SpaceX

Magic AI

SpaceX

SpaceX makes money three ways. First, launch services — companies and governments pay SpaceX to put their satellites into orbit.

A Falcon 9 launch costs about $67 million, which undercut the competition by 75% when it debuted. Second, Starlink — SpaceX's own satellite internet constellation, which now brings in well over $10 billion a year from more than 8 million subscribers and has quietly become the part of the business that funds everything else.

Third, government contracts — NASA pays SpaceX to ferry astronauts to the International Space Station and the DoD pays for national security launches.

The secret sauce is reusability. Before SpaceX, every rocket was used once and thrown into the ocean.

SpaceX figured out how to land the first stage booster back on Earth and fly it again. A single Falcon 9 booster has flown over 20 times.

That's like the difference between throwing away an airplane after every flight versus keeping it for decades.

Magic AI

Magic is building an AI that works like a software engineer, not a chatbot. The goal is a model you can hand a whole code base and a task, and it writes, reviews, and debugs like a teammate.

Their edge is memory. In 2024 they showed off LTM-2-mini, a model with a 100 million token context window.

To put that in perspective, that is about 10 million lines of code the model can look at in one shot, by far the biggest of any model at the time. The plan is to sell this as an AI coworker to companies that write a lot of software.

As of the 2024 rounds, the product was still not for sale.

SpaceX

In 2001 Elon Musk was already rich from selling Zip2, and the eBay deal that would hand him roughly $180 million after taxes for his PayPal stake was still a year away. Most people in that position buy an island.

Musk decided to buy rockets. His original idea was even weirder — he wanted to send a small greenhouse to Mars called "Mars Oasis" to reignite public interest in space exploration.

He flew to Russia three times to buy refurbished ICBMs. The Russians kept raising the price and at one point literally spat on him.

On the flight home from that last failed Russia trip, Musk opened a spreadsheet and started calculating the raw material costs of building a rocket from scratch. He realized the materials were only about 3% of the typical price of a rocket.

The rest was markup, inefficiency, and monopoly pricing by companies like Boeing and Lockheed Martin. He decided to build his own.

SpaceX was founded in June 2002 in a warehouse in El Segundo, California. Musk put in $100 million of his own money.

He hired Tom Mueller, a legendary rocket propulsion engineer who had been building rocket engines in his garage as a hobby. The first rocket, Falcon 1, was supposed to be the cheapest orbital rocket ever built.

It took six years and three spectacular explosions before it finally worked.

Magic AI

Eric Steinberger and Sebastian De Ro started Magic in 2022. Steinberger is an AI researcher who became convinced that the missing piece for a true AI software engineer was memory.

Chatbots forget what you told them a few thousand words ago. He wanted a model that could read your entire code base and remember all of it.

They raised a small seed in 2022, then a $23 million Series A in early 2023 led by Alphabet's growth fund CapitalG. From day one the pitch was not a chatbot.

It was an AI colleague that could handle real engineering work.

SpaceX

SpaceX's growth strategy was simple: be cheaper than everyone, then be better than everyone, then be the only option.

They started by undercutting the launch market. The United Launch Alliance (Boeing + Lockheed Martin joint venture) was charging $300-400 million per launch.

SpaceX offered $67 million. Government agencies and commercial satellite companies started lining up.

Reusability was the real game-changer. Landing a rocket booster looked like science fiction when SpaceX first attempted it in 2013.

They failed over and over — spectacular ocean landings, explosions on drone ships, near-misses. But in December 2015, a Falcon 9 first stage landed back at Cape Canaveral.

It was the first time an orbital-class rocket had ever landed after a mission. Now they do it routinely — it's almost boring.

Starlink created a completely new revenue stream. Instead of just launching other people's satellites, SpaceX launched thousands of its own.

Starlink passed 8 million subscribers and well over $10 billion in annual revenue. It turned SpaceX from a launch company into a telecom company.

Magic AI

Magic's strategy is research first, product later, which is both unusual and expensive. Instead of shipping an app and iterating in public, they raised huge sums to chase a technical breakthrough in model memory.

The headline move was the 100 million token context claim, which got the whole AI world talking. They also cut a deal with Google Cloud in 2024 to build two supercomputers, called Magic-G4 and Magic-G5, using NVIDIA chips, to train bigger models.

Star backers like Eric Schmidt served as validation. The bet is that if the tech is far enough ahead, the product and the customers follow.

SpaceX

The early days nearly killed the company. SpaceX's first three Falcon 1 launches all failed.

The first one in 2006 crashed 25 seconds after liftoff due to a corroded fuel line nut. The second in 2007 reached space but the second stage shut down early.

The third in 2008 failed because the first and second stages collided during separation. Musk had enough money for one more attempt.

If flight four failed, SpaceX was dead.

Flight four worked. On September 28, 2008, Falcon 1 became the first privately developed liquid-fuel rocket to reach orbit.

Musk has said he was so stressed during that period he was throwing up regularly.

The financial pressure was existential. Musk was simultaneously funding Tesla, which was also on the brink of bankruptcy in 2008.

He had to split his last $40 million between the two companies. He borrowed money for rent.

But right at the end of 2008, NASA awarded SpaceX a $1.6 billion contract to resupply the International Space Station. That contract saved the company.

Starship development has been its own saga. The rocket has exploded multiple times during testing.

Each failure costs hundreds of millions. But SpaceX treats failures as data — they move faster by blowing things up and iterating than competitors do by being cautious.

Magic AI

Magic has raised close to half a billion dollars and, years in, still has no product the public can buy. That is the whole problem in one sentence.

A giant context window is a great demo, but rivals like Cursor, GitHub Copilot, and Cognition's Devin already have real users and revenue. Magic has to prove that more memory actually makes a smarter engineer, not just a bigger number on a slide.

And it has to do that while burning enormous sums on compute against competitors who are already shipping. The clock and the cash both keep running.

SpaceX

Falcon 9 is the workhorse — the most-launched rocket in the world. It carries satellites to orbit and astronauts to the ISS, and the first stage lands itself for reuse.

Falcon Heavy is three Falcon 9 boosters strapped together — the most powerful operational rocket in the world until Starship came along. Dragon is the spacecraft that carries astronauts and cargo to the ISS.

It's the only American vehicle currently flying humans to space. Starlink is the satellite internet service — more than 8,000 satellites in orbit delivering broadband to 100+ countries, and the business that funds much of the rest.

Starship is the big one — the tallest and most powerful rocket ever built, designed to carry 100+ people to Mars. It has flown repeatedly, including full orbital-profile flights and booster catches.

Magic AI

The centerpiece is Magic's LTM, or Long-Term Memory, model family. In 2024 they unveiled LTM-2-mini, a model with a 100 million token context window, pitched as the key to an AI that can understand an entire code base at once.

The broader product is an AI software engineer, or coworker, that can write and change code across a whole project. The company also announced Magic-G4 and Magic-G5, two supercomputers built with Google Cloud and NVIDIA to train these models.

Much of it is still research rather than a shipping product.

SpaceX

Founders Fund, Draper Fisher Jurvetson, Google, Fidelity Investments, Valor Equity Partners, Baillie Gifford, a16z (Andreessen Horowitz), NASA (as customer/partner)

Magic AI

Magic's backers are a murderers' row of AI insiders. It started with a $5 million seed round in 2022 and a $23 million Series A in early 2023 led by Alphabet's CapitalG.

Then in February 2024 came a $117 million round led by Nat Friedman and Daniel Gross, two of the most connected investors in AI. The blowout was August 2024.

Magic raised $320 million with checks from former Google CEO Eric Schmidt, CapitalG again, Atlassian, Jane Street, Sequoia, and Elad Gil. That pushed total funding to around $465 million, all for a company still building toward its first sellable product.