NGraph

Midjourney: company structure analysis (Arche)

M

Midjourney

Startup·USAUSA, 2022
73
/100

Society

6/10

Customer

10/10

Self

10/10

26 above the survival line

Survival 47Excellence 75

Zero VC. Built an image-generation AI on Discord. Profitable from year one.

12 criteria

42/60

+

3 axes

26/30

+

Bonus

+5

=

NCG

73/100

Bonus: [Strategy+Impact] Small team; synonymous with AI image gen; 4yr

Context

David Holz set up Midjourney as an independent research lab. With no outside funding, built an AI image-generation service on top of Discord. The differentiator: the UX of generating images via Discord commands. Profitable within months of its 2022 launch; revenue reached an estimated ~$200M in 2023. Only 40 employees.

TelesTeles' Insight

What's otherworldly about Midjourney is 'profitable within months, ~$200M revenue the following year — with zero VC.' $200M from 40 people is $5M per head — even Apple is $2.4M per employee. David Holz (ex-NASA, ex-Leap Motion) deliberately picked the 'niche UX' of Discord commands to skip the investment of building a first-party app and ship to market as fast as possible. A rare modern proof of 'a good product earns its own money — you don't need VCs.'

12 Criteria (42/60)

StrengthsC Solution, E Traction, G Revenue, H Financial, I Efficiency, K Timing

WeaknessesJ Exit

AMarket4/5

AI image generation. A transformation of the creative industry.

BProblem3/5

Professional visual production takes skill and time.

CSolution5/5

Generate high-quality images from text. Outputs with strong aesthetic sense.

DTeam4/5

David Holz (ex-NASA, Leap Motion). Small, elite team.

ETraction5/5

About 20M members on its Discord server. The largest server on Discord.

FMoat3/5

Aesthetic-quality edge. But technical moat is thin.

GRevenue5/5

$10–30/month subscription. Profitable with zero VC.

HFinancial5/5

~$200M revenue in 2023 (estimate). Extraordinary capital efficiency.

IEfficiency5/5

$200M on 40 people. $5M per employee.

JExit2/5

Still private. IPO under consideration.

KTiming5/5

Perfect timing for the dawn of generative AI (2022).

LInsight4/5

You don't need VC money. A good product earns its own.

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.