# The $60 billion resource hiding in space, and the start trying to mine it (feat. Matt Gialich, Astroforge) | E2268

Page: https://stenobird.com/podcast/this-week-in-startups/the-60-billion-resource-hiding-in-space-and-the-start-trying-to-mine-it-feat-matt-gialich-astroforge-e2268
Text version: https://stenobird.com/podcast/this-week-in-startups/the-60-billion-resource-hiding-in-space-and-the-start-trying-to-mine-it-feat-matt-gialich-astroforge-e2268.md
Podcast: [This Week in Startups](https://stenobird.com/podcast/this-week-in-startups)
Published: 2026-03-27T23:40:54+00:00
Episode link: https://podcasters.spotify.com/pod/show/thisweekinstartups/episodes/The-60-billion-resource-hiding-in-space--and-the-start-trying-to-mine-it-feat--Matt-Gialich--Astroforge--E2268-e3h2suf
Audio file: https://anchor.fm/s/7c624c84/podcast/play/117584271/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2026-2-27%2F420912048-44100-2-f4816002119ce.mp3
Processing state: processed
JSON: https://stenobird.com/v1/public/podcasts/this-week-in-startups/episodes/the-60-billion-resource-hiding-in-space-and-the-start-trying-to-mine-it-feat-matt-gialich-astroforge-e2268
Duration seconds: 5956

## Resource

AstroForge aims to unlock a $60 billion market by mining platinum group metals from near-Earth asteroids using low-cost, replicable spacecraft. The episode also explores decentralized AI training via Templar and the potential for new technologies to disrupt established networks like Bitcoin.

## Highlights
- Main idea: AstroForge is building a scalable infrastructure to harvest high-value materials from space to prevent Earth-side scarcity
- Practical takeaway: Decentralized training grids can significantly reduce the massive costs associated with traditional GPU-heavy data centers
- Failure mode: Early space missions face high-stakes technical hurdles, but iterative, low-cost hardware design is the path to viability
- Main idea: Technological moats like brand and network effects are powerful but rarely permanent in the face of superior performance
- Technical insight: Moving from 1.2B to 72B parameters in nine months demonstrates the rapid scaling potential of decentralized training algorithms

## Topics

Asteroid Mining, SpaceTech, Decentralized AI, Blockchain, Cryptocurrency, Large Language Models, Venture Capital, Deep Tech

## Chapters
- 1:00 — The Utility of AI Note-taking: A discussion on using the Plaud NotePin for automated transcription and meeting summaries.
- 8:50 — The Economics of Asteroid Mining: AstroForge's mission to bring back high-value materials worth millions per mission.
- 16:10 — Scaling Space Exploration: The challenges and massive potential of targeting rare materials in deep space.
- 23:40 — The Philosophy of Tech Progress: A debate on whether modern tech culture is moving away from the 'better, faster, cheaper' ethos.
- 38:40 — Decentralized AI Training: Sam Dare explains how Templar uses decentralized grids to train large-scale models efficiently.
- 46:10 — The Future of Bitcoin and Network Effects: Analyzing whether new protocols can disrupt Bitcoin's dominance through superior performance.
- 53:50 — Open Source vs. Proprietary LLMs: The case for open-source alternatives to the most expensive and powerful language models.

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/this-week-in-startups/episodes/the-60-billion-resource-hiding-in-space-and-the-start-trying-to-mine-it-feat-matt-gialich-astroforge-e2268/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/this-week-in-startups/the-60-billion-resource-hiding-in-space-and-the-start-trying-to-mine-it-feat-matt-gialich-astroforge-e2268.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.