On April 21, 2026, a retail investor saw "CRWV IPO" trending on X and bought 50 shares at market open. They had never read the S-1. A week earlier, the same investor had ignored a much smaller AI-adjacent IPO because it "looked boring." The boring one was up 34%. The trending one was down 18%. The difference wasn't luck. It was 300 pages of plain text on SEC EDGAR that almost no retail investor reads.
An S-1 is the prenup a company writes before the IPO wedding. Every risk, every concentrated customer, every founder vote-inflation scheme, every revenue trick is in there because the SEC forces the company to disclose it. The problem is that S-1s are 250 to 500 pages of dense legal text, written by lawyers who want to satisfy regulators without scaring away buyers. That is exactly the kind of document Claude AI was built to read.
This guide shows the exact framework: a six-section prompt that turns any S-1 into a readable brief in under ten minutes. It uses two real filings from the last two years, Reddit Inc. (February 2024) and CoreWeave Inc. (March 2025), to demonstrate how the output stays comparable across very different business models. You will learn what Claude does well, where to verify, and one specific mistake that fools nearly every first-time IPO buyer.
What an S-1 Is and Why It Matters
The S-1 is the registration statement a private company files with the US Securities and Exchange Commission when it wants to sell shares to the public for the first time. It is the most information-dense document the company will ever produce about itself. Every subsequent 10-K, 10-Q, and 8-K is thinner and more routine. The S-1 sits in the public record on SEC EDGAR permanently and is free to read.
An S-1 contains: the full business description, three years of audited financials, a risk-factors section written by the company's own lawyers, a full customer and revenue concentration breakdown, executive compensation, insider share allocations and lockup schedules, the use-of-proceeds statement, and the voting-rights structure of the share classes being offered to the public.
Retail investors who skip the S-1 and buy on day-one price action are trading blind. That is not a moral judgment, it is an information-asymmetry problem. The institutional funds allocating into the IPO have already read every page, often twice. Reading the S-1 with Claude's help is the fastest way to close that gap before you commit capital.
Why Claude Changes the Math on S-1 Reading
Three reasons the combination of Claude plus a public S-1 is genuinely useful today, not just theoretically clever.
First, S-1s follow a templated structure set by SEC Regulation S-K. Sections always appear in the same order: Prospectus Summary, Risk Factors, Use of Proceeds, Capitalization, Dilution, Management Discussion and Analysis, Business, Management, Principal Stockholders, Financial Statements. That structural consistency means one well-crafted prompt produces comparable output across any US IPO.
Second, Claude pulls directly from SEC EDGAR when given the CIK number or ticker. EDGAR is a free, authoritative source. Unlike scraping a pay-walled research site, there are no copyright or access issues, and the data is as fresh as the company's last amendment.
Third, comparability across companies is the whole point. Your Reddit read and your CoreWeave read use the same six sections, so you can stack them side by side and see what genuinely differs: revenue concentration, debt load, voting structure, path to profitability. That comparison is where retail investors actually make better decisions.
$804.0M: Reddit's FY2023 revenue, up 21% year-over-year from $666.7M. Net loss narrowed 43% to $(90.8)M.
Claude's six-section S-1 brief for Reddit Inc., generated from the February 2024 filing on SEC EDGAR.
How to Read an S-1 with Claude in Ten Minutes
Open claude.ai in your browser. No special app, no API, no paid tier needed for this workflow. Copy-paste the six-section prompt below, swap in the target company name, and hit send.
Step 1: Pick the Target Company and CIK
Search the company on SEC EDGAR. Note the CIK number (Central Index Key) and the S-1 filing date. For a company that already IPO'd, the S-1 still exists on EDGAR and is worth reading before buying the public shares, especially in the first six months when the lockup is still active.
Step 2: Paste the Six-Section Prompt
Act as a senior IPO analyst. Use web search to pull [COMPANY NAME]'s S-1 registration statement from SEC EDGAR (CIK [XXXX]). Produce a six-section pre-IPO brief: (1) business in plain English, (2) three key financials with year-over-year change and page references, (3) revenue concentration by customer, segment, or geography, (4) top three risk factors de-jargoned with dollar-impact framing, (5) one red flag most retail investors will miss, (6) one-line investment thesis for a long-term holder. Cite EDGAR for every number. Under 400 words.
Step 3: Run the Same Prompt on a Comparable Company
The comparability is where the value compounds. Read Reddit and CoreWeave with the same prompt and you immediately see: Reddit is a mature-ish ad business with a loss-narrowing curve, CoreWeave is a hyper-growth infrastructure play with accelerating losses and one customer that is 62% of revenue. Those differences matter, and they jump off the page only because the output is structured the same way.
Step 4: Verify Three Numbers Against the S-1 Itself
Claude will occasionally hallucinate a specific number. The fix is simple: pick the three numbers that matter most (usually revenue, net income or loss, and the largest single customer concentration) and open the actual S-1 on EDGAR to check them. If all three tie out, trust the rest of the output enough to form a thesis. If any one is wrong, ask Claude to re-verify everything.
Step 5: Read the Red Flag Section Twice
Claude's fifth section consistently surfaces things that don't show up in the press coverage. For Reddit: Class B super-voting shares at ten votes each held by co-founders and early investors, with public buyers getting one vote. For CoreWeave: interest expense of $349.2M on $1.92B of revenue, a debt load that would cripple the company if customer contracts slip. The red flag is usually where your buy-or-pass decision actually sits.
Real Examples from Two 2024 and 2025 IPOs
Reddit Inc. (RDDT), S-1 filed February 22, 2024
Business: Reddit runs roughly 100,000 interest-based forums moderated by unpaid volunteers. It monetizes 73 million daily users by selling targeted ad placements to brands, and a newer line licenses its 19-year archive of human conversation to AI companies for model training. FY2023 revenue was $804.0M versus $666.7M, up 21% year-over-year. Net loss narrowed 43% to $(90.8)M from $(158.6)M. DAUq grew 27% to 73.1 million, while ARPU ticked down 2% to $3.42. Advertising was nearly all of revenue, with concentration outside the United States still small.
Top risks de-jargoned: moderator revolt (unpaid volunteers could walk and tank engagement, directly hitting the $804M ad base), advertiser concentration (losing two or three major brands would measurably hurt quarterly revenue), and Class B super-voting shares giving co-founders and early investors ten votes per share while public buyers get one.
CoreWeave Inc. (CRWV), S-1 filed March 3, 2025, IPO priced at $40 on March 27, 2025
Business: CoreWeave rents Nvidia GPU compute by the hour to AI labs and hyperscalers. Customers sign multi-year take-or-pay contracts, which CoreWeave pledges as collateral to borrow against, buys the GPUs, installs them in leased data centers, and bills usage. FY2024 revenue was $1.92B versus $228.9M, up 737% year-over-year, the fastest revenue ramp in any major US IPO in the last decade. Net loss widened 45% to $(863.4)M from $(593.7)M despite that revenue growth. Interest expense alone was $349.2M.
Top concentration risk: Microsoft represented roughly 62% of revenue. The business is geographically US-centric with UK and EU buildout underway. The red flag most retail investors missed on day one was that interest expense of $349.2M against $1.92B of revenue means 18 cents of every revenue dollar was already going to bondholders before any other cost. That is the whole investment thesis in one number.
Reddit vs CoreWeave: Same Framework, Very Different Answer
- Revenue scale: Reddit $804M FY2023, CoreWeave $1.92B FY2024 (more than double).
- Revenue growth: Reddit +21% YoY, CoreWeave +737% YoY.
- Net loss direction: Reddit narrowing 43%, CoreWeave widening 45%.
- Revenue concentration: Reddit diffuse across thousands of advertisers, CoreWeave 62% from Microsoft alone.
- Business model risk: Reddit depends on unpaid moderators, CoreWeave depends on one customer and a debt stack.
- Voting structure: Reddit Class B super-voting shares at 10 votes, CoreWeave also multi-class, read the S-1.
- Profitability path: Reddit within reach, CoreWeave blocked by interest expense until revenue scales.
- Best fit for reader: Reddit for ad-economy exposure, CoreWeave for leveraged AI-infra beta with tail risk.
62%: CoreWeave's FY2024 revenue from a single customer, Microsoft. $349.2M: interest expense that year.
Side-by-side Reddit vs CoreWeave output from the same six-section prompt. Comparability across companies is the whole point of the framework.
Common Mistakes That Cost You
Mistake 1: Reading the Prospectus Summary Only
The first 20 pages of any S-1 are marketing. They highlight the upside, downplay the risk, and use every growth metric flattering to the company. The real information lives in Risk Factors (pages 30 to 90), MD&A (pages 100 to 140), and the notes to the financial statements (pages 180 onward). Claude's six-section output goes straight to the substance, so you skip the summary and focus on what matters.
Mistake 2: Trusting Claude's Numbers Without Verification
Language models occasionally hallucinate a specific figure. Always open the S-1 PDF and spot-check the three numbers that matter most. For Reddit, that is revenue ($804.0M), DAUq (73.1 million), and ARPU ($3.42). For CoreWeave, revenue ($1.92B), Microsoft concentration (62%), and interest expense ($349.2M). Five minutes of verification saves you from acting on a wrong number.
Mistake 3: Buying on Day One Without Reading the Lockup
Insider lockups typically release 180 days after the IPO. That is when co-founders, employees, and early VCs are free to sell. Many IPOs see a meaningful price dip right around the lockup expiry because supply increases. The S-1 tells you exactly when the lockup ends. Ignore that section and you buy shares the week before your co-invest insiders dump theirs.
Mistake 4: Ignoring the Voting Structure
Dual-class share structures with super-voting founder shares are now standard in US tech IPOs. You are almost always buying a second-class citizen share. That is a legitimate governance choice for some companies, but it matters for long-term returns because entrenched founders can make decisions you disagree with, and you cannot vote them out. The S-1 spells this out; Claude's red-flag section surfaces it.
Mistake 5: Asking Claude for Price Targets
Do not ask an AI to price the IPO. The offering price is set by the underwriters based on institutional demand, and the first-day pop is driven by short-term allocations and retail FOMO. Ask Claude for framing, not for a number. Your job is to decide whether the business described in the S-1 is worth owning at any plausible price, then compare that to where the market is trading it.
Frequently Asked Questions
Is reading an S-1 with Claude legal and compliant?
Yes. S-1 filings are public documents on SEC EDGAR. You can read them directly, summarize them for personal use, and share summaries. Anything Claude pulls is from the same public record a professional analyst uses.
What if the company is not US-listed?
Non-US companies filing ADRs on US exchanges file a Form F-1 instead of S-1. The structure is nearly identical, and the same six-section prompt works. For purely domestic IPOs in other countries, substitute the local prospectus (for example, a Red Herring Prospectus filed with SEBI in India, or a Prospectus Regulation document in the EU).
How current is Claude's S-1 knowledge?
Claude pulls live from SEC EDGAR via web search at prompt time. Once a filing is posted, it is readable within hours. Amendments (S-1/A filings) update the record; always check whether the company has filed an amendment before trusting the earliest version.
Do I still need to read the full S-1 after running the prompt?
You do not need to read all 300 pages, but you should spend 15 minutes on the Risk Factors section and skim the MD&A. Claude's output is the scaffolding; your own reading fills in the nuance the model could not compress.
What to Watch Next
- v Does the IPO pipeline for late 2026 include a fintech or consumer name you want to test this framework on?
- v Does Reddit's next 10-K confirm the narrowing-loss trajectory implied in the S-1, or reverse it?
- v Does CoreWeave's Microsoft concentration (62%) fall below 50% by year-end, or does the business stay single-customer?
- v Does the SEC tighten disclosure rules on AI customer concentration, making the CoreWeave red flag mandatory going forward?
- v Does your personal hit rate on IPO reads improve over the next three filings when you run this framework?
Key Takeaways
- An S-1 is the single highest-signal document a company ever produces about itself, and it is free on SEC EDGAR.
- A six-section prompt (business, financials, concentration, risks, red flag, thesis) turns any S-1 into a comparable brief.
- Reddit's FY2023 S-1 showed $804.0M revenue, narrowing loss, and Class B super-voting share structure worth noting.
- CoreWeave's FY2024 S-1 showed $1.92B revenue, 62% Microsoft concentration, and $349.2M interest expense.
- Always verify the three highest-stakes numbers in the S-1 PDF before acting on any Claude output.
- Read the risk factors and the lockup schedule before buying on day one. Never skip.
- Ask Claude for framing, not for a price target. You still decide the trade.
References
SEC EDGAR full-text search: sec.gov/edgar
Reddit Inc. S-1 filing (CIK 0001713445): sec.gov Reddit filings
CoreWeave Inc. S-1 filing (CIK 0001769628): sec.gov CoreWeave filings
Investopedia primer on S-1 filings: investopedia.com/terms/s/sec-form-s-1
FINRA investor education on IPOs: finra.org