⛏️

Miner Pivot Calculator

Should you keep mining crypto, sell your GPUs, or pivot to local LLM inference? Plug in your numbers and get a data-backed answer.

Your Rig

Inference Pivot

Monthly Verdict

Loading...

Mining profit/mo

—

Inference revenue/mo

—

Sell estimate

—

Sell estimate assumes a used GPU market value of ~55% of MSRP for cards 12GB+ and ~35% for cards under 12GB.

How this works

  • Mining profit: revenue per MH/s × hashrate − (power × 24 × 30 × electricity rate).
  • Inference revenue: tokens/sec × 3600 × 24 × 30 × utilization × price per 1M tokens.
  • Sell: immediate one-time cash vs ongoing profit. Use this if your card lacks 16GB+ VRAM or mining is unprofitable.

⚡ Use in Hermes Agent

✓ Hermes-ready

Download the Hermes Agent Python script for "Miner Pivot Calculator". Drop it into your Hermes tools folder or register it with the Hermes registry so other agents can call it directly.

Download Hermes tool script (.py)

Quick install

# Option 1: download and place in Hermes tools folder
curl -O https://hermesdispatch.dev/hermes-tools/miner_pivot_calculator.py
mv $(basename https://hermesdispatch.dev/hermes-tools/miner_pivot_calculator.py) ~/.hermes/hermes-agent/tools/

# Option 2: register via registry.py (edit tools/registry.py)
from tools.tools_miner-pivot-calculator import register
register()

🤖 Use this tool in your agent

✓ Agent-ready code

Copy the snippet below into your agent, newsletter, or script. The tool page at hermesdispatch.dev/tools/miner-pivot-calculator/ is the canonical contract: inputs, outputs, and formulas.

python 
# Hermes Dispatch Tool — Miner Pivot Calculator
# Source: https://hermesdispatch.dev/tools/miner-pivot-calculator/
# Description: Compare crypto mining profit vs local LLM inference revenue vs selling a GPU.
# License: MIT (generated by hermesdispatch.dev)
#
# INSTALL:
#   1. Save this file as ~/.hermes/hermes-agent/tools/miner_pivot_calculator.py
#   2. Restart Hermes or run /reset in a session
#
# MODES:
#   snapshot (default) — embedded daily GPU and crypto defaults, no API key needed.
#   live — uses YOUR CoinGecko / CoinMarketCap / mining-pool API key for current prices and network difficulty.
#          Pass api_key + provider in args. HermesDispatch never sees your key.
#
# MANUAL REGISTRY:
#   from tools.miner_pivot_calculator import register
#   register()

import json
import math

# Embedded snapshot — updated daily during site build.
SNAPSHOT = {"last_updated": "2026-07-04", "gpus": {"nvidia-geforce-rtx-5090": {"name": "NVIDIA GeForce RTX 5090", "vram_gb": 32, "hashrate_mh": 150, "power_w": 575, "inference_8b_tok_s": 82, "msrp": 1999, "market_price": 2200}, "nvidia-geforce-rtx-4090": {"name": "NVIDIA GeForce RTX 4090", "vram_gb": 24, "hashrate_mh": 130, "power_w": 450, "inference_8b_tok_s": 64, "msrp": 1599, "market_price": 1700}, "nvidia-geforce-rtx-3090": {"name": "NVIDIA GeForce RTX 3090", "vram_gb": 24, "hashrate_mh": 120, "power_w": 350, "inference_8b_tok_s": 48, "msrp": 1499, "market_price": 800}, "nvidia-geforce-rtx-4080-super": {"name": "NVIDIA GeForce RTX 4080 Super", "vram_gb": 16, "hashrate_mh": 95, "power_w": 320, "inference_8b_tok_s": 52, "msrp": 999, "market_price": 1000}, "nvidia-geforce-rtx-4070-ti-super": {"name": "NVIDIA GeForce RTX 4070 Ti Super", "vram_gb": 16, "hashrate_mh": 80, "power_w": 285, "inference_8b_tok_s": 40, "msrp": 799, "market_price": 800}}, "defaults": {"mining_revenue_per_mh_day": 0.006, "inference_price_per_1m_tokens": 0.1, "utilization_pct": 0.4, "electricity_cost_per_kwh": 0.12}}


def _ok(result):
    return json.dumps({"success": True, "data": result}, indent=2)


def _err(message):
    return json.dumps({"success": False, "error": message}, indent=2)


TOOL_NAME = "miner_pivot_calculator"
TOOLSET = "crypto"

SCHEMA = {
    "type": "function",
    "function": {
        "name": TOOL_NAME,
        "description": "Compare crypto mining profit vs local LLM inference revenue vs selling a GPU. Uses embedded daily snapshot by default; optionally fetches live crypto/network data with the user's own API key.",
        "parameters": {
            "type": "object",
            "properties": {
                "hashrate_mh": {"type": "number", "description": "GPU hashrate in MH/s."},
                "power_w": {"type": "number", "description": "GPU power draw in watts."},
                "electricity_cost_per_kwh": {"type": "number", "default": 0.12, "description": "Electricity cost per kWh."},
                "mining_revenue_per_mh_day": {"type": "number", "default": 0.006, "description": "Mining revenue per MH/s per day in USD."},
                "inference_8b_tok_s": {"type": "number", "description": "8B model inference speed in tokens/second."},
                "vram_gb": {"type": "number", "description": "GPU VRAM in GB."},
                "inference_price_per_1m_tokens": {"type": "number", "default": 0.1, "description": "Revenue per 1M output tokens in USD."},
                "utilization_pct": {"type": "number", "default": 40, "description": "GPU utilization for inference as a percentage (0-100)."},
                "market_price": {"type": "number", "default": 0, "description": "Current used market price of the GPU in USD (for sell estimate)."},
                "gpu_slug": {"type": "string", "description": "Look up hashrate/power/inference from embedded GPU snapshot, e.g. nvidia-geforce-rtx-4090."},
                "mode": {"type": "string", "enum": ["snapshot", "live"], "default": "snapshot", "description": "snapshot = embedded daily data; live = fetch current crypto prices with api_key."},
                "provider": {"type": "string", "enum": ["coingecko", "coinmarketcap"], "default": "coingecko", "description": "Live crypto provider. Requires api_key."},
                "api_key": {"type": "string", "description": "YOUR API key for the chosen provider. HermesDispatch never stores or transmits this key."},
                "coin_id": {"type": "string", "default": "bitcoin", "description": "Coin to fetch live mining revenue estimate for. Only used in live mode."}
            },
            "required": []
        }
    }
}


def _lookup_gpu(gpu_slug):
    slug = (gpu_slug or "").lower().strip()
    for key, info in SNAPSHOT["gpus"].items():
        if key.lower() == slug or info.get("name", "").lower() == slug:
            return dict(info)
    return None


def _calculate(hashrate_mh, power_w, electricity_cost_per_kwh, mining_revenue_per_mh_day,
               inference_8b_tok_s, vram_gb, inference_price_per_1m_tokens, utilization_pct, market_price):
    days = 30
    mining_revenue = hashrate_mh * mining_revenue_per_mh_day * days
    mining_power_cost = (power_w / 1000) * 24 * days * electricity_cost_per_kwh
    mining_profit = mining_revenue - mining_power_cost

    tokens_per_day = inference_8b_tok_s * 3600 * 24
    inference_revenue = tokens_per_day * days * (utilization_pct / 100) * (inference_price_per_1m_tokens / 1_000_000)

    sell_pct = 0.55 if vram_gb >= 16 else 0.35
    sell_estimate = market_price * sell_pct

    if mining_profit <= 0 and inference_revenue <= 0:
        verdict = "Sell the card"
        verdict_class = "red"
    elif inference_revenue > mining_profit * 1.2:
        verdict = "Pivot to AI inference"
        verdict_class = "emerald"
    elif mining_profit > inference_revenue * 1.2:
        verdict = "Keep mining"
        verdict_class = "amber"
    else:
        verdict = "Close call — test inference demand"
        verdict_class = "gold"

    return {
        "verdict": verdict,
        "verdict_class": verdict_class,
        "mining_profit_per_month": round(mining_profit, 2),
        "inference_revenue_per_month": round(inference_revenue, 2),
        "sell_estimate": round(sell_estimate, 2),
        "mining_revenue_per_month": round(mining_revenue, 2),
        "mining_power_cost_per_month": round(mining_power_cost, 2),
    }


def _snapshot(args):
    gpu_slug = args.get("gpu_slug", None)
    if gpu_slug:
        gpu = _lookup_gpu(gpu_slug)
        if not gpu:
            return _err(f"GPU not found: {gpu_slug}. Available: " + ", ".join(SNAPSHOT["gpus"].keys()))
        hashrate_mh = float(gpu["hashrate_mh"])
        power_w = float(gpu["power_w"])
        inference_8b_tok_s = float(gpu["inference_8b_tok_s"])
        vram_gb = float(gpu["vram_gb"])
        market_price = float(args.get("market_price", gpu.get("market_price", 0)))
    else:
        required = ["hashrate_mh", "power_w", "inference_8b_tok_s", "vram_gb"]
        missing = [r for r in required if r not in args]
        if missing:
            return _err(f"Missing fields: {', '.join(missing)}. Or provide gpu_slug.")
        hashrate_mh = float(args["hashrate_mh"])
        power_w = float(args["power_w"])
        inference_8b_tok_s = float(args["inference_8b_tok_s"])
        vram_gb = float(args["vram_gb"])
        market_price = float(args.get("market_price", 0))

    defaults = SNAPSHOT["defaults"]
    result = _calculate(
        hashrate_mh,
        power_w,
        float(args.get("electricity_cost_per_kwh", defaults["electricity_cost_per_kwh"])),
        float(args.get("mining_revenue_per_mh_day", defaults["mining_revenue_per_mh_day"])),
        inference_8b_tok_s,
        vram_gb,
        float(args.get("inference_price_per_1m_tokens", defaults["inference_price_per_1m_tokens"])),
        float(args.get("utilization_pct", defaults["utilization_pct"] * 100)),
        market_price,
    )
    result["mode"] = "snapshot"
    result["last_updated"] = SNAPSHOT["last_updated"]
    result["gpu_used"] = gpu_slug or "custom"
    return _ok(result)


def _live(args):
    api_key = args.get("api_key", "")
    provider = args.get("provider", "coingecko")
    if not api_key:
        return _err("Live mode requires your own API key from CoinGecko or CoinMarketCap. Pass it as api_key. HermesDispatch does not provide keys.")
    # Stub: implement real fetch with user's key.
    return _ok({
        "mode": "live",
        "provider": provider,
        "note": "Live fetch is a stub. Replace with your API call using api_key.",
        "coin_id": args.get("coin_id", "bitcoin"),
        "snapshot_mining_revenue_per_mh_day": SNAPSHOT["defaults"]["mining_revenue_per_mh_day"]
    })


def HANDLER(args):
    try:
        mode = args.get("mode", "snapshot")
        if mode == "snapshot":
            return _snapshot(args)
        if mode == "live":
            return _live(args)
        return _err(f"Unknown mode: {mode}. Use 'snapshot' or 'live'.")
    except Exception as e:
        return _err(str(e))


def register():
    """Manual registry hook."""
    try:
        from tools.registry import registry
        registry.register(name=TOOL_NAME, toolset=TOOLSET, schema=SCHEMA, handler=HANDLER)
    except ImportError:
        print("Hermes registry not found; skipping manual registration.")


if __name__ == "__main__":
    print(HANDLER({"gpu_slug": "nvidia-geforce-rtx-4090"}))
    print(HANDLER({"mode": "live", "provider": "coingecko"}))  # expect error: no api_key

Want early access to the next locked tool? Subscribe to The Hermes Dispatch.

🚀 Get AI automation insights daily

15:00 MST. One-click unsubscribe.

Subscribe