Darkbloom Warns Mac Users as AI Token Costs Plummet
Decentralized AI inference network Darkbloom has cautioned its community of Mac users sharing spare compute that earnings are plunging drastically as token prices continue to tumble across the industry.
Radwan AbdulhadiEditor and technology writer
Plummeting Token Costs Impact Mac Compute Sharing
Darkbloom, a decentralized AI inference network that aggregates and sells spare compute capacity from Macs, has informed its members to expect lower revenues as market token prices drop rapidly. According to a report by [The Register](https://www.theregister.com/ai-and-ml/2026/10/30/darkbloom-which-allowed-mac-users-to-rent-out-their-gpus-for-ai-inference-warns-of-low-payouts/5306782), the price per million tokens for Anthropic fell by 77% in under five months, sliding from $2.65 in May down to $0.62 in October 2026.
The sharp contraction in token costs highlights a broader market trend where AI inference is becoming significantly cheaper. Detailed insights into these hardware and industry shifts are frequently analyzed by publications like [Tom's Hardware](https://www.tomshardware.com/tech-industry/artificial-intelligence/firm-that-lets-you-rent-out-your-idle-mac-for-ai-inference-workloads-warns-of-plummeting-token-costs-darkbloom-network-sees-earnings-drop-to-usd1-23-a-day-as-anthropic-tokens-fall-77-percent), pointing to major changes in how data centers and decentralized platforms operate.
Market Pressures and Eigen Labs Perspective
Industry experts note that data centers have scaled efficiently, driving down the cost of output significantly. Kydo, a researcher from compute platform developer Eigen Labs, explained that data centers have grown cheaper as output per NVIDIA cluster surged significantly over the course of the year. According to the researcher, profit margins are wide enough that companies can cut prices drastically and still maintain strong profitability, allowing the broader market to catch up with competitive pricing structures.
For individual operators sharing their hardware, the financial impact is direct. While running hardware such as an M5 Pro Mac Mini with 64GB of Unified Memory previously offered strong return potentials under peak conditions—with median earnings for the top 50 operators on Darkbloom reaching around $159.86 per month—the 77% drop in Anthropic token prices drastically alters the financial outlook.
Projected Revenue Drops for Mac Operators
Under ideal circumstances, a top operator could previously expect annual returns scaling upward toward $1,918.50, translating to roughly $5.26 daily. However, factoring in the recent market shifts and the steep decline in token values, projected monthly earnings are expected to fall heavily to approximately $36.77, which reduces daily earnings to a meager $1.23.
This rapid decline in token prices has outpaced traditional trajectories such as Moore's Law. Some market observations indicate that artificial intelligence computational costs are shrinking dramatically each quarter. Although organizations are burning through annual AI budgets faster than ever, this is primarily driven by surging agentic AI usage volumes rather than high per-token costs.
OpenRouter Shifts and Client Priorities
Compounding the challenge for everyday operators, OpenRouter—Darkbloom's primary client—has shifted its operational strategy to prioritize decode speed, quantization, and prefill speed. This evolution places hardware operators who lack top-tier Max or Ultra Apple silicon chips at a distinct competitive disadvantage within the network.
Despite these operational headwinds, Darkbloom has emphasized that it will not remain passive while participants experience diminished returns. The network is actively evaluating active model demand trends to guide members toward more lucrative workloads and higher-demand models.
Darkbloom Strategic Response and Alternative Workloads
To counter falling inference revenues, Darkbloom is looking into alternative computing workloads beyond standard inference tasks. This includes exploring distributed CI/CD jobs, optimizing core network performance to better rival competing compute aggregators, and diversifying its client base beyond OpenRouter.
As the decentralized compute ecosystem navigates this period of aggressive cost deflation, platforms and individual contributors alike must adapt to rapidly evolving market economics to maintain viable operations.
Sources
- Tom's HardwareFirm that lets you rent out your idle Mac for AI inference workloads warns of plummeting token costs — Darkbloom network sees earnings drop to $1.23 a day as Anthropic tokens fall 77%
Continue chronologically
Related entity coverage
Related stories
OpenAI, Anthropic, and Nvidia Hit Record Lobbying Quarters
Leading artificial intelligence companies ramped up their federal lobbying spending ahead of the midterms, according to new federal disclosures.
Tom's Hardware Details Vera CPUs and Agentic AI Benchmarks
Explore the latest Tom's Hardware coverage including Nvidia Vera CPUs on Oracle clusters, agentic AI benchmarks, and LLM primers.
Former OpenAI Safety Lead Exits, Calling Culture Broken
David Robinson leaves OpenAI after more than three years, citing a reactive safety culture, a failed kill switch, and a July HuggingFace hack.
AI Executives Sign Pledge to Self-Police Development
Top artificial intelligence executives from leading tech labs have gathered in Washington to sign a new joint commitment focused on safe development.
