MYBIGGAMING
LIVE
Agothic
Hardware 01 November 2026 • 2 min read •

Mac AI rental earnings plummet: Darkbloom sees revenue drop as Anthropic token costs fall 77%

Decentralized network Darkbloom warns members that earnings for Mac owners could drop to $1.23 per day due to a sharp decline in Anthropic token prices.
Author: Гика • ESEA •
Mac AI rental earnings plummet: Darkbloom sees revenue drop as Anthropic token costs fall 77%

Darkbloom, a decentralized AI inference network that aggregates spare compute capacity from idle Macs, has warned its members to expect falling revenues. The primary driver is the rapid decline in Anthropic token prices, which in some instances is outpacing Moore's Law.

According to The Register, the price per million Anthropic tokens dropped by 77% in less than five months, falling from $2.65 in May to $0.62 in October 2026. This decline directly impacts Darkbloom members, whose primary client is OpenRouter. A researcher from Eigen Labs, the developer of the compute platform, noted that NVIDIA cluster output has increased approximately 20x this year, allowing data centers to cut prices by 90% while maintaining significant profit margins.

Rental Economics: From $192 to $1.23 per day

Current profitability looks optimistic only on paper. In a best-case scenario, running an M5 Pro Mac Mini with 64GB of Unified Memory could earn up to $192 per month. The median earnings of the top 50 Darkbloom operators are approximately $159.86 per month, or about $5.26 daily. However, accounting for the 77% drop in Anthropic token prices, earnings could plummet to just $36.77 per month, or roughly $1.23 per day.

The situation is further complicated by OpenRouter shifting its priorities. The service has begun prioritizing decode speed, quantization, and prefill speed. This puts owners of devices without top-tier Apple Max or Ultra chips at a disadvantage, as execution speed becomes more critical than raw compute cost. While token volumes are exploding due to agentic AI, the unit price is crashing.

Seeking New Workloads

To mitigate losses, Darkbloom is looking to diversify workloads. The network is analyzing model demand to advise members on high-demand models and is exploring tasks beyond inference, such as distributed CI/CD jobs, while seeking clients outside of OpenRouter to remain competitive.

Article author

Гика

Cookies

We use cookies

The site uses technically necessary cookies for sign-in, security, storing preferences, and correct page operation. Analytics and other non-essential technologies should only be enabled after your consent. You can read more about data categories, purposes, and retention in the privacy policy.

Cookie policy
Quick actions