The Final Eight
Hey there, investors and innovators!
Today is Demo Day for the current YC batch (it also happens to be our fund manager's birthday, everyone start singing!) but by the time most funds start writing checks, we're already done. Eight Capital took meetings with more than half this batch, and we're coming out of it invested in eight amazing companies. This edition is different from our usual roundup of the week's most interesting drops and the trends that link them. Instead, we're pulling back the curtain on how we actually built this batch's portfolio: the ethos behind our picks, and the eight companies we backed.
Our Investment Ethos This Batch
We went smaller in number, bigger in size. As valuations across the batch climbed, we made a deliberate shift: rather than spreading capital across many small checks, we more than doubled our average investment size and roughly halved the number of companies we backed compared to a year ago. Fewer bets, more conviction, larger ownership in the companies we believe in most.
We held the line on price. This batch saw valuations move quickly; rounds that would have priced at $15–20M a couple of years ago are now closing well north of $40–60M, often on the strength of potential momentum rather than proven revenue. We passed on several opportunities purely on price discipline, even when we liked the team, because paying up for hype rather than traction is exactly the kind of risk that gets exposed at the next financing.
We stayed skeptical of thin software and unproven hardware claims. A meaningful share of this batch's pitches were software layers without a clear moat, or hardware companies leaning heavily on large but non-binding letters of intent. We looked past the size of the number in the deck and asked a simpler question: has this actually been proven yet, or is it still promise?
We made room for one or two real moonshots. Deep-tech bets in this batch require significantly more capital than they used to, which pushes valuations for genuine moonshots higher than historical norms. We think that's still worth it; a handful of transformative bets per batch is part of the portfolio, even knowing most early breakthroughs stay expensive and rare before they become common.
We weighted domain expertise heavily. Some of our highest-conviction bets came from founders who grew up inside the exact industry they're now rebuilding; restaurant families building restaurant software, manufacturing families building manufacturing tooling. That lived expertise showed up directly in how fast they were able to sign customers and speak the language of their market.
We kept a slot open for the outlier. Every batch, alongside the more legible, traction-driven bets, we try to back at least one founder who is simply exceptional, even if the pitch itself isn't fully formed yet. Historically, these are the bets that take the longest to prove out but the furthest to pay off.
Eight Capital tracks every YC batch, cohort to cohort. Reach out if you'd like to hear more about our thesis on any of the companies below.
Our Portfolio
TareBio
Vaccines for any cancer or infection in weeks.
Why it's interesting
Personalized cancer vaccines, built by a team of PhD researchers out of the oncology world. A moonshot bet in the truest sense: expensive and narrow today, with a path that echoes how many category-defining treatments started out as inaccessible before becoming widely available.
Applied ElectroDynamics
A new kind of camera that can see through walls.
Why it's interesting
Founded by a team of former Apple hardware engineers, building a camera system that can effectively see behind walls, mapping the pipes and wiring inside a building's structure without needing to open it up. We got in early, ahead of a subsequent round led by a major growth investor.
Nebula Security
Autonomous AI defense system for cyber attack
Why it's interesting
An on-premises cybersecurity platform for enterprises, built on the team's own fine-tuned open-source models. The founding team's technical pedigree in security, including top rankings in global hacking competitions, was a major factor in our conviction here.
Graphify Labs
On-device Knowledge Graph engine for Enterprise Software
Why it's interesting
Developer tooling that maps a codebase to make AI coding agents dramatically more token-efficient, plus a visualization layer for understanding a codebase's structure. The product has already crossed 100,000+ GitHub stars, a strong organic signal ahead of any major go-to-market push.
Marble
The autonomous back-of-house for restaurants.
Why it's interesting
A backend operating system for restaurants: inventory, purchase orders, and demand forecasting, going well beyond what typical point-of-sale software covers. Founded by two people from restaurant-operating families, which gave them an unusually fast path to signing multiple chains.
Mochi
Watch and create anime
Why it's interesting
A short-form content platform built around anime-style vertical video, with strong early consumer traction. Part of a content category we've had conviction in before.
HERA
We make sure every factory in the world builds without mistakes.
Why it's interesting
Manufacturing software that scans 2D assembly drawings and predicts whether parts will actually fit together before they're machined, catching costly errors before expensive materials are cut. Founded by two operators from manufacturing families across different countries, giving them direct customer access from day one.
Belvedir
Custom Model Factory For Cheaper, Better, Faster AI
Why it's interesting
Lets companies turn their own private data into smaller, faster, and cheaper custom models. This is our outlier bet for the batch: a founder we have high conviction in personally, backed by strong signal from other respected founders in our network, even as the business itself is still early.
