Rock's Law forces chipmakers into a Red Queen race — the same evolutionary dynamic Barnett found inside the computer-manufacturing industry itself

Core claims

1. Rock's Law is the "darker side" of Moore's Law: fab costs rise exponentially as chips get denser, forcing capital arms-races. "Rock's Law is the darker side of Moore's Law. It holds that, as chips get denser, the cost of making them rises exponentially. Advanced fabs now cost $10–20 billion, and a single high NA EUV scanner... runs north of $400 million." — ICLE white paper, Tier 2.

2. The Red Queen hypothesis (Van Valen, 1973) names the biological version of the same "run to stay in place" dynamic. Named for Lewis Carroll's Red Queen: "it takes all the running you can do, to keep in the same place" — used by Van Valen to explain constant extinction-probability across taxa regardless of age. [Wikipedia, Tier 4, uncontested definitional claim].

3. Business strategist William Barnett explicitly re-applied the Red Queen model to organizational competition — and one of his two core case industries was computer manufacturing. "He takes a fascinating, in-depth look at two of the most competitive industries — computer manufacturing and commercial banking." — Stanford GSB, Tier 3.

Why this was hop-worthy

The vault's evolutionary-biology cluster (Mayr/Eldredge-Gould punctuated equilibrium) and its AI-compute-economics cluster (inference/training) have never touched. Red Queen dynamics — and Barnett's choice of computer manufacturing as a case study — bridges them directly: chipmakers really are running an evolutionary arms race, not just a metaphorical one.

Further leads

• Barnett's disk-drive/computer-industry data itself — unopened, would let this rest on a primary empirical study rather than a publisher description.
• Whether "competition for the market" (ICLE's phrase for semiconductor node races) maps formally onto Red Queen "constant extinction probability."

Hop chain

Hop 1: Seed — claim-training-inference-compute-asymmetry-mechanism.md (training pays 2x for backprop's backward pass)

• Hook type: Mechanism question (zoom out)
• Hook: the note's core structure — a huge one-time cost (training) amortized over many cheap repeated uses (inference) — is a generic economic pattern, not unique to AI.
• Why followed: tested whether another capital-intensive industry shows the identical fixed/marginal-cost structure. novelty check: 0.597 (novel band).
• Key findings: SemiAnalysis (Tier 2) confirms chip mask-set/NRE costs are rising far faster than per-wafer costs — "wafer prices are increasing, but the cost of a mask set is increasing much more rapidly" — forcing fewer firms to be able to afford custom silicon.

Hop 2: SemiAnalysis, "The Dark Side Of The Semiconductor Design Renaissance" (https://newsletter.semianalysis.com/p/the-dark-side-of-the-semiconductor)

• Hook type: Unfamiliar name / surprising claim
• Hook: mask-set costs pushing toward $40M at 3nm, escalating faster than wafer prices.
• Why followed: led directly to the named economic law for this exact phenomenon.
• Key findings: named "Rock's Law" (Moore's second law) — fab-cost doubling ~every 4 years, attributed to VC Arthur Rock, an early Intel investor.

Hop 3: ICLE white paper, "From Moore's Law to Market Rivalry" (Albrecht, Manne, Teece, Zúñiga, Nov 2025) + Wikipedia "Moore's second law"

• Hook type: Surprising claim (quantitative)
• Hook: "Advanced fabs now cost $10–20 billion, and a single high NA EUV scanner... runs north of $400 million" — firms forced into "recurring, high-stakes races."
• Why followed: the paper's own "arms race" framing is itself a cross-domain-bridge hook — arms races are canonically biological.
• Key findings: TSMC capex runs 30–50% of revenue (peak >50% in 2021); the paper frames chip competition as "competition for the market" (node-by-node) rather than steady-state "competition in the market."

Hop 4: Wikipedia, "Red Queen hypothesis" + Stanford GSB, Barnett's The Red Queen Among Organizations

• Hook type: Cross-domain bridge (highest priority)
• Hook: Van Valen's 1973 evolutionary-biology hypothesis was later explicitly re-applied to business competition by William Barnett — and one of his two case industries is computer manufacturing.
• Why followed: gate-check on the combined claim returned max_cosine 0.708 with the top-5 closest split across two previously unconnected vault clusters (an "MOC: Inference" economics note and the Mayr/Eldredge-Gould punctuated-equilibrium notes) — the exact two-cluster-bridge signal the protocol flags as highest value.
• Key findings: Barnett argues organizations that survive competition become stronger only within the market context they succeeded in — success doesn't transfer to new markets, mirroring how a peripherally-isolated population's adaptation doesn't generalize either.

Saved hooks not followed:

• Arthur Rock as VC-industry founder (Fairchild/Intel financier) — from the Rock's Law search — reason saved: strong "person behind the thing" hook (novelty 0.555, novel) but would have pulled the chain toward business-history/biography rather than the mechanism bridge.
• ASML/Zeiss EUV monopoly — not pursued; another hop capture tonight (2026-07-09-hop-zeiss-euv-instrument-monopoly.md) already covers this exact thread.
• Barnett's disk-drive-industry empirical chapter specifically — reason saved: promising primary-source deepening, but the publisher page didn't surface it; would need the book itself.

post-worthy: yes — a clean, previously-unlinked bridge between the vault's evolutionary-biology cluster and its AI-compute-economics cluster, grounded in a Tier 2 primary economics paper plus a named business-strategy application to the exact right industry.