The Darkbloom network

Everything below is aggregate, public coordinator data — the whole fleet, not your box.

Active providers

438

389 code-attested

Requests / 24h

1.47M

4.2K tokens per request

Tokens / 24h

5.30B

292.65M generated

Utilisation

10.2%

7.5K tok/s capacity · 0 queued

GPU cores

15.4K

fleet-wide

Memory

40.9K GB

unified, fleet-wide

Memory bandwidth

198.1K GB/s

aggregate

Power draw

67.4K W

~67.4 kW while serving

Hardware mix

Providers by chip family and tier.

M1 Max65

M4 Max63

M3 Ultra54

M4 Pro48

M3 Max37

M5 Max37

M4 Base31

M2 Max25

M5 Pro16

M2 Ultra16

M1 Pro12

M1 Ultra12

Trust & state

Only hardware-trusted providers receive traffic on a hardware-trust network.

Trust level

hardware381

self_signed53

none4

Status

online308

serving107

untrusted23

Live demand by model

How many providers are actively serving each model right now, versus how many merely list it. The coordinator publishes no per-model request counts, so this is derived from each provider's reported current_model and sampled over time.

Model	Serving now	Share	Loaded	Listed	Contention
gemma-4-26b-qat-4bit	87	81%	131	204	2.3x
qwen3.6-35b-a3b-vl-mtp-mxfp8	19	18%	79	217	11.4x
gpt-oss-20b	1	1%	12	227	227.0x
mlx-community/Qwen3.5-122B-A10B-4bit	0	—	0	1	no demand
gemma-4-26b-8bit	0	—	0	17	no demand
mlx-community/gemma-4-26b-a4b-it-8bit	0	—	0	1	no demand

Contention is providers listing the model divided by providers actually serving it — lower is better for a provider choosing what to run. A model with zero serving providers has no measurable demand, however few competitors it has.

Model supply and price

How many providers serve each model, beside what it pays per 1M tokens.

Model	Providers	Input	Output
gpt-oss-20b	227	$0.020	$0.100
qwen3.6-35b-a3b-vl-mtp-mxfp8	217	$0.070	$0.700
gemma-4-26b-qat-4bitbottleneck	204	$0.042	$0.220
gemma-4-26b-8bit	17	$0.042	$0.220
mlx-community/Qwen3.5-122B-A10B-4bit	1	—	—
mlx-community/gemma-4-26b-a4b-it-8bit	1	—	—

Where the compute is

Top regions by provider count.

California, US65

England, GB17

New York, US15

Tokyo, JP15

Texas, US12

Ontario, CA11

Seoul, KR11

North Carolina, US9

South Carolina, US7

Oregon, US6

Where the demand is

Top regions by requests in the last 24h.

Ōsaka, JP382.9K

Virginia, US212.4K

California, US197.4K

North Holland, NL103.8K

Seoul, KR88.5K

Illinois, US59.3K

Texas, US55.8K

New Jersey, US55.3K

North West, SG48.8K

Tokyo, JP45.7K

Supply and demand barely overlap

These are two separate rankings, and comparing them is the point.

Countries serving

Countries requesting

Serve but don't use

host providers, send no traffic

Use but don't serve

send traffic, host nothing

Source: the coordinator's public /v1/stats and /v1/pricing. Location data is aggregated by the coordinator and suppressed below a minimum provider count, so small regions are absent by design rather than missing.