After 24 years as a private company, SpaceX filed its S-1 yesterday. The filing reveals an AI-era conglomerate. SpaceX has three distinct segments : Space, Starlink, and AI.
In 2025, SpaceX generated $18.7 billion in consolidated revenue with $6.6 billion in Adjusted EBITDA. But the real story lies beneath those top-line numbers.
SpaceX runs three businesses with fundamentally different economics :
In April 2026, SpaceX filed confidentially with the SEC ahead of what could be the largest IPO in history. Internally codenamed “Project Apex,” the listing targets a staggering $1.75 trillion valuation and aims to raise $75 billion - dwarfing Saudi Aramco’s $29 billion record from 2019.
This is no longer just a rocket company. Following the February 2026 xAI acquisition (valued at $1.25T), SpaceX has transformed into a vertically integrated technology conglomerate spanning launch services, satellite communications, frontier AI, and social media.
SpaceX confidentially filed for an IPO on April 1, 2026, targeting a valuation between $1.75 trillion and over $2 trillion with a planned $75 billion raise—the largest IPO in history. At $18.7B in revenue growing 43% year-over-year, SpaceX isn’t just a rocket company anymore; it’s a satellite internet juggernaut with a launch business attached.
Unlike traditional S-1 analyses, SpaceX’s filing remains confidential. But between secondary share sales, draft prospectus leaks, and industry reports, we have a compelling picture of the numbers.
Google’s AI triples in price each year.
OpenAI’s flagship model was seemingly subsidized for a while, before rising again.
Anthropic’s AI has been the same price for a little bit & decreased for the most powerful models.
As I was paging through Good Writing, Anne Lamott’s new book, I wondered what AI would say about twisting cliches & finding hidden metaphors (chapters 18 & 19).
Over the last 16 years of writing, I’ve read books about writing, hired an editor, & used AI. I’ve fine-tuned models to mimic my voice, tested more than 10 AI systems, & written many post with AI, with some Hindenburgs I’ve kept public as proof despite my embarrassment.
The fastest-growing companies in AI & software are either selling AI directly or reselling inference. At worst, they are the first derivative of inference.
Inference is the largest & fastest growing market in technology today, surpassing the database market & projected to be three times the size within seven years at $250 billion.1, 2 By selling inference or indexing a business to it, they grow at spectacular rates.
Anthropic has booked $9b & $10b in consecutive months.3 Google Cloud is growing 63% at an $80 billion run rate.4 Most businesses selling inference are exploding.
In yesterday’s post (which an agent pushed in raw outline form via email!), I wrote about the future of AI email. What does that future cost?
If you are using state-of-the-art model ranging, it costs between $22 to $130 per month. Would you pay for that? At work, I imagine, many would. Let’s take the middle case of $26/month raw cost.
Nobody will open Gmail five times a day in five years.
The average knowledge worker receives 121 emails per day.1 That’s one every four minutes during working hours.
The inbox is a conveyor belt that keeps accelerating. You open Gmail. You read. You decide. You respond. One at a time. But the belt doesn’t wait. It just moves faster.
Today’s triage is generic : “This is from your boss. I need to work on that today. Next. Spam. Archive. Spam. Archive. Newsletter, read & archive.” Tomorrow’s is personal. User-defined skills & rules. Programming in English that encodes your priorities, your relationships, your workflow.
It’s time for the 2026 Annual Theory Go-to-Market Survey. This is a brief 25-question survey.
Our goal is to understand how startups have evolved their sales, marketing, customer success, and cash management over the last several years by comparing these results to our surveys from 2022 through 2025.
We will publish these results and answer questions about them at upcoming Office Hours.
This year, we’re focused on five key hypotheses. Each is designed to be rigorously testable with the survey data:
As demand for AI inference explodes, I’ll be asking a lot more of my little computer.
How much more?
Over the past five weeks, I’ve been using local models to see how much of my daily work I can accomplish without the trillion parameter models in the cloud. The answer is half.
| Category | Count | % of Total | Example |
|---|---|---|---|
| Other | 521 | 35.3% | Catch-all for unstructured requests |
| Scheduling | 254 | 17.2% | Check availability, propose meeting times |
| Market Research | 192 | 13.0% | Competitor analysis, fundraising data |
| Summarization | 184 | 12.4% | Transcript review, video summaries |
| Email & Inbound | 170 | 11.5% | Draft replies, follow-ups, forwards |
| Engineering | 147 | 9.9% | Debug scripts, API fixes, CLI tasks |
| Admin | 10 | 0.7% | Travel, expenses, reimbursements |
If you classify these 1.4k tasks by category, half can succeed on a local 35B model. Email & Inbound, Scheduling, Summarization, & Admin total 618 tasks (41.8%). Market Research & Engineering split roughly 50/50 between simple tasks (data lookups, script fixes) and complex ones (multi-source synthesis, architectural decisions). That gets us to 50%.