ELON MUSK TAKES SPACEX PUBLIC, IMMEDIATELY BLOWS THE PIGGY BANK ON A $60 BILLION AI CODING STARTUP
Okay so here is what happened. SpaceX went public on June 12. Not a small IPO. Not a boring institutional affair where a handful of pension funds politely divide up shares. The largest IPO in American history. They raised $75 billion. Shares opened at $135 and jumped 19% on day one, making Elon Musk the world’s first official trillionaire. That is with a T. The stock started trading on Nasdaq under the ticker SPCX and the market just absolutely lost its mind over it.
Then, two trading days later, before the confetti had even been swept off the Nasdaq floor, SpaceX announced it was buying Cursor.
If you have not heard of Cursor, it is an AI coding tool built by a four-person team out of MIT called Anysphere. Developers use it to write code faster, debug things, understand unfamiliar codebases. It became one of the fastest-growing developer tools in history. People who use it tend to become evangelists about it. The kind of people who talk about it at parties and have to catch themselves.
SpaceX had quietly negotiated a call option earlier this year, back in April, giving it the right to buy Cursor within 30 days of its IPO. Turns out they only needed two trading days to decide. The price: $60 billion. In stock. The fresh IPO stock that had just surged 19%. So in a very literal sense, Elon used the momentum of his biggest financial moment ever to immediately acquire the biggest VC-backed startup ever acquired in human history.
The four MIT co-founders of Cursor are now billionaires. SpaceX gets the coding talent and the product that even some OpenAI employees apparently use internally. And Musk signals very loudly that his AI ambitions did not stop with xAI and Grok. He wants a piece of the developer tools market too, the market that right now is probably the single most valuable beachhead in enterprise AI.
The deal also tells you something about how public markets are functioning as an acquisition currency in this AI moment. Companies go public, stock rips, they immediately use the inflated shares to buy private companies at eye-watering valuations. It is the same playbook from 1999 except the underlying technology this time actually does something useful. Probably. Mostly. Ask me again in five years.
Sources: CNBC, TechCrunch, Fortune
SALESFORCE BUYS INTERCOM (NOW CALLED FIN) FOR $3.6 BILLION BECAUSE NOBODY HAS ENOUGH AI CUSTOMER SERVICE YET
Quick history lesson. Intercom was one of those companies that basically invented the little chat bubble in the corner of every website. You know the one. It pops up and says “Hi, I’m Alex, how can I help you today?” and you type your question and never get a useful answer. Intercom built that. They made a lot of money from that. Then AI arrived and they pivoted hard, rebranded the product as Fin, and now their main pitch is that the AI agent can actually resolve your questions without a human on the other end.
Salesforce looked at that and said, yes, we want that. They are paying $3.6 billion to get it.
This is not some random acquisition. Salesforce is in a full sprint right now. They have something called Agentforce, which is their attempt to own the agentic AI market for enterprise companies. The idea is that businesses buy Agentforce and use it to build custom AI agents that handle customer interactions, sales workflows, internal processes, whatever they need. The Fin acquisition is meant to supercharge that. Fin already has an AI model called Apex that handles chat, email, WhatsApp, SMS, phone calls, Slack messages. Add that infrastructure to Agentforce and suddenly Salesforce has a much more credible story when they walk into a boardroom and say your customer service team can be five times smaller.
Salesforce paid $3.6 billion for this. Which is either a bargain or a lot of money depending on your level of AI optimism. What is notable is that this is Salesforce competing by checkbook. They are not building all of this from scratch. They are buying speed. Intercom / Fin had customers, had live revenue, had a working AI model. Salesforce is absorbing all of that and calling it Agentforce.
The acquisition closes in fiscal Q4 2027. The customer service industry is watching very carefully. And the humans currently doing customer service jobs are watching even more carefully, but for different reasons.
Source: TechCrunch, CNBC
BASETEN WAS WORTH $5 BILLION IN JANUARY. IT IS NOW APPARENTLY WORTH $13 BILLION. THAT IS FIVE MONTHS APART.
There is a company called Baseten. Most people outside the developer world have never heard of it. It does AI inference, which means it runs your AI models fast after you have already trained them. Think of it this way: training an AI model is like building a car in a factory. Inference is actually driving the car on the road. Baseten is the company that makes driving the car very fast and cheap.
Here is the thing about inference: as AI moves from experiments to actual products, inference becomes the most important piece of the stack. Every time someone talks to ChatGPT, queries Claude, runs an AI agent, there is an inference provider somewhere making that happen. The better and cheaper you can run inference, the more the entire AI industry flows through you. Baseten figured this out earlier than most.
In January of this year, Baseten raised $300 million at a $5 billion valuation. That was considered a big deal. It was nine months after they had raised $150 million at a $2.5 billion valuation. So the valuation was already growing fast.
Now, five months later, they are reportedly closing a $1.5 billion round at a $13 billion valuation. That is a 160% jump in valuation in five months. The round is co-led by Spark Capital, Sands Capital, Altimeter Capital, and Wellington Management. There is also a split-price structure where some investors are coming in at $11 billion and some at $13 billion, which you only see in rounds where everyone wants in but the company has leverage over timing.
If you are keeping score at home: $2.5 billion, then $5 billion, now $13 billion. All within about a year. Baseten has not changed what it does. The market around it has just gone completely insane in the most productive possible way. The demand for reliable, fast, affordable AI inference is growing faster than anyone expected. Every new model, every new AI product launch, every enterprise that decides to actually deploy AI at scale, all of that runs through inference. Baseten is betting it will run through them specifically.
At $13 billion, they are now valued more than some airlines. For a company most people have never heard of. Welcome to 2026.
Source: TechCrunch
CHINA’S HOTTEST VIDEO AI COMPANY WANTS $2 BILLION FROM AMERICAN INVESTORS BEFORE IT GOES PUBLIC
Kling AI is the AI video generator from Kuaishou, which is the Chinese short-video platform that most Americans have never used but which is enormous in China. And Kling is very good. Not “kind of impressive” good. Actually good. It generates video from text prompts that is smooth, coherent, temporally consistent, which is the technical way of saying the people in the video do not suddenly grow extra arms or have their faces melt. It has been competing directly with Sora, which is OpenAI’s video generator, and by many accounts doing it credibly.
Kuaishou has decided to spin Kling out as its own entity and raise serious outside capital before taking it public. The target is $2 billion at a post-investment valuation of $18 billion. They wanted $20 billion originally, shopped it around, found the market’s appetite was a notch below that, and trimmed expectations accordingly. That is actually a sign of discipline you do not always see in this market.
The lead investor they are talking to is General Atlantic, the US growth equity firm. Having an American institution anchor the round is not accidental. Kling wants credibility with Western investors ahead of a potential IPO that could happen in the US or Hong Kong or both. Getting General Atlantic on board tells that story.
There is also a geopolitical subplot here that is impossible to ignore. Chinese AI companies have been walking a careful line trying to attract Western capital while navigating the political environment around AI and China. Kling’s video generation capability is impressive enough that pure performance arguments can win. But the regulatory and political questions around Chinese AI companies accessing Western capital markets are very much unresolved.
If the round closes at $18 billion, Kling becomes one of the most valuable pure-play video AI companies in the world. The Chinese AI industry, which most people assumed was permanently behind after the US chip export controls, is not acting like it is behind. It is acting like it is in the race. Whether Kling’s IPO actually gets done, and at what price, will tell us a lot about how the market is pricing Chinese AI risk right now.
Source: Bloomberg
THE SMART MONEY IS STARTING TO USE THE B-WORD. AS IN BUBBLE.
Here is a different kind of story to close out the week. Not a funding announcement. Not an acquisition. Just a notable signal that some of the most sophisticated investors in the world are starting to publicly worry about what all of this looks like from the outside.
Fortune ran a piece this week talking to the so-called Tiger Cubs. These are the hedge funds that grew out of Tiger Management, the legendary fund run by Julian Robertson. The Tiger Cubs include some of the best long-short equity investors in the game. People who have made enormous amounts of money over decades by being right about individual stocks before the broader market caught on. These are not permabears. These are not people who reflexively distrust new technology. These are people who have been buying AI stocks and making money on it.
And they are starting to get nervous.
The concern is not that AI is fake or that it does not work. The concern is valuation. The concern is that the capital flowing into this space is now so enormous, the expectations baked into these prices are now so astronomical, that the companies would need to grow at rates no company in history has ever sustained to justify where they are trading. SpaceX went public at a $1.75 trillion valuation, which requires growth at a rate that no company has ever achieved. Anthropic is worth nearly $1 trillion. OpenAI is worth $852 billion. Baseten just went from $5 billion to $13 billion in five months. And on and on.
The question is not whether AI will be transformative. Pretty much everyone agrees it will be. The question is whether the current prices assume more than even a genuinely transformative technology can deliver. Whether the capital has gotten so far ahead of the revenue that a correction becomes inevitable even if the technology works exactly as advertised.
The Tiger Cubs are not calling a crash. They are just noting that the math gets very uncomfortable very quickly at these valuations. And they are being more careful. That is a data point worth having in your pocket as you watch $60 billion deals get announced two days after IPOs and inference startups triple in value in five months.
The party is good. Possibly great. The smart guests are just quietly checking where the exits are.
Source: Fortune