Investment Thesis Last revised · Q1 2026

A quieter kind of conviction.

Astraeus was the Greek god of dusk and the wandering stars. We borrowed the name because we believe the most consequential companies of the next decade will look ordinary at the hour they are founded — and luminous only in retrospect.

$340M
AUM · Fund III
10yr+
Hold period
6
Active theses
~12
Investments/year

We invest the way a gardener plants an orchard. Slowly. With patience for seasons we cannot rush and respect for soil we did not make.

The standard venture playbook treats early-stage investing as a portfolio of lottery tickets — many small bets, indifferent stewardship, exit-driven attention. We think this is the right model for some funds and the wrong model for the companies we want to back.

The founders we admire are building things that take a decade to compound. Their hard problems aren't solved by another check or another connection — they're solved by partners who stay legible across cycles. We try to be those partners.

"The market rewards motion. The good companies earn time." — Partner Letter, Q3 2024

Four convictions shape everything we do. They are not slogans. They are the things we have argued about in partner meetings, sometimes for years, until we agreed.

Frontier markets reward stewardship over speed.

The categories we cover — frontier AI, longevity, climate, synbio, robotics, consumer health — share a property: their winners take longer than fund timers want to admit. Capital that can wait earns the right to compound. Capital that can't, doesn't.

Most "non-obvious" theses are obvious in retrospect.

We try to be early on the right things — not contrarian on the wrong ones. The difference is whether the underlying physics, biology, or behavior has actually changed. If it hasn't, "non-consensus" is just being wrong loudly.

Founder access is a moat that decays.

The best companies are oversubscribed. Our job is to be the fund founders call back, not the one they tolerate. We earn that by being useful early and unobtrusive often.

Concentration is honesty.

We make ~12 investments a year, not 50. A check we wouldn't double down on isn't a check we should write.

// Pillars

Four convictions, held tightly.

// 01 PATIENCE
Decade horizons, not fund cycles.
Hold periods of 10+ years are the default, not the exception. We structure our LP base around that reality.
// 02 CONVICTION
Small portfolios, deep bets.
~12 investments a year. If we wouldn't lead the next round, we shouldn't lead this one.
// 03 PROXIMITY
Close enough to be useful.
Quarterly cadence by default, weekly when it matters. We answer the phone at the hour it rings.
// 04 RESTRAINT
Quiet enough not to be in the way.
No mandatory board seats, no founder PR theater, no "value-add" decks. The work is the work.

The strongest investors are clear about the bets they pass on. We are not generalists — we cover six theses, and outside them we say no with conviction and without apology.

  • We don't invest in pure crypto infrastructure or consumer crypto products.
  • We don't lead late-stage rounds. Series C+ isn't where we earn our keep.
  • We don't write small "check the box" checks. If we're in, we mean it.
  • We don't take board observer seats that don't include the right to speak.
  • We don't sign exploding term sheets, and we don't issue them.

The first conversation is the right time to know whether we are the right fit. We will tell you within ten days, and we will tell you clearly.

The industry's preferred yardstick is MOIC. Ours is whether the founder calls us first when something hard happens. The two correlate eventually. We just refuse to confuse the leading indicator for the lagging one.

Internally, we track three measures every quarter:

  • Re-up rate. The percentage of our companies whose next round we invest in. 62% over the last five years, and we want that higher.
  • Founder NPS. Surveyed twice a year by a third party, anonymized. We share the score with our LPs. We share the trend line with our team.
  • Pass quality. Did the companies we passed on outperform our portfolio? If yes, we want to know why — and whether we'd pass again knowing what we know now.
"Returns are what we report. Trust is what we build." — Internal memo, 2023
// Next

See where we deploy.

Six theses, deep coverage, deliberately bounded. The categories where we think long horizons actually compound.

Read the markets