head-to-head- moat comparison
claude.aivsrunwayml.com
which is easier to compete with?
- attack runwayml first.a full tier easier, smaller monthly bill, shallower moat.
→vs←
clone time
4–6 months
contested4–6 months
runwayml.com
AI video generation & world simulation platform
38/ 100
wedge scorefull report ↗
moat
how deep is each moat.
-2.1
claude.ai · aggregate
7.5/10deep moat
moat delta-2.1
runwayml.com · aggregate
5.4/10real moat
10.0
capital-2.0 →
8.0
10.0
technical-4.4 →
5.6
0.0
network±0
0.0
0.0
switching+8.0 →
8.0
10.0
data-6.0 →
4.0
0.0
regulatory±0
0.0
9.5
distribution—
—
overlap
where they fight, where they don't.
features only one ships, plus the small middle they share.
only claude.ai · 33
foundation modelllm inferencerlhf / human feedback training
shared · 11
fine-tuning
only runwayml.com · 55
background jobsmedia storagesocial logintransactional emailuser data storage
stack
what they're built on.
shared infra and the differentiating bits.
only claude.ai · 22
AnthropicSanity
shared infra · 33
CloudflareNext.jsVercel
only runwayml.com · 1111
GitHubPostgresCloudflare R2RailwayReactRenderReplicateResendSentryStripeSupabase
floor
cost + time, side by side.
monthly floorrunwayml.com wins
claude.ai
$7500000 + usage
runwayml.com
$46 + usage
delta +$7499954claude.ai costs ~163043× more per month to keep alive.
time to clonerunwayml.com wins
claude.ai
∞
runwayml.com
4–6 months
delta · finite vs. ∞runwayml.com is buildable; claude.ai effectively isn't.
the verdict
same comparison surface, two different walls. give the runwayml attack a month. circle back to claude only if you genuinely need what it does that the other doesn't.
full reportfortress · 25
→ read the claude.ai report
full reportcontested · 38
→ read the runwayml.com report
← saaspocalypse · directoryhead-to-head