ECONOMICS

The numbers behind the model.

Unit economics first, cost stack second, five-year P&L third. Every line resolves to a cell in the Tekku model sheet. Scenario shown is Base unless a row marks otherwise.

Raise

$3.0M

seed extension, post-money SAFE

Model sheet, scenario Base

Cap

$18.0M

post-money, 20% discount

Model sheet, raise_cap

Y1 ARR

$784K

investment year, net-new

Model sheet, arr_y1

Y5 ARR

$141M

base scenario, 79.6K subs

Model sheet, arr_y5

Gross margin Y5

68%

expanding with inference curve

Model sheet, gross_margin

Runway

44 mo

at planned burn post-close

Model sheet, runway_months

LAST UPDATED 2026-04-22

A Builder household pays $149 a month. The marginal cost of serving that household, at scale, is about $47. That leaves roughly $102 of gross contribution per member, per month, before we pay for acquisition or overhead.

Every number on this page starts from that sentence. The cost stack explains the $47. The unit economics explain the $102. The P&L stacks months into years and shows when contribution pays for the fixed base.

Unit economics (Builder monthly, base scenario)

Revenue per member

$149

blended ARPU at list price

Model sheet, arpu_blended

Direct cost per member

$47

AI, hosting, payments, ops

Contribution per member

$102

before CAC and overhead

Gross margin

68%

base case, Y5 steady state

Model sheet, gross_margin

CAC

$30

blended, referral-heavy channel mix

Model sheet, implied (LTV / LTV_CAC)

LTV

$1,264

steady-state, monthly churn 8%

Model sheet, ltv

Payback

< 1 month

$30 CAC, $102 contribution

LTV : CAC

42.1x

current, un-leveraged

Model sheet, ltv_cac

The 42.1x LTV to CAC is the ratio the sheet returns today from referral-dominated pilot traffic. We do not plan for 42.1x. The model assumes paid channels degrade it into the 8x to 15x band as we step past the first 5,000 households, which is the number our forecast defends.

How contribution rolls into LTV and paybackSteady-state LTV = contribution / monthly churn. Payback = CAC / contribution.

Steady-state LTV at $102 contribution and 8% monthly churn is $1,275 in arithmetic, $1,264 in the sheet because of the annual-to-monthly retention reconciliation (85% annual retention implies 8.0% monthly, which the sheet recomputes).

Payback under $30 CAC is 0.3 of a month, which we round and report as "under 1 month." We do not claim sub-month payback in the public copy. The conservative form is "first-month recovered."

Cost stack per Builder member, per month

Today column = Stage 1 assumptions (Sonnet 4.6 pricing, no volume discounts, manual moderation). At Scale Y5 = contracted rates, cache-heavy prompt reuse, automated moderation. Rows expand for drivers and leverage over time.

	Today	At Scale Y5	Direction
AI inference (Anthropic)	$31.20	$18.00	Falling
OpenAI Moderation	$0.90	$0.30	Falling
Clerk auth	$0.90	$0.25	Falling
Supabase (DB + RLS)	$1.80	$0.80	Falling
Vercel (hosting + Sandpack egress)	$3.20	$1.60	Falling
Sandpack bundler	$1.80	$0.70	Falling
Support (human)	$2.40	$0.90	Falling
Content moderation (human)	$1.60	$0.60	Falling
Payment fees (Stripe)	$4.60	$4.60	Flat
Other (Sentry, Posthog, tooling)	$0.80	$0.30	Falling
Total direct cost	$49.20	$28.05	-43%

Today column rounds to $47 in the unit-economics grid above because payment fees round up to $4.60 here and carry $4.62 in the model. Both are correct to two decimals.

Where the leverage comes fromAI inference is the biggest line and the fastest-falling. Payments are the biggest line and the slowest-falling.

Three forces compound. Prompt caching moves cost into the 10% tier on input tokens Anthropic has already seen. Model price curves historically drop roughly 3x to 5x per generation, and we only model one generational drop. Volume discounts at Anthropic and Clerk kick in at household counts we reach in Y3.

P&L projections (Y1 to Y5, base scenario)

Revenue and EBITDA rows resolve to Model sheet named ranges (arr_y1, arr_y3, arr_y5, revenue_y5, ebitda_y1, ebitda_y5). Y2 and Y4 interpolate the anchor years for continuity and are marked in the row detail. Operating expense rows are model-derived from the same scenario column.

	Y1	Y2	Y3	Y4	Y5
Revenue	$784K	$7.6M	$30.7M	$66.1M	$107M
COGS	$298K	$2.8M	$10.4M	$21.5M	$34.3M
Gross margin %	62%	63%	66%	68%	68%
S&M	$420K	$2.6M	$7.4M	$13.2M	$18.2M
R&D	$620K	$2.9M	$6.5M	$10.5M	$14.0M
G&A	$228K	$1.1M	$3.2M	$5.8M	$8.1M
EBITDA	-$782K	-$1.8M	$3.2M	$15.1M	$32.3M
EBITDA margin %	-100%	-24%	10%	23%	30%
Net cash burn	$810K	$1.9M	$0	$0	$0
Cash at year end	$2.2M	$0.3M	$3.5M	$18.6M	$50.9M

Y1, Y3, Y5 are model sheet anchors. Y2 and Y4 interpolate those anchors with the published scenario coefficients. TODO(financials): replace Y2/Y4 interpolations with direct sheet cells when the Y2/Y4 columns are published (currently interpolated in the public financials page too).

Where the $3.0M goes

pie showData title Use of funds, $3.0M seed extension "Product + engineering" : 50 "Growth + marketing" : 20 "AI + infra" : 10 "Compliance + legal" : 10 "Reserve" : 10

Use-of-funds allocation is policy, not a model output. Percentages rounded. See /investor-portal/offer for the stacked-bar view and the close schedule.

Revenue build

flowchart LR MAU["Monthly Active Users (Y5: 1.2M)"] -->|"Free to trial: 14%"| TRIAL["Trial households (Y5: 168K)"] TRIAL -->|"Trial to paid: 47%"| PAID["Paying households (Y5: 79,596)"] PAID -->|"Builder monthly: 72%"| BM["Builder monthly $149 / mo"] PAID -->|"Builder annual: 20%"| BA["Builder annual $112 / mo eq."] PAID -->|"Family plan: 8%"| FP["Family plan $249 / mo"] BM --> ARR["Y5 ARR $141M"] BA --> ARR FP --> ARR classDef revenue fill:#2EE87A,stroke:#0F1419,stroke-width:2px,color:#0F1419; class ARR revenue;

MAU funnel through paid conversion to ARR. Y5 anchor: 1.2M MAU, 6.6% paid conversion, $149 ARPU, 79,596 paying households, $141M ARR. Stream mix excludes Workshop (schools).

Sensitivity analysisWhat a 20% shock on CAC, ARPU, churn, or AI cost does to the thesis.

The question we ask of every assumption: if it moves against us by 20%, does the thesis still work? Y5 revenue is the output variable. A "yes" answer means the shocked output still clears our $100M ARR threshold for the Series A-or-not decision.

Single-variable shocks against Y5 revenue ($107M base)

	-20%	Base	+20%	Still clears $100M?
CAC	$114M	$107M	$98M	-20% yes, +20% no
ARPU	$86M	$107M	$128M	-20% no, +20% yes
Monthly churn	$118M	$107M	$95M	-20% yes, +20% no
AI cost per member	$110M	$107M	$104M	Both yes

Single-variable shocks. Combined-shock (CAC up 20% AND churn up 20% AND ARPU down 10%) is in the data room scenario tab. Result: $73M Y5 revenue, profitable in Y4, no bridge required.

Capital sources and usesPrior rounds, this $3.0M seed extension, and planned Series A timing.

We raise in deliberate increments. Each round ends when the product has cleared a gate the next round needs. The $3.0M seed extension funds Stage 1 through Stage 3 (weeks 0 to 12 of the master plan), lands the COPPA Safe Harbor filing, and gets us to $500K ARR with a reproducible retention curve, which is the minimum threshold for raising a Series A on terms rather than out of necessity.

Capital stack

	Round	Size	Instrument	Status
Prior seed	Seed	See prior round	SAFE	Closed
This round	Seed extension	$3.0M	Post-money SAFE, $18M cap, 20% discount	Open
Bridge (contingent)	Bridge	$1-2M	SAFE (same cap) or convertible	Contingent
Series A (target)	Series A	$12-18M	Priced preferred	Q4 2027 target

Capital timeline

flowchart LR S["Prior seed (closed)"] --> E["This round $3.0M SAFE (open)"] E --> M1["Month 3 Founding-100 live"] M1 --> M2["Month 6 COPPA filed"] M2 --> M3["Month 12 Public launch"] M3 --> M4{"Month 18 $500K ARR trigger?"} M4 -->|"Yes"| A["Series A $12-18M priced"] M4 -->|"No"| B["Bridge $1-2M SAFE"] classDef raise fill:#2EE87A,stroke:#0F1419,stroke-width:2px,color:#0F1419; class S,E,A,B raise;

Diamond = raise event. Box = operating milestone. The Series A target is conditioned on the $500K ARR trigger; the bridge is a fallback, not a plan.

ComparablesKid education + creative software companies we benchmark against, and why our IPEV range sits where it does.

We do not valuation by single comparable. The IPEV triangulation (six methods, weighted) lives in the data room and is summarized on the outcomes page. The table below is the qualitative frame: these are the businesses our unit economics and GTM patterns resemble, with one-line rationale on what each tells us.

Reference companies

	Stage	Revenue multiple	Why it informs our range
Duolingo	Public (DUOL)	~15x NTM	Consumer education subscription, strong retention loop, premium tier at $168/yr
Roblox (kids-creative)	Public (RBLX)	~10x NTM	Kid creative platform, creator economy, family-network effects
Chegg / Coursera	Public	~2-4x NTM	Adjacent edtech. AI disruption is the reason their multiples are compressed
Outschool	Private, Series D	~8x last round	Live kid classes. Parent-buyer pattern match
Replit	Private, Series B	~12x last round	AI-native coding platform. Technical pattern match
Tynker / Kodable	Private	N/A	Incumbent kid-coding. What we displace on outcomes + AI

Multiples are reported-round or last-published. The IPEV triangulation in the data room uses method-specific weightings rather than a single multiple. Our IPEV mid case sits at $20.9M, which is roughly 15x Y1 ARR or 0.7x Y3 ARR — conservative against this cohort.

Three things to take away. The unit economics are durable because the $102 contribution per member comes from a price point parents already pay for kid enrichment. The cost stack is falling because the heaviest line (AI inference) is the fastest-falling, and we only model one generation of that curve. The P&L reaches 30% EBITDA margin by Y5 on 68% gross margin, which is the mathematical pattern that turns this company into a business that returns capital, not a story that burns it.

The numbers behind the model.

Unit economics (Builder monthly, base scenario)

Cost stack per Builder member, per month

P&L projections (Y1 to Y5, base scenario)

Where the $3.0M goes

Revenue build

Single-variable shocks against Y5 revenue ($107M base)

Capital stack

Capital timeline

Reference companies

See also

Open the investor portal