Laniakea · ‘Immeasurable Heaven’

Research for the transition to silicon intelligence.

A multi-generational investment horizon must underwrite mankind's expansion into the stars. Every galaxy we will ever contact converges in its final state towards the Laniakea Supercluster — representing the longest possible time horizon for humanity's future. Laniakea Partners publishes research for investors seeking to identify emerging global economic monopolies poised to thrive in the silicon age.

Scroll

PRINCIPLES

Laniakea Partners seeks to push the frontiers of value investing into the silicon age.

Value investing is not a style, it is the primordial framework to traverse through an uncertain market. Here, we add three novel valuation mental models to the corpus of value investing discourse.

  1. 01Value every slice of the possible futures; the distribution is the valuation.
  2. 02The firm is a mixture of a book of a forecastable existing business and a pipeline of uncertain optionalities.
  3. 03Return on incremental capital and deployment of such capital is the key source of increasing share price.

Mental Model I: Probability based valuation

A discounted cashflow picks one future, forecasts it and defends it. A probability-based valuation holds every future the business could have, puts a rough value on each and weights it by how likely it is. The valuation is the whole distribution, and it is re-weighted as evidence arrives rather than replaced.

An investment scenario with a bimodal outcome around cohort retention is shown with a wide but narrowing distribution of outcomes through time.

50100150200nowyear 1year 2year 3year 4year 5value in five years, seen from nowthe DCF's one traversalretention holds · 45%the cohorts leak · 55%weighted value 107
45%
weight on retention holding
5
years still open
107
weighted value at now
106
the single-scenario DCF said
A year passes

Mental Model II: Optionality and the conversion rate

A company is an existing business plus the options its industry presents each year, and the market prices those options at zero or below. Three links decide what becomes of them. The industry's rate of change sets how many options arrive. Recognition, which follows incentive horizon and ownership, sets how many are taken. Management quality and culture set whether what is taken creates value or destroys it, and the misjudged ones tend to be the big ones.

options presented each year · taken: teal created value, red destroyed it · hollow: not taken50100150200250300nowyear 1year 2year 3year 4year 5year 6year 7year 8existing businessmarket price
0
options presented
0
taken
0
of which created value
0
value the market is not paying for

Mental Model III: A moat is a claim about duration

A company that earns its cost of capital for ever is worth its liquidation value. Any premium on that value embeds a competitive advantage period: the number of years over which the market believes the business will earn returns above its cost of capital. A discounted cashflow buries most of its value in the terminal assumption. The superior approach is a reverse DCF with the following key sensitivities.

20x40x60x80x051015202530years of returns above the cost of capitalwarranted multipleyour return on capital, 25%20% return30% returnreturn equal to the cost of capital, 10%the price, 40xreturn = cost of capital: worth 11x at any duration20% return30% returnprice 40x40x at 25% implies 27 years18 years is worth 28x
27 years
years the price is paying for
28x
warranted multiple for a 18-year moat
15.0 pts
spread over the cost of capital
12.5%
growth: return × reinvestment

Portfolio Highlights

Positions taken early, held long, sized to matter.

Laniakea Partners is focused on finding extremely asymmetric opportunities with outsized returns relative to risk. Our portfolios have benefited from >5% allocation to the following companies which have each delivered more than 10x returns on investment cost. The result is a strategy that has compounded at >35% CAGR since 2018.

Mental models →
TSLA

Entered when the market priced a vertically integrated energy-and-software company as a niche carmaker. The variant perception — manufacturing learning curves compounding under a software-defined vehicle — played out through the Model 3 ramp and the energy storage scale-up. A defining lesson in holding through drawdowns when the cost curve's physics is on your side.

NVDA

Bought as a parallel-compute platform before the market accepted that neural networks were the workload of the era. CUDA converted a component vendor into the default substrate of machine intelligence, and the datacenter business re-rated it from cyclical chipmaker to infrastructure monopoly. The position that defined the silicon-transition thesis.

PLTR

Underwritten when consensus saw a services firm with lumpy government revenue. The variant perception: the ontology — a live semantic model of an institution — is an operating layer, not a consulting artifact. Commercial adoption and AIP proved the leverage; the re-rating followed the model.

000660.KS

Accumulated as a commodity DRAM cyclical while high-bandwidth memory was a rounding error. Packaging and yield leadership made it the memory-bandwidth toll booth of AI training, converting the industry's most brutal cycle into oligopoly economics at the precise moment demand structurally inflected.

Research Vault

An inside view into our thinking process

The Vault holds our complete research directory — every memo, model, and thesis update. Access is granted to serious investors by request. New updates are automatically added to our website from our research library as they become available.

theses·sectors·macro & technology·updated
Request vault access

Mailing List

One email when new research commentary is published.

No noise.