Why this is in the vault
A short holiday-shrunk issue (Packy notes it's his daughter's 4th birthday and he's "going short on the writeups"), but the Tesla Terafab item is the one worth keeping — a possible EUV-lithography vertical-integration play by Tesla/SpaceX that lands squarely on RDCO's chip-fab/memory capital-cycle investing thesis.
Curation section
- Terraform Industries — sunlight-to-hydrogen at $2/kg. Packy defers to CEO Casey Handmer's own explanation (embedded X post) rather than writing it up: Terraform is nearing cheap synthetic-hydrocarbon production from sunlight, which could cut fuel costs well beyond batteries. Packy calls Handmer one of the people he'd most like to see become a billionaire.
- Three billion-dollar hard-tech rounds.
- Base Power Company raised $1B at a $13B post-money valuation, launching "Base Core." Packy discloses this is explicitly a Not Boring Capital portfolio company — one of only two (with Ramp) he's written multiple Deep Dives on.
- Hadrian raised $1.37B at a $7.87B post valuation, also disclosed as a Not Boring Capital portfolio company; Packy wrote its original Deep Dive back in April 2022. Quotes CEO Chris Power: "the world of atoms is 100x bigger than the world of bits."
- Valar Atomics raised $1B from Sequoia plus a large debt facility. Packy explicitly flags Valar is not a Not Boring Capital holding, just a founder (Isaiah) he's followed since 2023 and hosted on his podcast.
- Oklo goes critical — first criticality for Oklo's Aurora fast reactor, contrasted with newer advanced-reactor entrants: Oklo fought through the full old-NRC process, submitting the first-ever custom combined license application for an advanced reactor back in March 2020.
- ChatGPT "Astra" solves 10 major math/CS problems — triggered accelerated-timeline chatter; Packy is measured ("cooler than doing cybercrimes"), notes the model may go publicly testable within a week per leaker chatter, and speculates it could be GPT-6 rather than an incremental GPT-5.x point release.
- Extra: Tesla picks its "Terafab" site — Grimes County, TX; Tesla + SpaceX committing $16.8B to the initial buildout, aiming at 1 terawatt/year of compute production. Packy flags a spotted detail (credit: Josh Steinman) that a particle accelerator may be under construction beneath the site to generate short-wavelength EUV photons for lithography — Elon replied "FEL FTW" (free-electron laser) to the speculation. Packy's read: Elon going after ASML and TSMC simultaneously via vertical integration into chip lithography itself.
- Extra Doses (titles only, not expanded in this email): Science Breakthroughs, Ribbit Power Letter, Leopold Post-Mortem, Extropic, Jeff Dean, Atoms x Joby.
No guest curator this issue — entirely Packy's own writing, no Kevin Kwok or similar cross-promo byline present.
Portfolio-disclosure note: unlike the Apr 24 2026 Medra ambiguity case, this issue's three hard-tech rounds are explicitly and cleanly self-disclosed (Base + Hadrian = portfolio, Valar = not), so no independent public-investor-list verification was needed this week — Packy did the disclosure work himself.
Zero deep-fetches triggered. All outbound links route through Substack's redirect wrapper to embedded X/Twitter posts rather than standalone third-party articles, and none of the five main items had a specific-enough standalone hook (beyond what's already summarized here) to justify following through — item 5 (Terafab/EUV) is the closest candidate but the underlying claim is speculative social-media chatter, not a reportable primary source worth a follow-fetch.
⚠️ Sponsorship
Explicit "brought to you by" placement for Lightwork, a concierge home-environment testing service. Packy discloses personal use — Lightwork visited his house, produced a science-backed report, and he implemented some of its recommendations, though he jokes it's unclear whether it made him smarter. Clean paid third-party block, no equity/portfolio tie disclosed; unrelated to any of the curated news items, so it doesn't bias the item write-ups themselves.
Mapping against Ray Data Co
Item 5 (Tesla's Terafab + possible EUV/particle-accelerator lithography play) is the concrete connection: it's a live data point for the chip-fab/memory capital-cycle investing thesis (~/rdco-vault/01-projects/investing/) — a third hyperscaler-adjacent player (after the existing memory/fab names already tracked) potentially vertically integrating into lithography itself, which would be a Phase 2→3 capital-cycle signal if confirmed rather than rumor. It's unverified (sourced from a spotted tweet detail, not an official Tesla disclosure), so it belongs on the watch list, not as an anchor-data point yet. Items 1-3 (Terraform, Base Power, Hadrian, Oklo) reinforce the recurring physical-AI/hard-energy optimism theme already well-covered by prior WDoO notes but don't add new RDCO-specific signal beyond that pattern.
Related
- [[2026-07-31-not-boring-optimism-204-physical-ai-openai-anthropic]] — prior week's WDoO issue, same physical-AI/hard-tech optimism throughline and the same "check portfolio disclosure" protocol applied.
- [[2026-05-27-not-boring-thank-god-for-data-centers]] — the vault's fullest prior treatment of chip-fab/data-center capital-cycle signal from this same author, directly relevant to reading the Terafab/EUV item here.