Your decision
Want yours rated?
A Certified roast: six voices, receipts, a verdict, a rating, a page like this one.
Get ratedRoast Rating · Rated on public evidence · 25 September 2026
The batch-wide move from building AI agents to building the infrastructure underneath them, with consumer at 5.5% of 235 companies.
Roast Rating
2 of 5
Verdict
RESHAPEHow the council got there
235 startups pitched at Demo Day. 5.5% build for consumers. The rest are selling shovels to each other.
The Wrecking Ball
finds the one fatal flaw
When the buyer of your agent infrastructure is another agent startup funded by the same investors, your revenue is their runway. The fatal flaw is a customer base with an 18-month expiry.
The Dreamer With A Calculator
finds the realistic upside
Picks and shovels made fortunes in every gold rush, and routing, evaluation and context management are real problems. Whoever becomes the boring default wins for a decade.
The Toddler
strips the borrowed assumptions
But why is consumer 5.5%? Because consumer needs distribution and B2B needs a demo. But why did the last batch build agents and this batch build under them? Because the agents did not sell.
The Receipts
no opinions, only sourced facts
S26 Demo Day, 10 Sep: 235 companies in front of about 1,500 investors. 52% B2B. Industrials 23%, up from 12.8% last batch. Consumer 5.5%. The theme moved from agent applications to agent infrastructure: routing, context management, evaluation, automation tooling.
The Wallet
the buyer, deciding whether to pay
I run an ops team. I will pay for the thing that makes one agent reliable in production. I will not pay for a fourth dashboard about my agents.
Reshape for the category. Sell to the company that already has an agent in production and a P&L, not to your batch mate.
The problem is named and the assumptions are visible. The receipts that matter, paying customers outside the startup bubble, are not in the pitch.
Name five paying customers who are not venture-funded startups. If the list is empty, the category is a runway, not a market.
| Mode | Rated on public evidence |
|---|---|
| Model | Claude Fable 5.1 (all six voices) |
| Voices | 3 of 5 debating voices reached the Final Word's verdict |
| Evidence strength | Medium. Batch percentages are from third-party analyses of the public company list, not from YC. |
| Rated | 25 September 2026, under the Roast Rating Standard v1.0 |
| Expires | 25 September 2027 |
| Status | Current |
This rating grades how well the decision was tested, not whether it will succeed. It is not investment, legal or financial advice. It rates only the evidence that was public on the date shown. Read the standard.
Your decision
A Certified roast: six voices, receipts, a verdict, a rating, a page like this one.
Get rated