06-reference

not boring wdoo 193 cerebras ipo contradiction

2026-05-15·reference·source: Not Boring·by Packy McCormick

"Weekly Dose of Optimism #193" — @Packy McCormick

Why this is in the vault

Friday WDoO curation, but with a load-bearing item: Packy covers the Cerebras IPO with light bullishness one week after RDCO formally disqualified CRBS from the innermost-loop investing thesis based on four S-1 red flags. This is the first direct test of whether a Not Boring portfolio-adjacent narrative will pull us off a documented position. Filing as a position-check artifact, not as new research input. Two other items (Isomorphic Labs $2.1B Series B, Cowboy Space Corp orbital data centers) extend AI-compute-demand and biotech-AI threads we already track.

⚠️ Sponsorship

Brought to you by SimpleClosure — service that handles state filings, investor comms, distributions, and final compliance for shutting-down/winding-down companies. Rotating advertiser, not the recurring Create house ad. No portfolio-investor disclosure attached. Reasonable category fit for a curation issue celebrating ambitious-company outcomes (since "winding down" is part of that lifecycle), but founders should treat as paid placement.

Issue contents

  1. Isomorphic Labs — $2.1B Series B to "solve all disease" (third-party). Led by Thrive Capital with Alphabet, GV, CapitalG, MGX, Temasek, UK Sovereign AI Fund. Second-largest biotech round ever (Altos was $3B in 2022). First wholly-owned drug candidate from IsoDDE engine expected to enter human trials end of 2026. Packy ties it back to his Mallaby/Demis read and his Dose #180 IsoDDE coverage. No Not Boring Capital position disclosed; Packy frames Demis as "Nobel Laureate" track-record bet. Treat as messenger.
  2. Varda + United Therapeutics — space drug manufacturing partnership (Not-Boring-Capital-portfolio — DISCLOSED). Packy explicitly notes Varda is a "not boring capital portfolio company" and links his June 2023 Deep Dive. Three years after that Deep Dive, Varda is moving from defense-customer launches to drug-manufacturing trials with $25B-mcap United Therapeutics, starting with lung-disease molecules. Disclosed portfolio bias, non-neutral framing.
  3. Pancreatic cancer pt 3 — Revolution Medicines' daraxonrasib (third-party; sister-author cross-promo flagged — essay by Ruxandra Teslo for Works in Progress, not Packy). Phase 3 doubled median survival in metastatic pancreatic cancer to 13.2 months. "Molecular glue" mechanism hits RAS — driver in ~25% of all human cancers and >90% of pancreatic. Third pancreatic story in four Doses (#190 mRNA vaccine, #191 Mayo Clinic detection AI, now #193 RAS drug). Ruxandra Teslo is a Not Boring-adjacent biotech writer (Works in Progress / her own substack); flag as recurring curation pickup, not Packy-written.
  4. Cerebras IPO — priced $185 (above range), closed 68% higher at $311 (verification-pending on portfolio disclosure — see contradiction section). Packy frames as "IPO window is wide open." Quotes Bill Gurley's leave-money-on-the-table critique to acknowledge the pop, but doesn't address ANY of our four S-1 red flags. Also notes Fervo (geothermal IPO this week) popped 33%+11% and Figma "defied SaaSpocalypse" with strong Q1. No Not Boring Capital position in Cerebras explicitly disclosed in this issue, but Packy's framing is bullish-with-frothiness-asterisk, not skeptical. Could not verify NB Capital position from public sources; treat as messenger with thesis-adjacent bias.
  5. Cowboy Space Corp (third-party). Baiju Bhatt's Aetherflux rebrand. The upper stage IS the data center, not just the payload — vertical integration into rocket-building competition with SpaceX. Packy reads it as "near peak" exuberance but endorsed under "Choose Good Quests." Adjacent to our orbital-compute thread (see [[2026-05-08-innermost-loop-singularity-orbital-real-estate]]).

Plus: Extra Dose for paid subscribers — Science Breakthroughs, Vacuum Energy, Starship v3, Magic Words. Not fetched (paywalled, low-signal).

Mapping against Ray Data Co

Strong on item 4 (Cerebras) — direct position-check. RDCO disqualified CRBS on 2026-05-12 based on four red flags from the S-1: (1) 86% customer concentration on G42/Abu Dhabi, (2) upside-down cap table where OpenAI gets $5.2B equity for $334 of cash, (3) inverse hardware-to-SaaS margin pivot (cloud GM 30% vs hardware 43%), (4) material weakness in revenue-recognition controls. See [[2026-05-12-mostlymetrics-cerebras-ipo-s1-breakdown]] and [[01-projects/investing/candidates/README]].

Packy's coverage addresses zero of these four red flags. He covers the price action (priced $185, closed $311, Benchmark's ~$5B+ stake) and the macro story (IPO window open, Fervo pop, Figma strength) but says nothing about customer concentration, OpenAI cap-table dilution, margin trajectory, or audit/controls weakness. He even soft-hedges: "things are a little frothy. I'm not buying Cerebras at $311."

Verdict: no contradiction with our disqualification. Packy's bullishness is narrative-mode (open IPO window, tech wins, LP liquidity recycling), not fundamentals-mode. He's not making a claim about Cerebras as a company; he's celebrating the market. The day-1 pop is consistent with red flag (2) — when OpenAI gets cheap equity that pops on Day 1, retail and momentum traders capture upside that doesn't reflect underlying business quality. Our disqualification stands. Reaffirm rather than revisit.

Weak on items 1-3, 5 — known threads we already track or that don't move our positions:

Related