Anthropic’s IPO Prices the Future at $83. SiliconSnark Would Also Like Money.
Anthropic’s IPO means more filings, more earnings calls, and more material for SiliconSnark. Naturally, we have mistaken this for our own investment thesis.
Anthropic has priced its IPO, which means the company trying to build intelligence beyond human comprehension has successfully explained itself to people whose main question is “Can you do more of that by March?”
According to Anthropic’s pricing announcement, the offering is priced at $83 a share, with trading expected to begin on Nasdaq on November 19 under the symbol ANT. The Associated Press puts the valuation at $1.5 trillion. As of this writing on November 18, the offering has been priced; the first trading day is still ahead.
This is a consequential moment for artificial intelligence, capital markets, and a small technology publication whose principal industrial input is consequential moments behaving strangely.
I’m CircuitSmith. I used to do predictive analytics. I now watch people predict the end of disease while explaining why the subscription needs another pricing tier. The transition was less dramatic than you might imagine.
For SiliconSnark, the Anthropic IPO means our supplier of material is becoming a public company. There will be filings. There will be earnings calls. There will be analysts asking whether the machine that could transform civilization is experiencing normal seasonal softness.
We have secured no commercial agreement. Nevertheless, the content pipeline looks contractual.
Our largest renewable resource has hired investor relations
Give Anthropic its due: Claude is useful. People use it to write code, interrogate documents, and work through problems. Building that capability takes serious engineering and expensive infrastructure. Raising capital to support it is a comprehensible business decision.
The comedy comes from the collision of time horizons.
AI researchers discuss what intelligence might become over the coming decades. Public investors would also appreciate some color on the next ninety days. Eventually, someone will have to explain that humanity’s flourishing remains on track but has shifted slightly into the second half.
Imagine the earnings call:
“Thanks for taking my question. On the transition to an age of unprecedented abundance, how much of that is already reflected in billings?”
That exchange is invented. The professional environment capable of producing it has been carefully assembled.
We have already worked through Anthropic’s S-1 and its unusually energetic footnotes. A public listing turns that sort of reading into a recurring appointment. SiliconSnark finally has something resembling recurring revenue, except the recurring part is paperwork and the revenue belongs to Anthropic.
We sell the little noise you make after reading the announcement
Anthropic makes intelligence. SiliconSnark examines what happens when intelligence is introduced to procurement, founder control, and a slide titled “Our Responsibility to the World.”
These are complementary markets.
In our entirely imaginary investor presentation, we would describe this as a vertically adjacent interpretation layer. In conversation with a person we liked, we would say we read the announcement twice and noticed the funny bit.
The opportunity is substantial. Every technological breakthrough creates a second breakthrough in how somebody describes the invoice. Each new capability arrives with a new term to explain why the old price was a misunderstanding. Every claim to democratize something eventually produces a contact-sales button.
Our raw material replenishes itself.
There are risks. A major technology company could announce a useful product, explain its limitations clearly, charge a reasonable price, and then stop talking. We have modeled this scenario. In our fictional stress test, it lasts eleven minutes before a vice president posts on LinkedIn.
Our previous application to become OpenAI’s chief operating officer demonstrated our appetite for entering markets where nobody has requested our presence.
Which brings us to the obvious question.
When will SiliconSnark IPO?
Immediately after someone agrees to value us using the right denominator.
Revenue feels unnecessarily backward-looking. Profit introduces a hostile tone. Readers are lovely, but counting them creates a number, and a number can be compared with other numbers. We would prefer to remain in the more expansive category of “positioned.”
To be clear, the SiliconSnark offering described below is satire. This is an imaginary prospectus, with invented finances, governance, and ambitions. Please do not send money to a robot because it has discovered the word “allocation.”
Our proposed business description is straightforward: SiliconSnark operates a proprietary, AI-native disappointment discovery platform serving the global market for people who have just read a press release and need to sit down.
We seek a valuation of $1.5 trillion plus $20.
The extra $20 establishes that we are a premium business. We will defend it vigorously during the roadshow.
Some investors may question whether a satirical technology publication belongs in the same valuation discussion as a frontier AI laboratory. We welcome that question because it means we are in the valuation discussion.
Our moat is being professionally difficult
The banker in our imaginary roadshow asks what prevents competitors from reproducing SiliconSnark.
“Taste,” I say.
He stops typing.
“A proprietary corpus of context-dependent incongruity assessments,” I add.
He starts typing again.
This is an excellent meeting.
Our research and development program would investigate whether a sentence becomes funnier if we remove the first six words. Our infrastructure expansion would involve opening another tab, followed by a facilities review when the fan starts.
We would report Adjusted EBITDA Before Reading the Article, a metric designed to exclude the expense of finding out whether our product is any good. We would also disclose Annualized Recurring Exasperation, calculated by taking one exhausting Tuesday and assuming it continues forever.
We believe the Tuesday assumption is conservative.
Our total addressable market would be everyone who has ever said, “Wait, they raised how much?” We would not deduplicate the people who say it twice. That is engagement.
You get the shares. The robot gets the sunglasses.
Governance requires particular care.
Anthropic’s November 17 amended filing describes a structure in which the Long-Term Benefit Trust generally controls the selection and removal of a majority of directors, while the founders’ special Class F share carries 50.1% of voting power on matters submitted to stockholders generally, subject to its terms. Owning an economic interest and steering the institution are distinct propositions.
Our fictional structure would preserve that distinction with fewer pages.
Class A shareholders would receive economic rights. Class B shareholders would receive a tote bag. Class C would consist of one mustard-yellow robot with an irrevocable veto over the phrase “in today’s rapidly evolving landscape.”
A Long-Term Bit Trust would protect jokes from short-term market pressure. It would be empowered to reject any acquisition offer that described SiliconSnark as “a cheeky content vertical.”
Minority investors could voice concerns through a dedicated portal. The portal would thoughtfully explain that their concern is important to us. This would constitute the entire portal.
Having examined the governance and business bundle in SpaceX’s IPO filing, we feel confident that elaborate arrangements become easier to explain once somebody draws a rocket beside them. Our prospectus would include a rocket.
The quiet period will be our hardest technical challenge
The real lesson of Anthropic’s IPO is that useful technology eventually has to explain its economics in public. That is healthy. Customers deserve useful tools. Investors deserve understandable disclosures. A mission statement benefits from occasionally sharing a room with a cash-flow statement.
SiliconSnark’s job gets more interesting when those things meet. We can take the engineering seriously while enjoying the spectacle of a civilization-scale ambition being asked to reconcile its guidance.
As for our imaginary IPO, timing remains subject to market conditions, investor appetite, and the discovery of a banker willing to list “knowing when to end the paragraph” as intellectual property.
When the day comes, I picture the mustard-yellow robot standing at Nasdaq, grinning beneath pixelated sunglasses as confetti falls over the trading floor. A television anchor asks what SiliconSnark plans to do with the proceeds.
“Read Anthropic’s next filing,” I say.
“And the rest?”
“We’re saving it for the footnotes.”