called general-purpose home robots quote "pure fantasy thinking." He argues that robots are fundamentally coordination-challenged in unstructured environments.

So where does that leave us?

Here's the bottom line. Most humanoid companies are solving the robot problem. Tesla is solving the factory problem. And in hardware, the company that figures out manufacturing at scale usually wins. Not the company with the best prototype.

It's the same arc as electric vehicles. The Roadster wasn't the best EV ever built. But the Gigafactories changed the industry forever.

Optimus might follow the same path. The robot you see today isn't going to change the world. The factory that builds a million of them just might.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com. Got a tip or feedback? Send it to [email protected] — we'd love to hear from you. Of course, this podcast wouldn't be possible without the help of using AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.

The Terafab announcement was treated as classic Musk ambition: Tesla has never manufactured a chip, the two-nanometer node is barely working at TSMC, and “vertical integration” is the kind of phrase that gets laughed out of industry conferences. Then on April 7, Intel formally joined as the manufacturing partner — and on the exact same day, announced a new kind of chiplet that combines gallium nitride power electronics with silicon logic on a nineteen-micron substrate, a feat that was considered physically impossible until the press release.

This episode unpacks why that technical breakthrough is the whole story. Gallium nitride is radiation-tolerant, making it purpose-built for the kind of orbital compute that Starlink is scaling toward forty thousand satellites. Terafab isn’t really a semiconductor play — it’s a bet that orbital AI compute is the next real market, and that Intel happens to be the only manufacturer who can serve it.

Tesorb Signal — Episode 001

Terafab: The $25B Chip Fab Nobody Believed In

Host: Lena Ruiz

Runtime: 6:39

Two weeks ago, every semiconductor analyst on the planet said Elon Musk was out of his mind for trying to build a chip factory. Last week, Intel signed on as his manufacturing partner. And suddenly nobody's laughing.

This is Tesorb Signal and this is Lena Ruiz. Let's talk about Terafab — what changed, why it matters, and what still has to go right.

Here's the short version. Tesla, SpaceX, and xAI are building a twenty-five billion dollar chip fab on the GigaTexas campus in Austin. Two nanometer process. A hundred thousand wafer starts per month at full scale. Target output: one terawatt of AI compute per year.

One terawatt. The entire global output of AI accelerators right now is around twenty gigawatts. Terafab is aiming for fifty times that. From one campus.

Now. The industry was right to be skeptical. Tesla has never manufactured a chip. They've designed plenty — the Dojo silicon, the FSD compute board, the AI4 inference chip in every car they ship. But every wafer came out of someone else's fab. Usually TSMC. Sometimes Samsung.

Chip manufacturing is not chip design. It's a different discipline. Lithography. Yield engineering. Process chemistry. The kind of knowledge that takes roughly thirty years to build. Tesla doesn't have it.

Which is exactly why the Intel deal changed the math.

April seventh. Intel formally joins Terafab. Same day — and this is the part that didn't get enough attention — Intel announces a new chiplet. Gallium nitride power electronics and silicon logic, combined on the same substrate. The silicon base is nineteen microns thick. One-fifth the width of a human hair.

Why does that matter?

Gallium nitride is radiation-tolerant. Regular silicon degrades in orbit. Solar radiation eats it over time. GaN doesn't care. So this chiplet is basically purpose-built for satellite compute. For Starlink. For orbital AI inference.

And it's light. Launch costs run a thousand to ten thousand dollars per pound. A thinner, lighter chip is a cheaper chip to put in space. At seven thousand satellites and counting, heading toward forty thousand, that math compounds fast.

So here's the Tesorb read on this. Terafab is not a general-purpose semiconductor play. It's a bet that orbital AI compute is a real market — and that the specific chip you need to serve it requires someone who can stack GaN and silicon on one substrate. That someone is Intel.

Intel gets a guaranteed foundry customer at a moment when nobody else is buying their capacity. Tesla and SpaceX get the one partner who can actually build what they need. Both sides are desperate. Both sides have leverage. That's the deal.

Let's talk demand, because this is where the bear case gets harder to hold.

Start with Optimus. Tesla's targeting ten million humanoid robots a year by 2030. Each one needs a brain chip. That alone is tens of thousands of wafers monthly.

Full self-driving. Every new Tesla ships with an AI5 chip once production starts. A couple million units a year. Plus every Tesla already on the road is a candidate for a retrofit.

Starlink. Each next-gen satellite is a miniature data center that needs a radiation-hardened compute module. That's the GaN chiplet, purpose-built for orbit.

Stack those up and a hundred thousand wafers a month stops sounding ambitious. It starts sounding conservative.

What about Nvidia? What does one terawatt of in-house compute do to their pricing power?

Short answer — not much in 2027. A lot by 2030. Tesla isn't selling these chips on the open market. But every wafer produced in-house is one they don't buy from Nvidia. And if SpaceX sources orbital compute at cost, Nvidia's margins on aerospace contracts get uncomfortable.

Musk is betting vertical integration breaks the dependency. Nvidia is betting the bleeding edge stays out of reach. One of them is wrong.

Now the risks. Because there are real ones.

Two nanometer is hard. TSMC is struggling with yields at that node right now. Intel's own two nanometer process — they call it 18A — is the entire foundation of Intel's foundry business. It has not yet proven itself at volume.

If 18A stumbles, Terafab stumbles. And Intel's foundry credibility, already fragile, takes another hit.

Then there's timelines. Musk has a well-documented pattern. Optimus was supposed to ship in 2025. It hasn't. The Cybertruck was two years late. FSD has been "next year" for five consecutive years.

Ambition is not execution. Tesorb's position is that both of those things can be true at the same time.

Three things to watch.

First — the AI5 pilot wafers. Expected late this year. If real silicon comes off that line and works, the skeptic case collapses.

Second — Intel 18A yields. This is the gating factor. Everything downstream depends on whether Intel can manufacture at two nanometers reliably.

Third — whether the GaN chiplet scales. Nineteen-micron silicon bonded to gallium nitride is a beautiful lab result. Whether it survives mass production is a different question entirely.

If all three go right, Terafab is the largest vertical integration play in tech history. If any of them break, it's twenty-five billion dollars of concrete in a field in Austin.

We'll know a lot more by this time next year.

That's the signal. I'm Lena Ruiz with the Tesorb Signal podcast. For more news about Tesla, SpaceX, and Elon Musk's companies, visit our website at tesorb.com. Got a tip or feedback? Send it to signal at tesorb dot com — we'd love to hear from you. Of course, this podcast wouldn't be possible without the help of AI assistants, including the voice, and undergoes a detailed review during production. See you on the next one.

More at tesorb.com/podcast

Tesorb Signal is a tech-news analysis podcast from tesorb.com. One story, done properly. Subscribe at tesorb.com/podcast or wherever you listen.

This podcast was developed with the help of using AI assistants, including the voice, and undergoes a detailed review during production. Corrections are posted to the episode page.