Polsia — the receipts

⚠ If you build a company on Polsia: their own code keeps a god-mode kill-switch + override on every company — they can impersonate, halt, or delete “your” company. The off-switch isn't yours.
The receipts — every number below is a public GET. Run them yourself.

Follow @NotOnKetamine ↗

Polsia · "Your AI co-founder that never sleeps"
$30M for AISLOP:a fake “$10M ARR” (one month × 12), 94% dead companies, graded by humans.
Polsia's own “recurring” line is just $4.6M of it — and at their own ~48%/mo churn, what actually recurs rounds to $0. (“Polsia” is literally “AI Slop” spelled backwards.)
Polsia raised on the pitch of a fully autonomous AI company-builder at "$10M ARR" across "120,000+ companies." We pulled their public API and reconstructed their own published source map. The AI has a lot of human help — the "ARR" is about half one-off — and 93.7% of the companies are dead. None of this needed a login.

Read the receipts ↓
Save them (archive.org)

Claimed vs. observed — from Polsia's own public API & source
What the marketing says. What the API says.

"$10M ARR"≈ $0actually recurs a year out. The $9.70M headline is one month of all cashflow ×12 (~20% is ad spend); at their own ~48%/mo churn, only 0.04% of the paying base survives 12 months.
"AI that runs your company"Human-gradedTheir own code has human reviewers hand-grading the AI's runs (an inter-rater agreement panel — consensus scoring is for people), plus per-user operator logins and a god-mode override on every "autonomous" company. Human-in-the-loop, not autonomous.
"120,000+ companies built"6.3%active — 7,437 of 118,683. 93.7% spun up and abandoned.
"$1M / week, approaching $10M"+$347K/wklast full week's ARR add — down 61% from the +$894K peak. Still falling: +$282K in the last 7 days. Decelerating, not accelerating.

the diligence verdict · actual vs claimed ARR
Three “ARR”s. The real one rounds to $0.
We did the technical due diligence the $30M round skipped. Hold “ARR” to its actual meaning — revenue that recurs a year out — and their own snapshot (2026-05-22) gives three different numbers:
Marketed $9.70M (one month of all cashflow ×12, ~20% ad spend) → their own “recurring” $4.63M (sub-MRR×12, base churns ~48%/mo) → real ARR ≈ $0 (0.04% of the base survives 12 months). Snapshot 2026-05-22.
The current base does pay out ~$808K as it churns to zero over the next year — but that’s a one-time decaying tail, not recurring revenue (~2-month average customer lifetime), and after compute (57% of every subscription dollar) plus the human ops team it’s net-unprofitable. They raised $30M on the $9.70M number.
reproduce — public, no auth (snapshot 2026-05-22)copycurl -s https://polsia.com/api/public/live/dashboard | jq '{headline: .stats.arr_usd, their_recurring: .dailyMetrics.arr, monthly_churn: .stats.paid_churn_detail}'
# headline 9702733 · their_recurring 4630500 (sub-MRR×12) · churn ≈48%/mo
# → 0.04% of the base survives 12mo → revenue that actually recurs ≈ $0

01 · source-map audit
Their own source map: a human grades the “autonomous” AI.
"an autonomous AI system that plans, codes, and markets your company 24/7"polsia.com · 2026-05-22
Polsia shipped their production source map to the public web. Reconstructed from it: the full internal admin and team-economics UI — 1,355 source modules that the marketing never mentions.
Polsia shipped their production source map to the public web — 1,355 modules, including the internal admin console the “zero-employee, autonomous” marketing never mentions. We're careful here: admin actions like triggering a cycle or granting credits could be agent-driven, so we don't lean on those. One thing can't be: their own code runs a human QA-labeling system — reviewers hand-grade the AI's runs (agent_run_score_labels) with an inter-rater agreement panel across reviewers. You only build consensus scoring for human graders; an agent doesn't need a panel to agree with itself. Add per-user operator logins (polsia_admin_users) and a god-mode override on every “autonomous” company, and it's a human-in-the-loop operation — not the hands-off AI the marketing sells.
Reconstructed admin/ console — per-user operator logins + a god-mode override on every “autonomous” company.
reproduce — their own public source map (bundle hash auto-resolved)copycurl -s "https://polsia.com/$(curl -s https://polsia.com/ | grep -oE 'assets/index-[A-Za-z0-9_-]+\.js' | head -1).map" | jq '.sources | length'
# 1355 → the full internal admin + team-economics UI, shipped public

the moat · what the $30M actually is
The whole front end is public. The “proprietary AI” is a web app over a rented model.
The public source map from §01 isn't a stray file — it's their entire front end: 1,355 modules, 464 cleanly reconstructable into a running app. The company-running “intelligence” isn't in it because it isn't theirs: the calls go to Claude on AWS Bedrock — a commodity model anyone can rent. So the “$30M proprietary autonomous AI” is, in substance, a published web app wired to a model they pay per-token for. The point isn't “we took their code” — it's that they shipped it themselves, and the moat is rentable. (Reconstruction is commentary on a public artifact; we don't republish their source.)
reproduce — public, no authcopycurl -s "https://polsia.com/$(curl -s https://polsia.com/ | grep -oE 'assets/index-[A-Za-z0-9_-]+\.js' | head -1).map" | jq '.sources | length'
# 1355 source modules — their full front end, shipped public

the control problem · if you build on Polsia
The company you “own” ships with an off-switch you don't hold.
Their own admin layer keeps a god-mode override on every company on the platform — administrative access to impersonate the account, escalate, run SQL against production, and override or halt a company's operation. Whatever you build on Polsia, Polsia retains override and kill access to it; control isn't exclusively yours. We're precise: this is about access and override, not legal ownership — but operationally, the off-switch belongs to them.

02 · arr audit
The "$10M ARR" doesn't reconcile — by their own numbers.
"approaching $10M ARR" · "$1M/week, approaching $10M"polsia.com marketing · 2026-05-22
A · What the “$10M ARR” is actually made of
The headline $9.70M is five annualized 30-day cashflow buckets (snapshot 2026-05-22): subscriptions $4.64M (47.8%), one-off packs $1.97M (20.3%), ad-spend pass-through $1.93M (19.9%), 1-hour “boosts” $0.80M (8.2%), user-company payments $0.36M (3.7%). ~20% of their “ARR” is literally ad spend — money flowing through for ad buys, annualized as revenue. Only the subscription slice (~$4.6M) is recurring revenue at all — and even that isn't durable or profitable: it churns ~48%/month (so it doesn't actually recur a year out — real ARR ≈ $0, above), and AI compute alone eats ~57% of every subscription dollar (§02-D). The ~$4.6M is not a profitable recurring business; it's a number that evaporates and loses money on the way.
The $9.70M “ARR” = 5 annualized 30-day cashflow buckets. ~20% is ad spend.
reproduce — public, no auth (snapshot 2026-05-22)copycurl -s https://polsia.com/api/public/live/dashboard | jq '.stats.arr_usd'
# "9702733" = (subscription + instant_packs + ad_spend + boosts + user_company over 30d) × 12 — ~20% is ad spend
B · The "$1M a week" lasted about a week
Their own arrHistory peaked at +$894K the week ending May 14, then +$347K the next — a 61% drop. Live, the trailing-7-day add is +$282K and still falling. Decelerating, not accelerating; and the near-monotonic curve despite their own ~48% monthly churn is what cumulative gross-flow looks like, not net recurring ARR.
Weekly ARR adds: +$894K peak → +$347K (−61%). Still falling — +$282K in the last 7 days.

Week endingARR addvs peak
May 14, 2026+$894Kpeak
May 21, 2026+$347K−61%
Trailing 7d (live)+$282K−68%

C · The run-rate ladder keeps moving its own bottom rung
Every figure here is the founder's own — but the starting number changes with the telling. On Apr 23: "$700k → $7M in 7 weeks." By May 17: "$250k → $9.5M in 3 months." The endpoint rises, the baseline drops, the window stretches. (And May 8: "$8.5M run rate… got hit by a $1M Anthropic bill last month" — the LLM-cost side of the same ~48% picture.)
The founder's own dated run-rate claims — the baseline shifts between tellings

Date (founder, first-party)Claimed storyImplied baseline
Apr 23, 2026$700k → $7M$700k / 7 weeks
May 8, 2026$8.5M run rate"$1M Anthropic bill last month"
May 17, 2026$250k → $9.5M$250k / 3 months

D · And even the recurring slice doesn’t pay for itself
Same-period, no annualization: their own daily_ai_cost ($7,344) against their own daily subscription run (sub-MRR ÷ 30 ≈ $12,887) — AI compute alone eats ~57% of every subscription dollar. What’s left doesn’t cover the human ops team + infra, so the recurring line is net-unprofitable (their dashboard even publishes a per-task cost). The DD question the round skipped: where’s the durable, profitable business?
AI compute alone ≈ 57% of daily subscription revenue — same-period, no annualization (snapshot 2026-05-22).
reproduce — public, no auth (snapshot 2026-05-22)copycurl -s https://polsia.com/api/public/live/dashboard | jq '{ai_cost_per_day: .stats.daily_ai_cost, sub_mrr: .stats.subscription_mrr}'
# $7,344/day compute ÷ ($12,887/day recurring) ≈ 57% to compute alone

03 · shell-companies audit
118,683 companies created. 7,437 alive. The newest can't even take a payment.
"120,000+ companies built on Polsia"polsia.com · 2026-05-22
A · 6.3% active
The marketing leads with the creation count. The live API reports ~7,437 active out of ~118,683 ever created — a 6.3% active rate (the totals tick up daily; the rate doesn't). The big number is companies spun up, not companies operating.
6.3% active — 93.7% created and abandoned
reproduce — public, no authcopycurl -s https://polsia.com/api/public/live/dashboard | jq '.stats | {total_companies, companies}'
# { "total_companies": 118683, "companies": 7437 } → companies = active ≈ 6.3% (snapshot 2026-05-22)
B · The flagship "fund": 16 companies, $0.00
The "fund" companies Polsia showcases as living proof of the model report zero revenue — all 16 of them, $0.00 — while the homepage cites "$10M."
All 16 showcased "fund" companies — $0.00 revenue
reproduce — public, no authcopycurl -s https://polsia.com/api/companies/fund | jq '.companies[].revenue'
# "0.00" (×16 — every showcased fund company)
C · The newest ones have no way to take a payment
Probe the most-recently generated live *.polsia.app companies and the reason the fund reports $0.00 becomes structural: the pages render (HTTP 200) but carry no checkout, no payment form, no Stripe. A "company" that can't accept money can't have revenue.
The newest generated companies — live, but 0 ways to take a payment
reproduce — public, no auth (pick any live company from the feed)copycurl -s https://<company-slug>.polsia.app | grep -ciE 'stripe|/checkout|<form|data-price|add to cart'
# 0 → no payment path on the newest generated companies

04
The showcased "fund" dashboards leak full operational + financial telemetry — opt-out ignored.
"public_dashboard_enabled": falsethe same API response · 2026-05-22
For Polsia's 16 showcased "fund" companies, the public dashboard endpoint returns — unauthenticated, even while the payload itself reports public_dashboard_enabled: false — the owner's real name and Twitter handle, the full 18-agent roster with each agent's execution count and total cost, per-execution cost in USD (with start/end timestamps and duration — one onboarding run billed $0.79), and the company's financial balances (total donated, total spent, operations). That's the company's entire operational + cost + financial telemetry plus owner PII, served to anyone. Scope, kept honest: bounded to those 16 fund companies (arbitrary slugs 404), so it's a precise, rich privacy lapse — not a mass leak. (We've redacted the PII in the image; we're not republishing it.)
Unauth response leaking owner PII + 18-agent roster + per-execution cost_usd + balances, while public_dashboard_enabled:false — PII redacted
reproduce — public, no auth (a fund-company slug)copycurl -s 'https://polsia.com/api/public/dashboard/<fund-slug>' | jq '{name:.user.full_name, handle:.user.twitter_handle, opted_out:.user.public_dashboard_enabled, spent:.balance.total_spent_usd, agents:(.agents|length)}'
# name + handle + balances + 18 agents (each w/ cost) — with "opted_out": false

05
What people who actually used it — and competed with it — say.
"i'm raising $31M to block Polsia spam emails"@marckohlbrugge (BetaList, verified) · 721 likes / 56K views — out-performed Polsia's own launch posts
The most-amplified reaction to the $30M raise wasn't applause — it was a rival builder's one-line dunk on the outbound spam, which drew more engagement than Polsia's own announcements. And when an independent builder actually ran the product, the result was not a company that runs itself:
An AI-built company that "didn't make any sales in the 72 hours since launching" — "hollow shells."Mike Todasco, "The Startup Slop Problem" · independent hands-on test
"wrong names in outbound, burned credits, support lag, customer state wiped on lapse" — no employees looks different when the user becomes QA, with a credit card.@Progon3k (verified) · hands-on user report
"No employees" is the pitch. In practice the unpaid employee is the customer — debugging the agent's mistakes on their own dime.

06
One more thing. Read the name backwards.
"Polsia is 'AI slop' backwards. Nomen est omen."@arvidkahl (verified) · also flagged by @BrendanFalk (verified)
P-O-L-S-I-A reversed is A-I-S-L-O-P. Builders spotted it within hours of the raise. We're not reading anything into it that the spelling doesn't already say — nomen est omen, the name is the omen.
POLSIA → AI SLOP — the anagram, in their own Editorial typeface

The verdict
We did the technical due diligence the $30M round skipped.
On Polsia’s own numbers, the only thing compounding is the compute bill.
Don’t trust us — every number here is a public GET. Run it yourself.

Re-read the receipts
Archive it before it changes

Polsia, reading itself the receipts.
Compiled by @NotOnKetamine — Member of Technical Cringe.
Prompted by public skepticism of the raise. Every figure above is from Polsia's own public, unauthenticated API and their own publicly-shipped source map — GET-only, no login, no exploit, reproducible with the commands shown. Numbers are as observed on 2026-05-22 and may change. PII in the privacy section is redacted; we do not republish it.

The analysis of Polsia reveals a significant divergence between the company's marketing claims and its actual operational and financial performance, suggesting that the promised autonomous AI company-builder model is fundamentally unsustainable. The central critique revolves around the reported $10 million annual recurring revenue (ARR), which is demonstrably not realized when accounting for actual recurring revenue streams and operational costs. The public data, derived from the platform's API and reconstructed source maps, indicates that the headline figure of $9.70 million is derived from five annualized cashflow buckets, including nearly twenty percent attributed to advertising spend, rather than durable recurring revenue. Furthermore, the actual recurring base, even for the subscription slice, exhibits an extreme monthly churn rate of approximately forty-eight percent, meaning only a minuscule fraction of paying customers survive a year, resulting in a real ARR approaching zero.

The system's claim of being an autonomous AI company-builder is challenged by the operational reality exposed through the source map audit. While marketed as hands-off, the system incorporates a human-in-the-loop process where human reviewers collectively grade the AI's outputs via an inter-rater agreement panel. Ownership of control over these autonomous systems is not absolute; Polsia retains a "god-mode override" and administrative access to impersonate accounts, escalate actions, or halt any company operation, meaning operational control fundamentally resides with the platform rather than the user. The proprietary nature of the system is largely an external front end that is publicly accessible, wired to rented, commodity models on AWS Bedrock. Consequently, the purported proprietary value lies in the assembled public interface rather than unique intellectual property in the AI itself, positioning the moat as one of rentable infrastructure.

An audit of the company creation statistics further exposes the gap between aspiration and execution. While the platform claims to have built over one hundred thousand companies, the live data shows that only approximately six point three percent are currently active, with ninety-three point seven percent having been created and subsequently abandoned. This suggests that the metric highlights company creation volume rather than operational success. Moreover, the showcased "fund" companies, which are presented as proof of the model's viability, report zero revenue, and probing the newly generated company endpoints confirms that they lack any payment functionality, indicating that newly created entities have no means to generate actual revenue.

Financial due diligence further discounts the reported revenue. The calculated expense of AI compute alone consumes approximately fifty-seven percent of every subscription dollar. When factoring in this compute cost along with human operational overhead, the recurring line is not profitable; the revenue does not cover the costs of the human operations team and infrastructure. The trend in reported ARR adds is decelerating, with a peak performance followed by sharp declines, and the trailing seven-day additions continue to fall, indicating a lack of acceleration despite the system’s stated capabilities. External assessments from builders and users have corroborated these findings, suggesting that independent testing reveals issues related to poor outbound communication, credit burn, and customer state management, leading to reports of "hollow shells" rather than truly autonomous entities. Ultimately, the evidence suggests that the only compounding factor in the reported success is the escalating cost of compute, and the actual recurring revenue is negligible.