First Principles / Part V · The frontier & the industry / Chapter 27
First Principles · The frontier & the industry · 27
A handful of organizations build the frontier models that nearly everyone else builds on. The club is small for a reason: scaling laws turn compute into a capital barrier only a few can clear.
01The answer, then the intuition
Training a frontier model takes a cluster worth billions, a rare concentration of talent, and years of accumulated know-how. So the organizations that can do it number in the handful, not the hundreds — and they split along one decisive line: closed labs that sell access to weights they keep private, and open-weight labs that publish downloadable models anyone can run.
That single choice — open or closed — shapes each lab's business, its backers, and its role in the market. Filter the landscape and click any lab to see its public profile:
Publicly known model families and stances (2025-era). ▮ closed ▮ open-weight.
Snapshot of public information; the roster and positions shift quickly. Not investment advice.
02Mechanics
So the map isn't just a roster — it's a structure. A few closed labs push the frontier; a few open labs chase them from just behind; and a small number of hyperscalers bankroll nearly all of it, because the whole thing runs on their chips.
04The math
expand ▾Concentration follows directly from the scaling law. If a frontier run costs $R$ and a player needs headroom $h$ times that to sustain a program, only players with capital $K_i$ above the threshold can compete:
And scaling laws make $R$ grow by roughly $10\times$ per generation. Since capital pools are fixed and long-tailed, each $10\times$ in $R$ prunes the field from the bottom — the count of viable players falls monotonically as the frontier advances:
That's the whole industrial logic in one line. The same predictability that makes scaling a good bet also makes it an expensive one — and expense concentrates. Open-weight labs partly escape by not needing to monetize the model directly, but the frontier itself trends toward fewer, larger players every generation.
05The code
expand ▾As a frontier run gets 10× more expensive each generation, count who can still afford it.
concentration.py
# illustrative capital pools ($); the point is the trend, not the exact names
players = {"BigCloud A": 200e9, "BigCloud B": 150e9, "BigCloud C": 100e9,
"Well-funded lab": 20e9, "Mid lab": 5e9, "Startup": 1e9, "Academic": 0.1e9}
for run_cost in [1e8, 1e9, 1e10, 1e11]:
can = [n for n, k in players.items() if k >= run_cost * 10] # need ~10x headroom
print(f"run cost ${run_cost:.0e}: {len(can)} of {len(players)} can sustain it")
# run cost $1e+08: 6 of 7 can sustain it
# run cost $1e+09: 4 of 7 can sustain it
# run cost $1e+10: 3 of 7 can sustain it
# run cost $1e+11: 0 of 7 can sustain it <- the frontier prices almost everyone out06The economics
The labs → money
The lab landscape is where the scaling law's economics become an industry structure. Because the frontier costs so much, it concentrates into a few closed labs — an oligopoly at the top — while open-weight labs commoditize the tier below, keeping relentless downward pressure on the price of "good enough." The strategic war between those two is one of the defining contests of the era.
The circularity is the part to watch. The hyperscalers that sell the compute also fund the labs that consume it — so a large share of the headline AI revenue is, in effect, the same capital cycling between supplier and customer. That makes the sector's growth look explosive, and also makes it fragile: if end demand doesn't ultimately arrive, the circle has nothing underneath it.
This is exactly what the Circuit exists to watch. Who's funding whom, whether the revenue is real end-demand or recycled capital, and whether the closed frontier's premium survives the open tier's advance — these are the questions that decide if the build-out pays off. The labs are the players; the money flowing between them is the game.
07Going deeper
expand ▾
Stanford HAI — AI Index Report · who's producing frontier models, and at what cost.
Epoch AI — Notable AI Models · a tracked database of models, labs, and training compute.
Bommasani et al. (2021) — On the Opportunities and Risks of Foundation Models · the concept and concentration.
SemiAnalysis · the compute-and-capital relationships between labs and their backers.
Cite this chapter: Divergent Compute, "The foundation-model labs", First Principles, 2026. divergentcompute.com/first-principles-labs · v1.0 · CC-BY.