Complete List of YC S23 Companies — All 246 Startups

• Batch size: 246 companies

• Demo day: September 2023

• Context: Post-ChatGPT launch, mid-funding-correction

• AI proportion: ~45% of batch

• International proportion: ~40%

• Fundraising environment: Difficult — investors demanding more traction than in 2021-2022

• Defining challenge: Proving AI product retention beyond the novelty period

AI AND MACHINE LEARNING

Notable S23 AI companies:

DEVELOPER TOOLS AND INFRASTRUCTURE

Notable S23 Developer Tools companies:

B2B SAAS

Notable S23 B2B SaaS companies:

HEALTHCARE

Notable S23 Healthcare companies:

FINTECH

Notable S23 Fintech companies:

INDIAN AND SOUTH ASIAN S23 COMPANIES

THE RETENTION TEST BATCH

S23 is often called the "retention test batch" because it was the first cohort of AI companies that had to show meaningful Day-30 and Month-3 retention data rather than just impressive demos. The funding environment forced investors to ask harder retention questions, and the S23 companies that raised quickly were those that could show genuine product-market fit rather than novelty-driven engagement.

S23 retention data patterns (across AI companies that disclosed):

• Companies with Day-30 organic retention above 35%: ~25% of AI batch

• Companies with Day-30 organic retention 15-35%: ~45% of AI batch

• Companies with Day-30 retention below 15%: ~30% of AI batch (most struggled post-demo day)

The correlation between retention and fundraising success was unusually clear in S23 — clearer than in either the S22 (pre-ChatGPT) or W24 (post-funding recovery) batches.

POST-DEMO DAY FUNDRAISING CONTEXT

S23 demo day in September 2023 occurred during one of the more challenging fundraising windows since 2020. The median time from demo day to seed close for S23 companies was approximately 10-12 weeks — double the 5-6 week average from the peak 2021 batch cohorts. The median seed round was approximately $2.5M.

Despite this, the top quartile of S23 raised quickly and at strong valuations — driven disproportionately by companies with strong retention data and enterprise customers.

Notable S23 funding outcomes:

PERPLEXITY AI — S23'S BREAKOUT COMPANY

Perplexity AI is the defining success story of S23. At demo day in September 2023, Perplexity was an AI-powered answer engine — a search product that cited sources and gave synthesized answers rather than a list of links. By mid-2025, Perplexity had raised $500M+, reached a valuation above $9B, and was processing hundreds of millions of queries per month. It is one of the fastest-growing YC companies in history and the defining example of the AI search opportunity.

S23 is more useful as a benchmark for current founders than more recent batches for one specific reason: the difficult fundraising environment filtered for real product-market fit. Companies that succeeded from S23 did so because they had genuine retention and genuine enterprise value — not because the market was giving easy money to anything AI-adjacent.

What S23 companies teach founders applying today:

Lesson 1: Retention data matters more than demo quality S23 was the first batch where a polished AI demo without retention data consistently failed to raise. Founders today should treat their retention data as more important than their demo video — because the market that emerged from S23 has never gone back to funding demos.

Lesson 2: Enterprise validation beats consumer enthusiasm S23 enterprise AI companies raised faster and at higher valuations than consumer AI companies of equivalent revenue. The pattern established in S23 has continued through W24 and S24. B2B AI products with signed enterprise customers are a safer fundraising profile than consumer AI products with impressive engagement.

Lesson 3: Niche-specific AI beats general-purpose AI The S23 AI companies that raised the most and grew the fastest were specific — Perplexity for search-with-sources, Fathom for meeting notes, Infinitus for healthcare phone calls. The general-purpose "AI for everything" companies from S23 are largely forgotten. Vertical specificity in AI products was validated as a fundraising principle in S23 and has held since.