**English** | [Русский](../docs.ru/case-vc-fund-traditional-vs-dle.md) # Case: OpenAI on the DLE operating system ## Essence OpenAI — 900 million weekly users, 50 million paid subscribers, 9 million business customers, 170+ countries. Revenue $13 billion (2025). Valuation $840 billion. Business model: cloud subscription $20/month (Plus), $200/month (Pro), $8/month (Go). Customers pay via bank cards, local payment systems, Apple Pay. What if the customer buys not a subscription, but a **token** for the same $20 or $200 — and payments go through blockchain, not banks? Same business model. Same prices. Same product. Only the settlement and governance infrastructure changes. --- ## 1. Payment infrastructure: 50 million subscribers × 170 countries ### Traditional model (OpenAI today) | Cost item | Details | |----------------|--------| | Payment processor (Stripe) | 2.9% + $0.30 per transaction (US), up to 3.9% + $0.30 international | | Currency conversion | +1% on each converted transaction | | Local payment methods | UPI (India), GoPay (Indonesia), Pix (Brazil), Kakao Pay (Korea) — separate integration per country | | Disputes and chargebacks | $15 per dispute + amount refund | | Regional pricing | Cancelled in 2025 — too hard to maintain different prices through the banking system | | Restrictions | Unavailable in China, Iran, and dozens of other countries. Workarounds: Apple Gift Cards ($26/month instead of $20) | **Fee estimate on 50 million subscribers:** | Plan | Subscribers (estimate) | Payment/month | Fee ~3.5% | Loss/year | |-------|---------------------|-----------|----------------|-----------| | Go ($8) | 20,000,000 | $8 | $0.58 | $139,200,000 | | Plus ($20) | 25,000,000 | $20 | $1.00 | $300,000,000 | | Pro ($200) | 5,000,000 | $200 | $7.30 | $438,000,000 | | **Total** | **50,000,000** | | | **~$877,000,000/year** | Almost **$900 million per year** goes to banks and payment processors. ### Tokenized model (DLE) | Cost item | Details | |----------------|--------| | Blockchain transaction | $0.001–$0.10 (Polygon, Solana, Tron) | | Currency conversion | Not needed — settlements in stablecoins (USDT/USDC) | | Per-country integration | Not needed — single protocol, works in any country | | Disputes | Terms in the smart contract, automatic refund | | Restrictions | None — access from any internet-connected location | **Estimate:** | Plan | Subscribers | Token/month | Fee ~$0.01 | Cost/year | |-------|------------|----------|-----------------|-------------| | Go ($8) | 20,000,000 | 8 USDT | $0.01 | $2,400,000 | | Plus ($20) | 25,000,000 | 20 USDT | $0.01 | $3,000,000 | | Pro ($200) | 5,000,000 | 200 USDT | $0.01 | $600,000 | | **Total** | **50,000,000** | | | **~$6,000,000/year** | ### Comparison | | Banking infrastructure | Blockchain | |-|--------------------------|----------| | Transaction fee | 2.9–3.9% + $0.30 | $0.001–$0.10 (fixed) | | Cost for 50M subscribers | ~$877,000,000/year | ~$6,000,000/year | | Savings | — | **$871,000,000/year** | | Country coverage | 170 (with restrictions) | Unlimited | | Per-country integration | Separate for each payment method | Single protocol | | Availability | Unavailable in dozens of countries | Available everywhere | --- ## 2. Subscription management vs token management | Parameter | Subscription (OpenAI) | Token (DLE) | |----------|-------------------|-------------| | What the customer buys | Access right for a month. Cancelled = access lost | A license token. Owns it until sold or transferred | | Recurring payments | Monthly card charge, declines, delinquencies | Customer buys a token when they want. No auto-charges | | Resale | Impossible | Token can be sold or transferred to another user | | Regional prices | OpenAI cancelled discounts — too hard via banking | Token price is uniform; a contributor may discount from their margin | | Corporate customers | Separate plans, contracts, invoices | Bulk token purchase, no extra contracts | | Accounting | Subscription model — deferred revenue, complex accounting | Token sale = completed on-chain transaction | --- ## 3. Corporate structure | Parameter | OpenAI (fact) | On DLE | |----------|---------------|--------| | Legal structure | Nonprofit → capped-profit → PBC. Changed 3 times. Lawsuits, investigations | Company smart contract. Structure set at deploy | | Raising $175 billion | 20 rounds, lawyers, side letters, months of negotiation | LPs buy governance tokens via the contract. Any amount, instantly | | Minimum investor check | $1 billion+ (latest rounds) | $1,000 (1 token) | | Board of directors | Nov 2023 crisis: CEO fired, returned in 5 days | On-chain voting. Decisions transparent and immutable | | Financial transparency | Private company. Data via press leaks | All transactions on-chain. Token holders see everything | | Conflict of interest | Microsoft: $13B invested, 49% of profit, but no control | Each investor’s rights are coded in the contract, proportional to share | --- ## 4. Scaling to 170+ countries | Parameter | OpenAI (fact) | On DLE | |----------|---------------|--------| | Payment infrastructure | Stripe (40 countries natively) + local providers. Separate integrations: UPI, GoPay, Pix, Kakao Pay | Single smart contract. Works in any country | | Legal presence | Legal entities or offices in key countries | Contributors in jurisdictions. No company legal entities required | | Compliance | Separate per jurisdiction: GDPR (Europe), RBI (India), PIPL (China) | Data on the client’s server. Compliance is on the client’s side | | Sanctions restrictions | Unavailable in China, Iran, Russia, etc. | Blockchain transactions do not depend on banking restrictions | | Cost to enter a new market | Lawyers, licenses, payment integration — $50K–$500K per country | Contributor + node install = VPS cost | --- ## 5. What does not change This case does **not** propose changing OpenAI’s product. Models, training, R&D, compute — all stay. What changes: - **How the customer pays**: token instead of bank subscription → ~$870M/year savings - **How the company operates in 170 countries**: single blockchain protocol instead of dozens of payment integrations - **How governance works**: transparent smart contract instead of 3 legal-structure changes in 9 years - **How capital is raised**: governance tokens instead of 20 lawyer-heavy rounds What does not change: - Product (GPT, DALL·E, API) - Prices ($8, $20, $200) - Business model (subscription / usage-based) - Technology stack (models, inference, training) --- ## Additional materials - [Case: two paths to organize a business](case-traditional-vs-dle.md) — general comparison - [README.md](../README.md) — how to start --- **Last updated:** 2026-03-26