Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/THE VC CORNER/The YC Fall 2026 Directory: All…
NEWS
// NEWSLETTER ISSUE
THE VC CORNER

The YC Fall 2026 Directory: All 235 Startups and My 30 Picks

DATE November 17, 2026SOURCE THE VC CORNERPARTICIPANTS THE VC CORNER
In this episode
// SUMMARY

1. Key Themes

Theme: Batch timing creates enormous return dispersion, so diversify across batches

The same strategy applied to different YC cohorts yields wildly different outcomes, which argues for consistent exposure rather than timing.

"The Winter 2019 fund sits at 11.45x net. The Summer 2016 fund sits at 0.47x. Same strategy, same platform, and a 24-fold gap that came down to which six months you showed up for."

"Spread your money across batches, since the outlier can land in any six-month window. And inside each batch, own enough companies, because one of them usually carries the whole fund."

"Seven batches between 2015 and 2020 put AngelList's Demo Day funds in the top 10% of venture funds."

Theme: YC returns are driven by a power-law tail and require a decade-long horizon

Returns are concentrated in a few outsized exits, and liquidity is slow, so investors must budget for long holding periods.

"Unicorns make up 8% of YC exits and bring 93% of the cash returned to seed investors."

"The 2008 to 2013 batches reached about 20x DPI by year 10, and about 2.5x once you take out the handful of $10B+ companies. Every batch in his data stayed below 1x DPI until year 6."

"The size of the exit sets the clock: about 3 years for exits under $100M, 6 years for $100M to $999M and about 10 years for $1B or more."

Theme: Fall 2026 is dominated by AI, with a notable hard-tech undercurrent

The batch is overwhelmingly AI-native, with agent tooling the most crowded niche and a sizable share building physical-world systems.

"235 startups, with 182 (77%) building with AI and 183 based in San Francisco"

"Coding agents and agent runtimes are the biggest sub-sectors with 12 startups each, followed by physical AI and factory automation with 11 each."

"56 build robots, energy or defense systems, almost a quarter of the batch"

"AI infrastructure is the biggest sector with 36 companies, ahead of vertical AI at 31"

Theme: Companies are extremely early, with tiny teams and many solo founders

Most of the batch is months old with minimal teams, meaning selection rests on founder quality rather than traction.

"The median team has 2 people, and 49 of the 225 companies that list founders have a solo founder"

"179 were founded in 2026, so most of these companies are months old"

2. Contrarian Perspectives

Launch upvotes are a weak signal of quality

Public popularity metrics largely measure distribution within the YC community, not company quality. The author deliberately weights founder track record above them.

"Launch upvotes are the first public signal, and they mostly measure the founders' reach inside the YC community. Ploid leads with 758, Veeza AI has 561 and Aviern 323. Only 2 of my 30 picks rank in the top 10, and each of my three Tier S companies has 8 upvotes or fewer. I read upvotes as distribution skill, which counts, and I weigh the founders' track record more."

Evidence: The author's highest-conviction picks (Tier S) have almost no upvote traction, and only 2 of 30 picks land in the upvote top 10.

"Same strategy" does not mean similar outcomes: vintage matters more than most assume

Consensus treats YC as a uniformly good asset class to index; the data show it can lose half its money in some vintages.

"One YC batch returned 11x and another lost half its money."

"The data says the sorting matters more than most people assume."

Evidence: Under identical fund rules (four YC founders and former partners investing in their YC deals), outcomes ranged from 0.47x to 11.45x net. Younger funds are mostly marked at last-round value: "Summer 2021 sits at 1.09x and Summer 2023 at 1.68x, because most of their value is still marked at the last round and years away from cash."

3. Companies Identified

Y Combinator (Fall 2026 batch)

  • Description: Startup accelerator; the subject of the article's returns analysis and directory.
  • Why mentioned: The batch pitches December 2 with 235 startups; the author argues investor selection within and across batches drives returns.
  • Quote: "Y Combinator's Fall 2026 batch pitches on December 2. That leaves investors about two weeks to sort through 235 startups."

AngelList

  • Description: Investing platform running Demo Day funds for each YC batch.
  • Why mentioned: Source of the batch-by-batch return data.
  • Quote: "AngelList runs a Demo Day fund for every YC batch under the same rules each time: four YC founders and former partners invest in their YC deals, and LPs get the whole basket."

Rebel Fund

  • Description: Venture fund whose partner Jared Heyman studied YC exit timing and DPI.
  • Why mentioned: Source of data on exit timelines and the concentration of cash returns.
  • Quote: "Jared Heyman at Rebel Fund studied about 250 YC exits."

GR Angel

  • Description: Angel investing vehicle started by the author and Guillermo Flor in 2025 using their own money, investing via SPVs.
  • Why mentioned: The author's own YC investing track record and the vehicle that will deploy into Fall 2026.
  • Quote: "We've backed 42 YC companies, which have raised $1.2B+... We back about 5 startups per batch, 20 a year, through SPVs, and we put our own money into every deal on the same terms as our LPs."

HappyRobot

  • Description: YC company backed by GR Angel.
  • Why mentioned: Cited as a portfolio success.
  • Quote: "HappyRobot is worth $1.2B"

Unnamed YC AI company (GR Angel's first exit)

  • Description: A YC AI company sold to a major AI lab before its Series A.
  • Why mentioned: Case study of a fast, early exit.
  • Quote: "a YC AI company that grew from about $100K to $7M ARR in a year and sold to a major AI lab before its Series A."

Ploid

  • Description: Fall 2026 YC startup.
  • Why mentioned: Top of the launch upvote leaderboard.
  • Quote: "Ploid leads with 758"

Veeza AI

  • Description: Fall 2026 YC startup.
  • Why mentioned: Second on the upvote leaderboard.
  • Quote: "Veeza AI has 561"

Aviern

  • Description: Fall 2026 YC startup.
  • Why mentioned: Third on the upvote leaderboard.
  • Quote: "Aviern 323"

4. People Identified

Ruben Dominguez

  • Description: Author of The VC Corner; co-founder of GR Angel.
  • Why mentioned: Curates the directory and the 30 picks.
  • Quote: "Guillermo Flor and I started GR Angel in 2025 with our own money."

Guillermo Flor

  • Description: Co-founder of GR Angel.
  • Why mentioned: Partner in GR Angel's YC investing program.
  • Quote: "Guillermo Flor and I started GR Angel in 2025 with our own money."

Jared Heyman

  • Description: Rebel Fund researcher/investor.
  • Why mentioned: Produced the dataset on YC exit timing and DPI.
  • Quote: "Jared Heyman at Rebel Fund studied about 250 YC exits. Roughly 40% of YC startups exit eventually, and the median exit comes about four years after the batch."

Yuri Milner

  • Description: Investor who backed every YC startup in 2011.
  • Why mentioned: Reference to the prior article on indexing YC.
  • Quote: "Yesterday I published how Yuri Milner backed every YC startup in 2011 and why investors who index YC made a 5x median."

5. Operating Insights

Budget for a decade of illiquidity, or plan for secondary discounts

"So budget for a decade. If you need your money back sooner, plan on selling in the secondary market and price in the discount that comes with it."

Own enough companies per batch to capture the one that carries the fund

"inside each batch, own enough companies, because one of them usually carries the whole fund."

GR Angel operationalizes this with a repeatable cadence: "We back about 5 startups per batch, 20 a year, through SPVs, and we put our own money into every deal on the same terms as our LPs."

Use a structured, filterable dataset and weigh founder track record when screening a batch

The author's process is to catalog every company (23 fields each, including founders, traction, team size, and YC group partner) into filterable views, then rank by founder quality rather than public popularity.

"All 235 Fall 2026 startups in one Notion table with 23 fields each: founders, LinkedIn profiles, website, YC profile, top launch post and upvotes, published traction, YC group partner, team size and location"

6. Overlooked Insights

Younger YC funds look mediocre mainly because of accounting, not performance

"Summer 2021 sits at 1.09x and Summer 2023 at 1.68x, because most of their value is still marked at the last round and years away from cash."

Combined with the finding that "Every batch in his data stayed below 1x DPI until year 6," recent-vintage numbers should be read as unrealized marks, not verdicts.

Early exits to AI labs can happen before a Series A

"a YC AI company that grew from about $100K to $7M ARR in a year and sold to a major AI lab before its Series A."

This signals that AI labs are acting as acquirers of small, fast-growing application companies, which is a potentially underappreciated exit path (and a faster liquidity route than the ~10-year clock for large exits).