How to Find Recently Funded Companies Free With SEC Form D

Every list of recently funded companies you can buy is downstream of a public document. Funding databases, "newly funded startups" newsletters, the lists agencies sell you: a large share of what they know starts life as a regulatory filing anyone can read for free. I spent years as a data scientist before I started building revenue systems, and this pattern repeats everywhere in GTM data: the paid product is mostly packaging around a public source.

This guide goes to the source. You will learn what SEC Form D is, the three free ways to pull it, which fields on the filing matter for prospecting, how to filter a raw feed that is mostly noise into a tight list of operating companies, and how to turn that list into qualified pipeline. We built this exact pipeline for our own outbound at TrueAdvertize, so every step below is something we run, not something we read about.

Timing beats copy. A mediocre email that lands the week a founder is planning how to deploy new capital will outperform a brilliant email that lands during a hiring freeze. Recently funded companies sit in a short, predictable window where three things are true at once:

• They have budget that is explicitly allocated for growth. A raise comes with a plan and a board that expects the plan to be executed. Spending decisions that took quarters now take weeks.
• They are hiring, which means they are building process. New sales hires, new tools, new systems. A company adding its first go-to-market roles is actively deciding how revenue will work.
• The pressure is on. Runway is a clock. Founders who just raised are specifically the people who cannot afford another year of "the product will sell itself." If you sell anything that compresses time to revenue, the belief you are working against, that a great product grows on its own, is weakest in the months right after a raise.

The problem with acting on this window has never been the logic. It has been access. Crunchbase charges for it. Sales databases gate funding filters behind enterprise tiers. Newsletters curate a fraction and share it with thousands of your competitors on the same morning. The primary source, meanwhile, sits in public view with almost nobody in GTM reading it.

Form D is a notice, not an application. When a US company sells securities without registering a public offering, it usually relies on an exemption under Regulation D, and the rules require it to notify the SEC by filing Form D within 15 calendar days after the first sale. That covers the everyday instruments of private fundraising: priced equity rounds, SAFEs, convertible notes, private placements under rules 504, 506(b), and 506(c).

Two properties make this filing unusually valuable for prospecting:

1. It is fast. Fifteen days from first sale is often earlier than the press release, when there is a press release at all. Databases that wait for announcements are structurally behind the filing.
2. It is public by design. Electronic filing has been mandatory since March 2009, so the full history sits in EDGAR, the SEC's disclosure system, readable by anyone.

The filing is short: identity of the issuer, where it is incorporated, an industry classification, the executives and directors, how much is being raised, how much has been sold, the date of the first sale, and a few more fields. No pitch deck, no cap table, no investor names beyond related persons. But for building a list of recently funded companies, short is enough. The filing tells you who raised, roughly how much, exactly when, and which humans run the company. That is a cold list with the two hardest columns already filled in: the trigger event and the decision makers.

Same data, three access patterns. Pick the door that matches how much engineering you want to do.

Door one: EDGAR full-text search. Go to EDGAR search, filter by form type "D", and set a date range. This is the zero-setup option: you can see this week's filings in your browser in under a minute. Good for getting a feel for the data and for spot checks. Bad for volume, because you are paging through results one screen at a time.

Door two: the daily and quarterly indexes. EDGAR publishes index files of every filing received, which automation can walk to pull each new Form D as it lands. This is the door for engineers who want fresh filings on a daily rhythm. One practical detail that stops most first attempts: sec.gov rejects requests that do not declare who is asking. Send a User-Agent header that identifies you and includes a contact email, exactly as the SEC's automated-access guidance asks, or you will see nothing but 403 errors. We learned that one by hitting the wall ourselves.

Door three: the quarterly structured data sets. The SEC's economists publish Form D data sets as downloadable bundles: every filing for the quarter, already parsed into structured tables covering the issuer, the offering, and the related persons. No scraping, no parsing XML, just load and query. If you want history or volume, start here. A full quarter of national fundraising activity fits in a spreadsheet-sized download.

Here is what actually happens when you open door three, so the first download does not surprise you. Each quarter ships as one zip archive named by period, in the pattern of 2026q2_d. Inside are tab-separated tables, not one flat file, and three of them carry the prospecting weight. FORMDSUBMISSION holds one row per filing: the accession number that identifies it, the filing date, and the submission type that tells you whether you are looking at an original D or a D/A amendment. ISSUERS holds the company identity, and it can hold more rows than there are filings because a raise with co-issuers lists each entity. OFFERING holds the money fields: TOTALOFFERINGAMOUNT, TOTALAMOUNTSOLD, the date of first sale, and the industry group you will filter on.

Everything joins on the accession number, so the working shape is one query or spreadsheet merge away: filings joined to their issuer and their offering, filtered by submission type and date. Two format details save you an hour of confusion. Dates arrive as text in DD-MON-YYYY form, 08-APR-2026 style, so parse them deliberately instead of letting a spreadsheet guess. And the amount columns are not purely numeric, because Indefinite is a lawful value in TOTALOFFERINGAMOUNT; force the column to numbers and you silently corrupt exactly the open-ended raises you might care about. One quarter of this, loaded and joined, is a national funding census on your own disk.

If you later outgrow all three and want real-time queries without maintaining your own pipeline, commercial layers such as sec-api.io index new filings within seconds and expose every field as an API. That is a convenience purchase, not a data purchase. The data itself never stops being free.

A Form D filing has dozens of fields. Nine of them do the work.

The related-persons table deserves emphasis. Every filing names the people who run the company, on a document they signed. No enrichment vendor, no guessing which "VP Growth" on LinkedIn actually has authority. For early-stage companies, the related persons are usually the founders themselves, which for most B2B sellers is exactly the buyer.

The "Indefinite" offering amount is worth a note because it trips up first-time parsers. Issuers running open-ended raises can lawfully state an indefinite total. If your pipeline treats that as malformed data and drops the row, you silently lose real companies. Validate against what the form allows, not against what you expected the data to look like.

The raw feed will disappoint you before it pays you. Pull a quarter of Form D filings and the majority of rows are not startups at all: they are hedge funds, venture funds, and private equity vehicles raising from limited partners under the same exemption. Add the amendments, the trusts, and the non-US issuers, and the signal you want is a minority of the file. That is why the filtering is the method. Here is the sequence we use, in order, because each step shrinks the file for the next one:

1. Set the recency window first. Filter on date of first sale, not filing date, and take a trailing window that matches your motion. A 90-day window catches companies still inside the post-raise spending period.
2. Remove pooled investment funds. The industry group field labels them directly. This single filter removes most of the noise in one pass. If the industry group says hedge fund, private equity fund, venture capital fund, or other investment fund, it is not a prospect.
3. Keep operating-company entity types. Corporations and LLCs stay. Trusts and fund-adjacent structures go.
4. Apply geography. If your motion is US-only, drop foreign primary addresses now, before you spend anything on enrichment.
5. Band by offering size. Match the raise size to your deal size. A company that raised $1.5M buys differently from one that raised $40M. Set a floor and a ceiling that describe your actual buyer, and be honest about the ceiling: above a certain raise size, companies hire the problem away instead of buying your product.
6. Deduplicate amendments. Collapse D/A filings onto their original notice so each raise appears once. While you are at it, collapse multiple filings from the same issuer inside your window.

Run those six steps and the file flips from mostly noise to mostly signal: a list of operating companies, in your geography and size band, that sold securities inside your window, each with named officers. As an illustrative example of the shape, a quarter of raw filings measured in the tens of thousands typically reduces to hundreds of in-scope operating companies once the funds, amendments, and out-of-band rows are gone. The exact numbers depend entirely on your filters, which is the point: the list is engineered to your definition of a buyer, not bought off a shelf built to someone else's.

Log what you exclude and why. It sounds like bureaucracy until the first time a good-looking company is missing from your list and you need to know whether that was a decision or a bug. A one-line exclusion record per dropped row turns your list from a black box into a system you can debug and defend. That habit is engineering rigor applied to prospecting, and it is the difference between a list you trust and a CSV you hope about.

A filing is a fact about paperwork. Before a company goes into a sequence, confirm it is a fact about a business:

• Resolve the real website and confirm the company operates. The filing address can be a registered agent or a law firm. Free-first resolution works: search the exact issuer name, match officers' names against the site's team page, and hold anything ambiguous rather than guessing. A wrong-domain hold beats a wrong-person email every time.
• Look for the announcement. Search the company name against funding news. If the raise was covered, the coverage gives you round type, investors, and stated plans, all useful personalization material.
• Notice when there is no announcement. This is the quiet edge of the whole method. A company that filed but never announced is invisible to every database that waits for press. They raised, they are deploying capital, and nobody else's list has them yet. Some companies stay quiet on purpose; respect that in your tone, but understand what it means competitively: the filed-but-unannounced segment is the least contacted list of funded companies that exists.

A list of funded companies is still just a list. What we actually want is the subset in a buying situation, and the raise is only the first of several observable signs. This is where the census becomes a system.

For each surviving company, collect a small set of qualifying observations from public evidence. The pattern we use asks four questions: is the raise verified by a second source; is a founder still personally carrying the commercial motion; is there a dedicated sales or revenue-operations bench already in place, or an open role for one; and is there public evidence of a working outbound function. The answers come from team pages, careers pages, LinkedIn presence, and job postings, all public. Which combination counts as "qualified" depends on what you sell. We look for founder-led motions with no sales bench because that is specifically the situation our work addresses. If you sell recruiting, you would invert the hiring question. The structure carries; the thresholds are yours. Defining that structure precisely is ICP work, and it pays for itself in every later step.

Then enrich the survivors only. Contact data costs money per row, so it comes last, after every free filter has run: a provider waterfall for emails, verification before anything sends, and the officers' names from the filing as your seed so you are enriching a known person rather than guessing a title. Sequencing costs this way, free evidence first and paid enrichment last, is what keeps the whole engine cheap enough to run weekly.

Finally, score and route. Recency of the raise, strength of the qualifying observations, and size band give you a simple priority order, and the top of that order feeds your outbound system at whatever pace your sending infrastructure honestly supports. The point of the system is not blast volume. A tight, signal-based list of companies in a verified buying situation is the input that makes every downstream number better, and it is the difference between outbound that compounds and the experience so many founders describe to us: doing cold outbound but not getting positive replies, because the list underneath was somebody else's stale export.

Fair reading of that table: the paid tools sell convenience, and convenience has a price beyond the invoice. The list they give you is identical to the list they give every other subscriber targeting your market, on the same day. The primary source costs effort instead of money, includes rounds the databases never see, and produces an asset you own: your filters, your history, your definition of a buyer. If outbound is a core motion for you rather than an occasional experiment, owning that layer compounds. If you just need ten funded companies for a one-off campaign, buy the shortcut without guilt.

Every data source earns trust by admitting its edges. Form D has five you should design around:

1. It is not a census of all raises. Regulation D covers most private US fundraising, but not all of it. Other exemptions exist, some issuers file late, and non-US raises without US securities sales never enter the system. Treat the feed as the largest free sample, not the whole population.
2. Addresses lie politely. Registered agents, law firms, and mail drops appear as issuer addresses. Geography filters on the filing alone will misplace some companies; the website and team evidence settle it.
3. Revenue is optional. The revenue-range field allows "Decline to Disclose," and many issuers do. Never build a filter that silently requires a field the form does not require.
4. The filing does not describe the product. Industry groups are broad. "Other Technology" spans everything from dev tools to drone hardware. Product understanding comes from the company's own surfaces, in their language, which is where your personalization should come from anyway.
5. Fifteen days is typical, not guaranteed. Late filings happen. A recency window based on first-sale date absorbs most of this, but the occasional raise will surface later than the rule implies.

None of these edges break the method. They define where the free source ends and your verification layer begins, which is exactly the boundary a well-engineered system makes explicit instead of discovering in production.

We did not write this guide from the outside. TrueAdvertize's own prospecting census is built on this exact source: SEC EDGAR and the quarterly Form D data sets, filtered through the sequence above, cross-checked against independently researched funding announcements, with every exclusion logged and every surviving company qualified against observable evidence before a human writes a word to them. We run it on ourselves because we sell the same discipline: revenue systems built on signals you can verify, owned by the company that runs them.

That last part is the actual position. The engine matters more than any single list. A funded-companies census is one signal source among several that a working revenue engine consumes, alongside your CRM history, your inbound, and the technographic and hiring signals specific to your market. Building that engine so your team owns it, runs it, and improves it without a vendor in the loop is the work. If you are stuck scaling because the pipeline depends on a founder's personal effort, and this guide reads like the kind of machinery you wish existed under your motion, that is the conversation we are built for.

Book a Revenue Engine Diagnostic. 30 minutes, founder-led, no pitch. Bring your current list source and your motion as it stands; we will map where a primary-source census like this one fits, what the rest of your signal layer should be, and whether building an owned engine is the right next step for your stage. If what you have is working, we will tell you that too.