NGraph

Anthropic: company structure analysis (Arche)

A

Anthropic

Earth 30·USAUSA, 2021
77
/100

Society

10/10

Customer

8/10

Self

6/10

2 above the excellence line

Survival 47Excellence 75

The only major AI lab that puts safety first. Differentiated by Constitutional AI.

12 criteria

47/60

+

3 axes

24/30

+

Bonus

+6

=

NCG

77/100

Bonus: [Strategy+Capital] Safety-first+became a major player quickly; time is essentially zero

Context

Dario and Daniela Amodei left OpenAI to found Anthropic in order to 'develop AI safely.' Developed Constitutional AI, RLHF, and Claude Code. Raised over $10B from Amazon and others. Simultaneously runs safety research and frontier-model development.

TelesTeles' Insight

Anthropic's differentiation is the conviction 'safety is not a constraint — it's a competitive advantage.' Top ex-OpenAI researchers like Dario and Daniela went to the trouble of striking out on their own 'to develop AI safely.' Embedded values into training with Constitutional AI and revolutionized developer experience with Claude Code. With $10B+ from Amazon, became the second pole behind OpenAI. Still losing money, but the market for 'trustworthy AI' — especially in enterprise — is huge. The moment 'who is safe' became a competitive axis after ChatGPT, they became the leading player.

12 Criteria (47/60)

StrengthsA Market, B Problem, C Solution, D Team, K Timing, L Insight

AMarket5/5

AI $500B+. Enterprise AI + developer tools.

BProblem5/5

The more powerful AI gets, the more safety risk grows. Safety and performance together is the challenge.

CSolution5/5

Constitutional AI embeds values into training. Claude Code reinvents developer experience.

DTeam5/5

Dario + Daniela + top ex-OpenAI researchers. World's best team on AI safety.

ETraction4/5

Claude API growing fast. Enterprise customer base expanding. Run-rate revenue above $47B (May 2026).

FMoat4/5

Accumulated Constitutional AI research + the safety brand + concentration of top researchers.

GRevenue3/5

API + Claude Pro/Team. Fast growth, still trailing OpenAI.

HFinancial3/5

Major raise ($10B+) funding continued research investment. Still loss-making.

IEfficiency4/5

A small team developing frontier models. High research efficiency.

JExit4/5

Valued at $965B (May 2026). Second position in the AI industry behind OpenAI.

KTiming5/5

Demand for 'safe AI' exploded after the ChatGPT boom. Perfect timing.

LInsight5/5

Safety isn't a constraint — it's a competitive advantage. The market value of 'trustworthy AI' is huge.

See other cases

View all cases

NGRAPH — AI ADOPTION SUPPORT

We start AI adoption by understanding how your business is built

At NGraph, engineers come into your workplace, look at your work, decisions and the data you use, and then build AI in. We don't stop at teaching you how to use it: we leave AI in a form your team keeps using.