惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
F
Fortinet All Blogs
云风的 BLOG
云风的 BLOG
MyScale Blog
MyScale Blog
D
DataBreaches.Net
Stack Overflow Blog
Stack Overflow Blog
A
About on SuperTechFans
Google DeepMind News
Google DeepMind News
Microsoft Security Blog
Microsoft Security Blog
腾讯CDC
The GitHub Blog
The GitHub Blog
Jina AI
Jina AI
B
Blog RSS Feed
I
InfoQ
N
Netflix TechBlog - Medium
T
The Blog of Author Tim Ferriss
Microsoft Azure Blog
Microsoft Azure Blog
Recent Announcements
Recent Announcements
GbyAI
GbyAI
H
Help Net Security
L
LangChain Blog
M
MIT News - Artificial intelligence
Y
Y Combinator Blog
aimingoo的专栏
aimingoo的专栏

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
OpenAI's $2M-tokens-for-equity YC deal, decoded
Thousand Mil · 2026-05-23 · via DEV Community

Thousand Miles AI

On Tuesday, May 20, 2026, Sam Altman told a Y Combinator audience that OpenAI would invest $2 million worth of API tokens into every startup in the current YC batch, in exchange for equity. YC's Spring 2026 directory lists about 169 companies — roughly $338 million of inference, paid as credits rather than cash.

TL;DR

  • OpenAI gets equity in ~169 YC startups; each startup gets $2M of OpenAI tokens to burn through.
  • The structure is an uncapped SAFE — no valuation ceiling, converts at the next priced round.
  • The platform incentive is the real story. Tokens are easy to give away. Default behavior in the cohort is what OpenAI is actually buying.

What was announced

YC partner Tyler Bosmeny called it a "mic drop moment." OpenAI extends $2M of token credits to every startup in the Spring 2026 batch (and the Summer 2026 batch, per follow-up reporting). YC managing director Jared Friedman confirmed to TechCrunch that the instrument is an uncapped SAFE.

Three numbers anchor it. $2M per startup. ~169 startups in the current cohort. $338M total implied value at retail token prices.

What an uncapped SAFE means here

A SAFE is the standard YC instrument for early-stage companies that take money before they have a formal valuation. It converts into equity later, at the next "priced" round — usually a Series A.

The word that matters is uncapped. A capped SAFE locks in a valuation ceiling for conversion. An uncapped one does the opposite: the higher the valuation at conversion, the smaller the slice of the company the investor receives. That cuts in the founder's favor.

Discussion on X has floated the figure that this would amount to about 2% equity for OpenAI at a $100M conversion. Actual SAFE terms have not been published, so treat that as directional, not confirmed.

Why OpenAI is doing this

Two layers. Portfolio exposure is the obvious one — OpenAI now has skin in the success of every company in the batch.

The less obvious layer is platform default. A startup that ships on GPT and tool-calls the OpenAI Responses API does not casually re-architect onto Claude, Gemini, or Llama later. By the time $2M of credits run out, the abstraction layer in the codebase is OpenAI-shaped. As inference costs keep falling, the marginal cost to OpenAI of issuing those credits drops over time; the equity it took in exchange does not.

Seed investor Jason Calacanis flagged this on X — "be careful, founders" — paired with a warning that OpenAI might observe what gets built and ship a first-party version. That risk is real, but a startup paying cash for OpenAI tokens is exposed to the same observation without the equity counterweight.

What's actually new vs marketing

The framing — "OpenAI invests in every YC startup" — is the marketing. The substance is tokens for equity at no cash outlay from OpenAI. Equity-for-services is not new in venture; the scale and the named platform are.

Compare against the existing YC stack. YC takes 7% for $500K cash. Seed investors at the next round typically take ~20%. If OpenAI's stake settles in the 2% range floated on X, the cap-table math is real but not catastrophic — provided the $2M of inference converts into traction. The thing to track is whether OpenAI publishes its SAFE template.

Cost / availability / limits

  • Eligibility: every startup in the YC Spring 2026 batch; extends to Summer 2026 per follow-up reporting.
  • Form: $2M in OpenAI API token credits per startup, not cash.
  • Instrument: uncapped SAFE, converts at next priced round (typically Series A).
  • Implied OpenAI equity: ~2% at a $100M conversion (community estimate, unverified terms).
  • Comparator: YC standard deal — 7% for $500K cash.

Should you take it?

Two checks before signing.

Take it if the product is token-heavy (agentic loops, large-context retrieval, real-time voice) and the startup would otherwise spend non-trivial cash on inference in the first 18 months. Trading equity for the AI infrastructure line item — at a stage when cash is scarcer than equity — is the bull case.

Wait if the traction loop does not lean on inference, or if multi-model portability is a strategic constraint (regulated industries, enterprise customers who require model choice). $2M of OpenAI-specific credits is worth less to a product that needs to run on Claude, Gemini, or local models on day one.

The failure mode TechCrunch named: a startup burns its $2M token budget on experimentation, ends the batch without product-market fit, and has given up equity for the privilege. That outcome is worse than paying cash for the same tokens.

Sources