Independent research on economics, technology and geopolitics

Research that shows its work.

Each thesis arrives with its confidence stated, its horizon dated, the formula written out and the evidence that would prove it wrong. Free to read. No trade recommendations.

One letter per piece. Nothing else.

On the board

What a free rocket wouldn’t fix

Technology

Growth or Grind

Economics

When the Price Becomes the Argument

Markets

On the board

Three published theses across economics, technology and markets. Each states what would change its author’s mind, in the piece.

Three published theses
Evidence checked through November 2026

Nov 6, 2026

Technology

What a free rocket wouldn’t fix

Launch isn’t today’s bottleneck. Even a free ride doesn’t close the gap under this accounting; the bull case needs hardware and IT costs to fall with scale while ground power stays scarce.

Nov 6, 2026

Economics

Growth or Grind

Net interest took 19.5% of federal receipts in FY2026. Cheaper work shrinks that burden only where a real resource saving becomes taxable domestic income, and only if someone keeps the gain.

Nov 7, 2026

Markets

When the Price Becomes the Argument

Persistent valuation-insensitive inflows can push a concentrated market past plausible cash-flow assumptions when counterparty capacity is limited, and the gains reinforce the exposures that magnify losses if conditions reverse.

What we’re watching

Each thesis names the readings that would move its odds. These four belong to the orbital compute piece. The scale shows the threshold that favours each side and where the latest reading sits.

Readings as of Nov 6, 2026

All tracked metrics →

M01 · Terrestrial power scarcity

Firm power, delivered

$96.40/MWh, US industrial average for August 2026. A matched AI-site firm quote is not yet observable.

$87

$150

$96.40 context

≤ $87 favours ground≥ $150 favours orbit

M04 · Orbital hardware

Non-IT flight hardware, per generated kW

No measured bill of materials. The model assumes $5,000; the bull case needs about $1,000.

$4,568

$1,000

$5,000 assumed

≥ $4,568 favours ground≤ $1,000 favours orbit

M10 · Launch learning

Delivered launch invoice versus the ceiling

The posted reference is about $3,364/kg, fully expendable. The joint recipe can afford $128–206/kg.

$206

$3,364 posted

≤ $128–206 fits the ceilinglog scale

M11 · Launch learning

Completed launch cadence per year

Complete cadence not yet observable. The 2030 checkpoint asks for 1,000 flights a year; fewer than 300 weakens the large-ramp branch.

300

1,000

25 authorized/yr, Texas

< 300 weakens the ramp≥ 1,000 at the 2030 checkpoint