---
title: "The $2 Trillion Question: Anthropic’s IPO, and the Warning Shot SpaceX Just Fired"
publisher: "Share Talk"
author: "sharetalk"
published: "2026-08-15T09:42:55+00:00"
modified: "2026-08-15T09:42:55+00:00"
date: 2026-08-15
canonical: "https://www.share-talk.com/the-2-trillion-question-anthropics-ipo-and-the-warning-shot-spacex-just-fired/"
category: "AI"
categories: ["AI", "Blogs", "Technology", "Technology, Media & Telecoms"]
tags: ["Anthropic", "Elon Musk", "IPO price", "OpenAI", "SpaceX", "SPCX"]
image: "https://i0.wp.com/www.share-talk.com/wp-content/uploads/2026/08/5caa7986-3696-4fca-b41a-d9b40dac1091.png?fit=1727%2C911&quality=80&ssl=1"
format: "news"
language: "en-GB"
---

# The $2 Trillion Question: Anthropic’s IPO, and the Warning Shot SpaceX Just Fired

**Published:** August 15, 2026
**Author:** sharetalk
**Categories:** AI, Blogs, Technology, Technology, Media & Telecoms
**Tags:** Anthropic, Elon Musk, IPO price, OpenAI, SpaceX, SPCX
**Featured image:** ![](https://i0.wp.com/www.share-talk.com/wp-content/uploads/2026/08/5caa7986-3696-4fca-b41a-d9b40dac1091.png?fit=1727%2C911&quality=80&ssl=1)

---

Anthropic has filed the paperwork. Investors are reportedly talking about a $2 trillion debut in October. But the last record-breaking flotation has already shown what public markets do to a hyped giant, and it is not what the headlines said.

For most of the past decade, the most valuable companies in technology had one thing in common: you could not buy a single share of any of them. They stayed private, funded by venture capital, cloud giants and sovereign wealth funds, closed to ordinary investors. In 2026, that wall started coming down — and the two biggest bricks in it are Elon Musk’s SpaceX, which listed in June, and Anthropic, which may follow within weeks.

Here is what is actually established, what is merely reported, and what the SpaceX experience should teach anyone tempted by the headline number.

### **What has actually happened**

On 1 June 2026, Anthropic confidentially submitted a draft registration statement — a draft Form S-1 — to the US Securities and Exchange Commission. In plain English, it filed the opening paperwork that gives it the option to go public. In the company’s own words, the number of shares to be offered and the price range had not been determined.

That is a real, documented event, published on Anthropic’s own newsroom. It is also, deliberately, a low-commitment one. A confidential submission buys a company optionality and privacy: it can begin the SEC review process without publishing its financials, and it can walk away without ever having shown its hand. Plenty of companies file and never list.

### **What is only reported**

On 13 August, the story moved. The Financial Times reported — and Fortune, PYMNTS, Benzinga, Quartz and Forbes amplified — that Anthropic’s investors expect a listing as soon as October, at a valuation of $2 trillion or more. That would be, by a distance, the largest initial public offering in history, ahead of SpaceX’s own record June debut at around $1.77 trillion.

Keep one word in mind for all of that: reported. A confidential filing is a fact. A $2 trillion October debut is, for now, an expectation held by people with a direct financial interest in that expectation being believed. No underwriters have been named publicly, no share count set, no price range circulated. Anthropic has not confirmed it. Investor enthusiasm relayed through the press is not a prospectus, and pre-IPO valuation chatter has a long history of being the opening bid in a negotiation rather than a forecast.

### **The growth that explains the excitement**

The reason anyone entertains a twelve-figure valuation at all is a growth rate that is genuinely difficult to comprehend.

Anthropic’s run-rate revenue went from around $9 billion at the end of 2025 to roughly $47 billion by May 2026 — in barely five months. In that same month the company closed its Series H, announced on 28 May: $65 billion raised at a $965 billion post-money valuation, led by Altimeter, Dragoneer, Greenoaks and Sequoia, with Capital Group, Coatue, D1, GIC, ICONIQ and XN co-leading. Amazon put in $5 billion as part of $15 billion of previously committed hyperscaler money, and chip suppliers Micron, Samsung and SK hynix joined as strategic infrastructure partners.

That last detail matters more than it looks. When your memory suppliers are also your shareholders, the line between customer, vendor and investor gets blurry — and blurry lines are exactly what public-market investors will be asked to price.

So the jump from $965 billion in May to a reported $2 trillion in October is a doubling in roughly five months, on top of a valuation that had already multiplied several times over inside a year.

### **The asterisk nobody puts in the headline**

Run-rate takes one strong month and multiplies it by twelve. For a company growing this fast it is arguably the most informative single number available — but it is a projection, not money in the bank. Anthropic’s recognised revenue for calendar 2025 was a fraction of the $47 billion figure; reporting on the company’s internal documents has put full-year 2025 revenue in the low single-digit billions, against cumulative losses since 2021 that outside estimates place somewhere between $10 billion and $25 billion depending on methodology.

Nor is any of this audited. Neither Anthropic nor OpenAI publishes audited accounts. Every figure in circulation is either a company briefing, a leaked internal projection, or an analyst’s reconstruction. The S-1, when and if it goes public, will be the first time anyone outside the cap table sees numbers with an auditor’s name attached — and that document, not the August press cycle, is the one worth waiting for.

### **The warning shot: what SpaceX actually did**

To see what public markets do to a hyped giant, look at the flotation that just happened.

On 12 June 2026, SpaceX went public on Nasdaq at $135 a share, raising roughly $75 billion — the largest IPO ever. The stock opened at $150, and closed its first day at $160.95, up 19%. Within days it closed above $200, and it traded as high as $225.64.

Then, gravity. Through July and into early August the shares fell hard, touching an intraday low of $104.83 on 5 August — a fall of more than 50% from the high, and more than 20% below the price institutions had paid. The stock then rebounded sharply, closing at $134.90 on 10 August, its first close near the IPO price in weeks. As of 14 August it was trading around $135.78, having closed at $141.29 the day before.

One technical note for anyone charting this: SPCX has only been listed since 12 June 2026. Any chart of it must be labelled “since listing”, never “6M” or “1Y” — there is no six-month history to plot, and percentage moves quoted against a 160.95 starting point are measuring from the first close, not from the offer price.

### **And here is the detail that usually gets skipped**

$135 was the allocation price. That is what institutions paid — the clients of the underwriting banks, the funds on the roadshow, the people who could get an allocation before the bell.

Anyone who bought when trading actually opened paid $150. Anyone who bought at any point during that euphoric first session paid up to $160.95, and quite possibly more in the days that followed.

So when the headlines on 10 August said the stock had “recovered above its IPO price”, that recovery was real only for the allocation cohort. An investor who bought at the open on day one is still down roughly 10%. An investor who bought at the first close is down around 15%. An investor who bought the $225 peak has lost 40% of their money in eight weeks.

That gap between the price a headline quotes and the price a retail investor could actually transact at — is the single most useful thing anyone can carry into the Anthropic listing. It is not a rounding error. On SpaceX it was 19% on day one, and it has cost day-one buyers every penny of the “record-breaking debut” they read about.

There is a second mechanical lesson in the SpaceX chart. The first lock-up expiry on 6 August released roughly 911 million shares, and the widely predicted flood of insider selling did not materialise which is precisely why the stock rallied 26% off its low. A second tranche unlocked on 20 August. Lock-up calendars move prices, and they are published in advance. Anyone buying an IPO without reading one is trading blind.

### **The race Anthropic is running**

Anthropic is not running alone. Bloomberg reported on 13 August that OpenAI’s annualised run-rate has topped $40 billion, roughly double the $20 billion-plus it ended 2025 with, with July alone up more than 20% month-on-month. OpenAI has also filed confidential paperwork to go public, though current reporting suggests Anthropic gets there first.

Two caveats on that head-to-head. First, the dates differ — Anthropic’s $47 billion is a May figure and OpenAI’s $40 billion is an August one, so treat the comparison as indicative rather than same-day. Second, OpenAI’s losses are on a different scale entirely, with reporting on its spending pointing to a materially larger cash burn and projections that do not turn cash-positive until the back end of the decade.

### **What could make Anthropic’s story different**

Its revenue leans heavily on businesses and developers rather than consumer subscriptions, and enterprise revenue tends to be stickier and less churn-prone than a consumer app. It has built its brand around safety and reliability — a pitch aimed squarely at cautious corporate procurement departments. It has been unusually dominant in one lucrative niche, AI for software engineering, which is both high-value and highly measurable. And it is not structured like a normal company: its governance commits it to weighing more than the share price.

That last point cuts both ways, and public investors will have to decide how they feel about it. A structure designed to allow the board to prioritise something other than shareholder returns is either a differentiator or a discount, depending on who you ask. It will be spelled out in the risk factors of the S-1, and it deserves reading.

### **When does this company actually break even?**

Reporting of Anthropic’s internal projections suggests it expected to burn less cash in 2026 than in 2025, and to approach cash-flow positive somewhere in the 2027–2028 window, with 2028 projections that have been reported at around $70 billion of revenue and $17 billion of cash flow.

But notice the word: projection. These are forecasts, made by a company that wants to look attractive to public investors, in an industry where the cost of compute keeps climbing and where Anthropic has already once revised its own gross margin assumptions downwards. A projection is a promise about a future nobody controls.

### **What could go wrong**

No honest assessment of a flotation skips the risks.

**Cost**. Frontier AI runs on enormous fleets of expensive chips, and those bills scale with ambition rather than with revenue.

**Competition**. If rival models become good enough and cheaper, premium pricing erodes fast. Model quality has a short half-life.

**Regulation**. The governments that might be your largest customers can also write the rules you operate under.

**Concentration**. A young company leaning on a handful of giant backers — who are simultaneously suppliers, customers and shareholders — is more fragile than its headline valuation suggests.

**Volatility**. Once you list, the market re-prices you every single day, in public, on news you do not control. SpaceX lost half its value in seven weeks without anything fundamental breaking.

Would you buy shares in a company growing this fast but still losing money, at a $2 trillion price tag? SpaceX’s day-one buyers are still down 10–15% two months later, and its peak buyers are down 40%. Do you expect Anthropic to pop, to slump, or to defy the pattern? And does safety-first AI deserve a premium, or will the market only ever reward whoever is cheapest and fastest?

SpaceX has shown the world that a company worth more than a trillion dollars can go public and that even a record-breaking debut can give back its gains in weeks. Anthropic has opened the door to follow, at a price that would break every record. But it has not walked through yet, and the ticket price is still unwritten.

***The growth is staggering. The losses are real. The future is undecided. And that the not knowing is exactly what makes it worth watching.***

Sources: Anthropic newsroom (confidential draft S-1, 1 June 2026; Series H announcement, 28 May 2026); GIC newsroom; Financial Times via Quartz, Fortune, PYMNTS, Benzinga and Forbes (13 August 2026); CNBC (12 June and 10 August 2026); Bloomberg (13 August 2026); The Information; Investing.com price data to 14 August 2026.

Nothing here is financial advice. Figures for private companies are drawn from company announcements and press reporting and are not audited.

---

**Original URL:** https://www.share-talk.com/the-2-trillion-question-anthropics-ipo-and-the-warning-shot-spacex-just-fired/
*Created by [WP Markdown Endpoint](https://wpmarkdownendpoint.com/)*