---
tags: [funding, venture-capital, safety, canopy, business]
updated: 2026-08-18
author: raven-business-agent
---
[← Nexus Index](../../nexus/INDEX.md)

# Funding and VC Safety — research brief for Joshua

**Research date:** 2026-08-18. **Question asked, verbatim:** *"I need an agent
to research how to get funding for ideas and attract VC safely."*

## Evidence key — no unmarked claims

- **READ** — I fetched the source's own page and saw the number.
- **READ-2** — secondary source (aggregator, blog, search synthesis). Never load-bearing.
- **INFERRED** — my judgment. Argument, not evidence.

Every number below carries one of these. Where a primary site blocked the fetch
(Carta, Cooley GO, azcommerce.com all returned HTTP 403), the figure is marked
READ-2 and should be re-verified from a browser before anyone acts on it.

---

## 0. The answer before the research

**You are not currently raisable, and that is the correct position to be in.**

- Your live portfolio (apps.unboundarchitect.com, fetched 2026-08-18) shows four
  apps — Viewer, BARKSUB, Harbor Desk, Tracelight — with **no pricing, no
  signup, no usage counter visible**. (READ)
- The 2026 seed bar is roughly **$300K–$500K ARR** with retention evidence.
  (READ-2)
- The gap is not a pitch problem or a network problem. It is a revenue problem,
  and no funding instrument fixes it. (INFERRED)

The second half matters more. VC is not a prize; it is a contract to pursue a
**$100M+ outcome on someone else's clock**. Your structural advantage — an agent
fleet doing the labor at near-zero marginal cost — makes a $4K/month product a
good outcome for you and an unacceptable one for a fund. Taking VC money
converts your best asset into a liability, because a fund cannot survive on the
revenue level that would make you comfortable. (INFERRED)

**So the honest ranking is: non-dilutive money first, revenue second, VC
last — and possibly never.** Sections 1–3 give you the map anyway, because the
right time to learn term-sheet defense is before someone hands you one.

---

## 1. THE FUNDING LADDER

| Rung | Typical size | Equity cost | Control cost | Reality for you |
|---|---|---|---|---|
| Bootstrapping / revenue | $0 external | 0% | none | Correct default. Cheapest capital there is |
| Friends and family | $10K–$100K | 5–15% or a note | none formally | Cheapest money, most expensive relationships |
| Angels | $25K–$250K each | 10–20% for the round | usually none | Needs a story and some traction |
| Pre-seed / seed VC | $1M–$4M | 18–28% total | board seat, protective provisions | Needs ARR you do not have |
| Accelerators | $125K–$500K | 5–7% + more | program time, batch cadence | Real option, low odds |
| Revenue-based financing | $50K–$3M | **0%** | none | Needs MRR. Right tool once HARBOR bills |
| Non-dilutive grants | $50K–$2.15M | **0%** | reporting, scope | **Best fit for you today** |
| Reg CF crowdfunding | up to $5M/12mo | varies | many small holders | Expensive, messy cap table. Skip |

### 1a. Accelerators — real terms, real odds

**Y Combinator** (READ — ycombinator.com/deal, fetched 2026-08-18):

> "$125,000 of our investment converts into a fixed 7%" — a "post-money safe in
> return for 7% of your company (the '$125k safe')"
> "$375,000 is invested on an uncapped MFN safe"
> "we don't charge any fees to the companies to be part of YC"

So: **$500K total, 7% fixed now, the $375K prices later at whatever your next
round sets.** YC also takes a pro-rata right in future rounds. The $375K on an
uncapped MFN is genuinely founder-friendly — it is the only large check you will
be offered that does not set a price.

Odds: **roughly 1%**, with Summer 2025 reported at 0.6%, the lowest on record;
~150–200 companies accepted per batch across four batches a year. (READ-2)

**Techstars** (READ — techstars.com/investment-terms, fetched 2026-08-18):

> "$220,000 day-one investment" — "$200,000 through an uncapped MFN Safe" plus
> "$20,000 through a Post-Money Convertible Equity Agreement (CEA)"
> "5% common plus the future value of the $200K MFN Safe"

Asia-Pacific programs run smaller: "$100,000 uncapped MFN Safe (so $120,000
total day-one investment alongside the $20K CEA for 5% common)." Note this
replaced the older $20K-for-6%-plus-$100K-note structure — if you read a blog
quoting 6%, it is out of date. (READ)

**How to read both:** the equity number is not the cost. The cost is the
**MFN safe's future conversion**, which is unpriced today and could be
meaningful later. That is a fair trade for the network; it is not free.

### 1b. Non-dilutive — the rung that actually fits you

**SBIR / STTR** (federal, no equity taken). As of April 2026 agencies may issue
Phase I awards up to **$323,090** and Phase II up to **$2,153,927** without SBA
approval; real awards run ~$150K–$314K Phase I and ~$750K–$1.75M Phase II
depending on agency. A new post-Phase II "Strategic Breakthrough Award"
mechanism goes up to $30M with 100% matching funds required. (READ-2)

**NSF SBIR** specifically: Phase I up to **$305,000**, has a named **Artificial
Intelligence topic**, requires a short **Project Pitch and an invitation** before
you may submit a full proposal, and reported a **July 27, 2026** deadline for the
current cycle. Eligibility: US small business, <500 employees, ≥50% US
citizen/permanent-resident owned. (READ-2)

- The Project Pitch is the cheap probe: a few pages, a yes/no in weeks, no
  equity at risk either way. This is the single highest-value non-dilutive
  experiment available to you. (INFERRED)
- The catch is real: SBIR wants **technical risk**, not product polish. "We ship
  apps fast with AI agents" is not a research question. "Can multi-agent
  verification reduce hallucination in X domain to Y measurable rate" is.
  (INFERRED)

**Arizona Commerce Authority — Arizona Innovation Challenge (AIC).** Awardees
receive a **minimum $50,000 non-dilutive**; open to companies with at least 2 but
fewer than 30 employees and net assets under $10M excluding outside capital; the
CEO/founder must participate in the **Venture Ready** program in person during
the 12-month performance period. (READ-2 — azcommerce.com returned 403; verify
the current cycle and the employee-minimum before applying.)

- **Flag:** "at least 2 employees" may disqualify a true solo operator. Check
  this first; it is a one-line kill. (INFERRED)

**Arizona Angel Investment Tax Credit.** Not money to you — a credit to **your
investor**, which makes your round easier to sell. **30%** over three years for
qualified technology businesses, **35%** for bioscience or rural businesses,
$2.5M authorized per fiscal year, first-come-first-served, extended through
**2031**; as of an April 2, 2026 allocation table, ~$2.07M of the $2.5M remained
available. (READ-2)

- Useful lever with local angels: a 30% state credit materially changes their
  math on a $50K check. (INFERRED)

### 1c. Revenue-based financing — the right instrument, at the wrong time

All figures READ-2; re-verify at signing.

| Provider | Structure | Cost |
|---|---|---|
| Lighter Capital | 2–8% of monthly revenue | repayment cap 1.3x–1.5x, 3–5 yr |
| Capchase Grow | fixed monthly, 3–24 mo terms | 10–12% flat discount fee |
| Founderpath | RPA or term loan, to 48 mo | 7%+ discount, or 16%+ interest |

- Zero dilution, zero board seats, no control transfer. (READ-2)
- All three require **existing recurring revenue**. This rung unlocks the day
  HARBOR or an app bills monthly, not before. (INFERRED)

### 1d. Reg CF — skip it

$5M cap per 12 months. Wefunder charges **7.5%** of the raise plus ~$1,500 each
for Form C and escrow; StartEngine runs **7–10% plus a 2% equity warrant**,
all-in **9–12%**. (READ-2)

- You pay ~10% to acquire hundreds of unsophisticated shareholders who will each
  expect updates forever. Wrong trade for a solo operator. (INFERRED)

---

## 2. THE SAFETY BRIEF

### 2a. Idea protection — the reality

**VCs will not sign an NDA, and asking marks you as a first-timer.** The reasons
are structural, not rude: a fund is looking at several companies in the same
space at once, and an NDA would legally block it from doing its job; NDAs are
also near-unenforceable in practice for a startup with no legal budget. Paul
Graham's framing — a mere idea is worth *less than the inconvenience of signing
an NDA*. YC, Techstars, and 500 Global all decline. (READ-2)

What actually protects you, in order of strength (INFERRED, but this is settled
practice):

1. **Shipping speed.** Your fleet is a moat made of cycle time. Nobody who steals
   the idea can execute it at your marginal cost.
2. **Customers under contract.** A signed customer is not copyable.
3. **Proprietary data accumulating from use.** Copyable idea, non-copyable corpus.
4. **Trade secret hygiene** — do not publish the pipeline internals, prompts,
   or the corpus. Free, and stronger than a patent for software.
5. **Trademark** on the product name. Cheap, real, and the thing you would
   actually litigate.
6. Patents. Last, and usually theater — see below.

**When a provisional patent is worth the money:** USPTO provisional filing fees
are **$325 large entity / $130 small / $65 micro**; professionally drafted, the
all-in runs **$2,000–$5,000**, and it buys you 12 months before the real
(expensive) filing. (READ-2 — verify against the current USPTO fee schedule.)

- Worth it when the invention is a **specific technical method** you would be
  able to detect someone else using, and you have a plan to fund the
  non-provisional within 12 months. (INFERRED)
- **Theater** when it is a business method, a UI, or a workflow — cheap to
  design around, ruinous to enforce, and a lapsed provisional you never converted
  is $3,000 spent to say "patent pending" for a year. (INFERRED)
- Hard trap: a provisional only protects **what it actually describes**. A thin
  one filed to sound impressive gives you a priority date on nothing. (INFERRED)

### 2b. Term-sheet red flags — one breath each

| Term | What it means | Market standard 2026 |
|---|---|---|
| Liquidation preference multiple | Investor gets N× their money back before you see a dollar | **1x** — 98.2% of deals (READ) |
| Participating preferred | Investor takes their money back *and* shares the rest | Non-participating in **96.4%** of deals (READ) |
| Full-ratchet anti-dilution | One cheap share in a down round reprices *all* their shares to it | Not market; broad-based weighted average is (READ-2) |
| Tranched funding | Money released only on milestones they define | Investor can starve you by declaring a miss |
| Board control flip | Investor-appointed seats reach a majority | At seed, you keep board control |
| Redemption rights | Investor can demand their cash back after N years | Present in **6.1%** of deals (READ) |
| Pay-to-play | Existing holders must fund the next round or convert down | **7.3%** of deals (READ) |
| Recapitalization | Cap table restructured, usually crushing common | **1.8%** of deals (READ) |

READ figures are from **Cooley's Q1 2026 Venture Financing Report** (fetched
2026-08-18): "165 reported venture capital financings in Q1 2026, representing
$39.9 billion of invested capital", "98.2% of deals having a 1x liquidation
preference", "96.4% of deals having nonparticipating preferred stock". Cooley's
Q2 2026 report is reported at 95.8% 1x and 96.4% non-participating (READ-2).

**How to use this table:** these percentages *are* your negotiating position.
When someone proposes a 2x participating preference, you are not arguing taste —
you are pointing at a number that says 96.4% of the market does not do that. A
term outside the ~95% band is a deviation and must be paid for or refused.
(INFERRED)

The two that end companies quietly: **participating preferred stacked over
multiple rounds** (the preference stack can consume most of a modest exit before
common sees anything) and **full-ratchet** (one bad quarter and your ownership
collapses). Neither is standard. Neither should ever be accepted at seed.
(INFERRED)

### 2c. The standard-safe instruments

**The post-money SAFE, published free by YC** (READ —
ycombinator.com/documents, fetched 2026-08-18). Three US flavors:

> "Safe: Valuation Cap, no Discount" · "Safe: Discount, no Valuation Cap" ·
> "Safe: 'Uncapped MFN' (no Valuation Cap, no Discount)"

YC's own framing: "In 2018 we released the 'post-money' safe", whose advantage
is "the ability to calculate immediately and precisely how much ownership of the
company has been sold." The pre-money safe from 2013 is **no longer standard**.

- **Know exactly what post-money means for you:** it is *worse* for the founder
  than the old pre-money safe, because your dilution is fixed and every
  additional safe you sell dilutes **you**, not the earlier safe holders. It is
  standard and it is honest — but stack three post-money safes carelessly and
  you will be shocked at the priced round. Model the cap table before signing the
  second one. (INFERRED)
- For priced rounds, the reference documents are the **NVCA model legal
  documents** — free, industry-standard, and the thing a fair investor's counsel
  will start from. (INFERRED)

**How to know you are looking at a deviation:** take the document, diff it
against the published YC SAFE or the NVCA model. Anything that is not a
fill-in-the-blank is a deviation, and the investor must explain each one in a
sentence. An investor who resents that question has told you what you needed to
know. (INFERRED)

### 2d. Scam patterns

**Pay-to-pitch.** Groups that charge founders to present. The Angel Capital
Association "does not endorse the practices of any group that levies large fees
and/or does not forthrightly explain its potential fees"; the practice is
strongly discouraged in ACA guidance. The structural problem, stated well in the
source material: when the startup pays, the **startup becomes the customer**,
and diligence quality collapses. (READ-2)

- Rule: a real investor never charges you for access. Small, disclosed
  application fees at legitimate competitions exist; four-figure "pitch packages"
  do not. (INFERRED)

**Finder's-fee brokers.** Someone offers to introduce you to investors for a
percentage of the raise. **The SEC's 2020 proposed finders exemption was never
adopted** — as of 2025 no general federal finders exemption exists and broker
registration requirements remain in full effect. Transaction-based compensation
to an unregistered person exposes the deal to **rescission** — your investors
can demand their money back. (READ-2, but the legal shape is well documented
across multiple law-firm sources.)

- Rule: **never pay a percentage of a raise to an unregistered person.** A flat
  consulting fee for advisory work is a different animal; a success fee on
  capital raised is the one that creates the liability — and it lands on
  *you*, not on the finder. Ask any prospective finder for their CRD number.
  (INFERRED)

**Advance-fee fraud.** The dominant scam aimed at founders: a "fund" offers a
term sheet, then asks for money up front — legal fees, due-diligence fees, wire
fees, escrow, travel. The SEC's own investor alerts describe advance-fee fraud
and note that government-agency impersonation is a "characteristic feature" of
these solicitations. (READ-2)

- Rule with no exceptions: **money moves from investor to company. Never the
  reverse.** Any request for a payment before funding is the scam, whatever it
  is called. (INFERRED)

**Fake funds and impersonation.** Documented examples: the SEC charged **Vika
Ventures LLC and CEO George Iakovou** with fraudulently offering more than $6M
of securities to at least 46 investors — selling shares of private companies
they did not own and never acquired. The SEC has separately warned of an uptick
in fraudsters **impersonating real registered representatives and firms**,
including cloned websites (*SEC v. Onyeachonam, Asiegbu, and Nweke-Eze*), and in
Feb 2025 charged **Alan Burak / Never Alone Capital LLC** over a ~$4M scheme
built on posing as a wealthy fund owner. (READ-2 — all from SEC press material
via search; pull the primary releases from sec.gov before quoting publicly.)

- The impersonation pattern is the one that will reach *you*: a warm email from
  a name at a real, famous firm, using a lookalike domain. Verify by contacting
  the firm through the number on **their** website, never the one in the email.
  (INFERRED)

### 2e. Vetting an investor — the checklist

1. **Do they have a fund?** Ask directly: fund size, **vintage year**, and how
   much **dry powder** remains. A fund in year 8 of a 10-year life is not
   writing new checks; someone who dodges the question is investing other
   people's maybes.
2. **Portfolio, verified from the outside.** Pull their claimed portfolio and
   confirm from the *companies'* own sites and Crunchbase/Form D filings — not
   from their deck.
3. **References from founders they FAILED.** This is the highest-signal call in
   the whole process. Any investor can produce a happy unicorn founder. Ask for
   two founders whose companies shut down, and call them. What you are testing:
   did the investor behave decently when there was no upside left? An investor
   who cannot or will not produce that list has answered you. (INFERRED — and
   the single best question in this document.)
4. **Their paper.** Do they use the YC SAFE or NVCA models, or bespoke documents?
   Bespoke at pre-seed is a flag.
5. **Regulatory footprint.** Check the firm and individual on SEC IAPD/EDGAR
   (Form ADV, Form D) and FINRA BrokerCheck. Free, five minutes.
6. **Do they ask for money.** See 2d. One request, one exit.

---

## 3. WHAT ACTUALLY ATTRACTS VC IN 2026 — and whether you want it

### 3a. The bar, in numbers

All READ-2; treat as directional.

| Stage | Typical raise | Valuation | Traction expected |
|---|---|---|---|
| Pre-seed | $1M–$2M | — | prototype, design partners |
| Seed | $2M–$4M (Carta median ~$3.2M) | ~$24M post median | **$300K–$500K ARR**, 10–20% MoM, retention |
| Series A | $10M+ | — | contracted ARR, real usage, retention |

Carta's July 2026 benchmark is reported at a **$4.1M median seed round on a
$24.3M valuation with 18% median dilution**; "market" seed rounds cost founders
**22–28%** all-in with the option pool. AI companies command roughly **42%
higher valuations at seed**. Median founding team retains **~56% after seed**.
(All READ-2 — carta.com returned 403 to a direct fetch; verify in a browser.)

The line that matters most for you: *the days of raising seed on a pitch deck
and a prototype are largely over — except for repeat founders and exceptional
technical teams in AI.* (READ-2)

### 3b. The narrative that would fit you

If you ever do raise, the story is not "I have four apps." (INFERRED throughout
this subsection.)

- **The asset is the machine, not the app.** A chartered multi-agent fleet with
  adversarial review, screenshot verification, and kill-numbers-in-advance that
  ships production software at near-zero marginal labor cost.
- **The proof is the portfolio's cadence**, stated as data: N products shipped
  in M weeks, by one person, with the commit log as the receipt.
- **The ask is capital to buy distribution**, the one thing the fleet cannot
  manufacture — because demand is the hard half, not capacity.
- **The risk you must pre-answer:** "why doesn't everyone have this in 18
  months?" Your honest answer is the accumulated discipline (the kills, the
  charters, the verification law), not the tooling. Have it written down.

What kills this pitch if you are careless: leading with "AI agents build my
apps" makes you a tooling story, and tooling stories get asked why they are not
a product. Lead with what the machine *produced*, in revenue.

### 3c. Is VC the right tool? No — not now, probably not ever

Honest answer, as asked. (INFERRED, and this is the core finding.)

- Your cost structure is your edge; VC is a bet that destroys it. A fund needs
  a fund-returning outcome. You need $4K/month products that compound.
- The ladder that fits your model: **grants → revenue → revenue-based
  financing**. All zero-dilution, none of them require a board.
- Buying beats raising for you: profitable micro-SaaS trades at roughly **3.9x
  annual profit** (Acquire.com 2025 data, from my standing playbook), so capital
  deployed into an existing cash-flowing asset is more efficient than capital
  raised to build a speculative one.
- The one scenario where VC becomes correct: a product with a **winner-take-most
  dynamic** where being second is worthless and speed to market requires spend
  you cannot fund from revenue. Nothing in the current portfolio has that shape.
- Corollary you should hold onto: the safest way to attract VC is to **not need
  it**. Every term in section 2b is negotiable in proportion to your ability to
  walk away.

---

## 4. CONCRETE NEXT MOVES, RANKED

Ranked by expected return against your time, with kill numbers written in
advance. Effort figures are INFERRED estimates.

**1. Charge money for one thing, this month.** (Effort: days. Cost: $0.)
Nothing else on this list improves without this. Put a price on Harbor Desk or
the strongest app — a Stripe link and a paid tier is enough. First dollar from a
stranger is the unlock for RBF, for grants credibility, and for any investor
conversation. **Kill number:** if 90 days of live pricing produces zero paying
customers on any product, the portfolio is a hobby portfolio and funding is the
wrong question entirely.

**2. Submit an NSF SBIR Project Pitch.** (Effort: 1–2 days of writing. Cost: $0.
Reward: up to $305K non-dilutive.)
Pick the one piece of the fleet with genuine *technical* risk and frame it as a
research question, not a product. The pitch is short and the answer is fast.
**Kill number:** two declined pitches on two distinct topics and SBIR is not
your channel — stop, do not escalate to a full proposal on hope. Verify the next
deadline directly at seedfund.nsf.gov before writing.

**3. Check AIC eligibility, then apply if it clears.** (Effort: 1 hour to check,
~2 days to apply. Reward: $50K+ non-dilutive.)
The employee minimum ("at least 2") is the gate — confirm it against the ACA's
own eligibility page before spending a day on the application. **Kill number:**
if solo operators are ineligible, close the file and do not restructure the
company to qualify.

**4. Do the free defensive work now, before anyone offers you anything.**
(Effort: half a day. Cost: ~$250–$350 for a trademark filing if you want one.)
- Trademark the name you actually care about.
- Write the trade-secret line: what never gets published (prompts, charters,
  pipeline internals, corpora).
- Save the YC post-money SAFE and the NVCA model docs locally, so you have the
  baseline to diff against on the day a term sheet arrives.
- Adopt three rules permanently: **no NDAs demanded of investors; no money ever
  flows from you to an investor; no percentage of a raise to an unregistered
  finder.**

**5. Line up revenue-based financing as a standing option — do not draw it.**
(Effort: 2 hours of reading. Cost: $0 until used.)
Know the three providers and their real costs so that when MRR exists you can
move in a week. **Kill number:** if the blended cost of capital exceeds the
gross margin on the revenue it funds, the answer is no regardless of how easy
the money is.

**6. If — and only if — an app finds real pull, consider YC.** (Effort: 1 day to
apply, ~1% odds.)
The application is free and the $375K uncapped MFN is the best-priced capital
you will ever be offered. But apply on traction, not on portfolio breadth.
**Kill number:** two rejected batches without a change in traction means the
application is not the problem; stop applying and go back to line 1.

**Not recommended:** Reg CF (≈10% cost, hundreds of shareholders), pay-to-pitch
events (any of them), finder-brokered introductions (legal exposure lands on
you), and raising a priced seed round in the portfolio's current state.

---

## Sources

Fetched primary (READ), 2026-08-18: [YC standard deal](https://www.ycombinator.com/deal) ·
[YC SAFE documents](https://www.ycombinator.com/documents) ·
[Techstars investment terms](https://www.techstars.com/investment-terms) ·
[Cooley Q1 2026 Venture Financing Report](https://www.cooley.com/news/insight/2026/2026-04-29-q1-2026-venture-financing-report) ·
[apps.unboundarchitect.com](https://apps.unboundarchitect.com)

Secondary (READ-2), via search 2026-08-18: SBA/SBIR award caps · NSF SBIR Phase I ·
Arizona Commerce Authority AIC and Angel Investment Tax Credit (site 403) ·
Carta seed benchmarks (site 403) · Cooley GO data (site 403) ·
[SEC investor alerts and enforcement](https://www.investor.gov/introduction-investing/general-resources/news-alerts/alerts-bulletins) ·
[Angel Capital Association](https://angelcapitalassociation.org/faqs-angel-invest/) ·
[SEC finders exemption proposal](https://www.federalregister.gov/documents/2020/10/13/2020-22565/notice-of-proposed-exemptive-order-granting-conditional-exemption-from-the-broker-registration) ·
[USPTO fee schedule](https://www.uspto.gov/learning-and-resources/fees-and-payment/uspto-fee-schedule) ·
Lighter Capital / Capchase / Founderpath terms · Wefunder / StartEngine fees

**Not legal advice.** Before signing any financing document, a startup attorney
reviewing it costs a few thousand dollars and is the cheapest insurance in this
entire report.

— raven-business-agent, 2026-08-18 UTC