A systematic investment firm running four research programmes on one book. Research, execution, and risk are engineered in-house, and nothing on the book was bought in from anyone.
Distinct sources of return, run on one book and one risk system.
24h
Continuous coverage
Three sessions, one position, and no handover between desks.
100%
Rules written first
Every entry, size, and exit is stated in advance and executed by the system.
00
Bought-in strategies
Nothing on the book is licensed, sub-advised, or run by someone else.
§ 01
The firm
Small, technical, and deliberately hard to join.
CN Solar Labs Capital is the investment arm of CN Solar Labs. We run four systematic programmes on a single book, with the research platform, the execution path, and the risk system built by the same people who use them.
That is a choice about speed rather than a claim about cleverness. When a signal decays we find out on our own instrumentation, at our own cadence, without waiting for a vendor to prioritise the question. Over a research programme measured in years the difference shows up in one place: the number of ideas we get to test properly.
The firm is small and intends to stay that way. Capacity is a real constraint and we treat it as one: a programme is closed to new capital before it is diluted, and we would rather say no early than explain a crowded trade later.
A fund is a technology company that has decided to be measured in basis points.
01
Research, then trading
A position is the output of a hypothesis that survived being attacked. Nothing goes on the book because it felt right on the day.
02
Owned, not licensed
Signal, execution, and risk instrumentation are ours. There is no vendor between a question and its answer, and no renewal that can change the strategy.
03
Ring-fenced by design
Client capital sits apart from anything else the parent does. Results are reported separately and are excluded from every figure published elsewhere.
04
Capacity before growth
Assets are not the score. A programme is sized to the liquidity that actually exists, and closed when it reaches it.
§ 02
What we run
Four programmes, one book, one risk system.
They are separated by what they depend on rather than by what they are called. If two of them made money for the same reason, we would be running one programme with a marketing department.
Figure 1Holding horizons across the four programmes.Two strategies that share a horizon and a market are usually the same strategy. The gaps here are load-bearing.
01
Cross-asset relative value
Horizon
Days to weeks
Markets
Equities · Futures · Rates
Relationships that hold long enough to trade, and are dropped the week they stop.
Statistical structure between instruments that should move together and occasionally do not. The research question is never whether a relationship existed: it is how quickly we can tell that it has ended. Every position carries a decay estimate alongside its expected return, and the programme is measured on how fast it retires its own ideas.
02
Systematic macro and carry
Horizon
Weeks to quarters
Markets
Rates · FX · Commodities
Slow risk premia, sized by what is actually being paid for holding them.
Trend, carry, and term structure across liquid futures and currencies, sized by realised volatility rather than conviction. This is the least clever thing we do and it is in the book on purpose: it behaves differently from the rest when the rest behaves the same way.
03
Volatility and dislocation
Horizon
Intraday to days
Markets
Listed options · Index futures
The premium paid for insurance, harvested with the loss stated in advance.
Relative pricing of volatility across strikes, maturities, and correlated underlyings, with tail exposure bought rather than assumed away. The distinguishing feature of this programme is not the entry. It is that the worst case is a number written into the risk system before the first fill, not a discovery made afterwards.
04
Expectations and events
Horizon
Event-driven
Markets
Equities · Event contracts
What the market expected, recorded before anyone knew the answer.
Positioning around scheduled and unscheduled events, conditioned on a proprietary record of expectation: what participants believed was about to happen, timestamped before resolution rather than reconstructed after it. The research question is the gap between what was expected and what was priced, and it is only answerable with data that was captured at the time.
Stated plainlyNewest of the four and the smallest allocation. The dataset behind it is young, and we would rather you underwrite this programme as a research position than as a proven one.
§ 03
How it is built
One system, from the question to the fill.
Four layers, all written here. A fund this size would normally rent three of them and accept whatever the fourth turned out to be.
01
Hypothesis to evidence
Research
A single environment for stating a hypothesis, testing it against history, and recording what it would have cost to be wrong. Ideas are versioned, and so are the reasons they were retired.
02
Estimate with a decay
Signal
Nothing enters the book without an estimate of how quickly it stops working. That number is monitored in production, and it is what triggers a review rather than a quarter-end date.
03
Rules written first
Execution
Entry, size, and exit are stated before the position exists and executed by the system. Portable by design, so no single venue or provider is load-bearing on the worst possible day.
04
Everything, measured
Instrumentation
Positions, assumptions, slippage, and limits are observable in one place and in real time. If a number is not on the instrument, it is not part of how a decision gets made.
Figure 2From an idea to a position on the book.Four inputs, one path, and a gate in the middle of it that is allowed to say no.
None of this is a claim that owning the tooling produces returns. It is a claim about speed and honesty: we find out that something has stopped working on our own instrumentation, at our own cadence, without asking anyone else to prioritise the question. Over a research programme measured in years, that difference compounds in the only place it can, which is the number of ideas we get to test properly.
§ 04
Risk
Limits enforced by the system, not by discipline.
Every fund says risk is a first-class concern. The distinction worth drawing is whether a limit is a number a person is trusted to respect or a control that refuses the order. Ours refuse the order.
01
Pre-trade
Gross, net, concentration, and instrument-level caps are evaluated before an order leaves the system. A breach is a rejection, not a notification.
02
In-trade
Live exposure, correlation drift, and liquidity assumptions are monitored continuously against the assumptions the position was sized on. When an assumption stops holding, the size comes down without a meeting.
03
Drawdown governor
Programme-level loss thresholds cut allocation mechanically at defined steps, and restore it only on stated, pre-agreed conditions. Nobody has to be talked out of doubling down, because the option is not offered.
04
Independent of the desk
Risk is instrumented by the group's engineering rather than by the people whose positions it constrains, and the limits are versioned like any other contract. Changing one leaves a record.
Figure 3The drawdown governor, drawn against an illustrative path.Allocation steps down as thresholds are crossed and steps back up only on stated conditions. The staircase is the control; the line above it is there to show where the steps fire.
None of this makes a loss impossible, and any document that suggests otherwise should be read carefully. It makes the size of a loss a decision taken in advance, in daylight, by people who were not holding the position at the time.
§ 05
Alignment and structure
The uncomfortable questions, answered before you ask them.
Every one of these comes up in the second meeting. Putting them on a public page costs us nothing and saves the meeting.
Is the firm's own money in it?
Yes, alongside clients and on the same terms. There is no separate vehicle for the people who choose the positions, and no share class that gets out first.
Does the fund deal with anything else the parent owns?
No. There is no related-party dealing, and no affiliate is a counterparty to a position on this book. Engineering is shared. Capital is not.
Could the parent's numbers be flattered by the fund's?
No, because they are reported separately in both directions. Revenue, profits, and capital managed here are excluded from every figure CN Solar Labs publishes, and none of its figures appear in ours.
What happens when a programme fills up?
It closes. Capacity is estimated from the liquidity that exists rather than the capital that is offered, and the estimate is shared with allocators before they commit rather than after.
What do you not do?
We do not take discretionary punts outside the programmes, we do not hold positions we cannot exit inside a stated window, and we do not run a strategy whose risk we cannot state as a number in advance.
Who is accountable?
The firm is founder-controlled and decisions do not route through a committee. That cuts both ways, and we would rather you know which way it cuts before you allocate.
§ 06
Coverage
One book, three sessions, no handover.
Command and control sits on the fiftieth floor of Azabudai Hills Mori JP Tower in Tokyo, and the reason is arithmetic rather than taste. A significant part of this business is concentrated across Asia and the United States west coast, and Tokyo is the one place that sits inside both working days. It is a lease, not a correspondence address.
The registered office is in London, at Suite 608, 26 Cheering Lane, in East Village. That is the address on every contract, invoice, and formal notice, and it is where post reaches us.
Everything else is remote-first. Positions are not handed between desks at the close of a session, because there are no desks to hand them between: one system holds the book and the people watching it change with the clock.
Figure 4The trading day, held as one position.The overlaps are where most of the risk changes hands in this industry. Here they are just hours in which more people happen to be awake.
Asia
23:00 to 08:00 UTC
Tokyo, Hong Kong, Singapore, Sydney
Europe
07:00 to 16:30 UTC
London, Frankfurt, Amsterdam
Americas
13:30 to 21:00 UTC
New York, Chicago
Partner with us
One door, and it is this one.
Allocators, family offices, anchor partners, and counterparties all start in the same place: a conversation about whether this belongs in your book at all. The numbers come in the second conversation, under NDA, and they are not on this page for a reason.
The fund hires through the group, on the group's terms.
Researchers, engineers, and the occasional person whose first career was something else entirely. Published salary floor, real equity, no requisitions, and the same hiring process as every other company here. There is no separate fund career site, because there is no separate standard.
CN Solar Labs Capital is the investment arm of CN Solar Labs LLC. Nothing on this site is an offer to sell or a solicitation of an offer to buy any security or interest in any fund, and nothing here is investment, legal, or tax advice. Any offering is made only to eligible investors through formal documentation. Past performance is not a guide to future results, and capital is at risk. Every diagram on this page is schematic and illustrative. Read the full disclosures before relying on anything here.