The Crypto AI Gem Score™ methodology

Explosive upside potential

The Gem Score answers one question: how attractive is this asset TODAY as a potential high-upside hidden gem? It deliberately does not answer “how big or reputable is this project?”. Alongside the six research criteria — which together form the project quality score — we compute an independent upside score from market-cap headroom, fully diluted valuation, circulating supply and unlock overhang, distance from the previous all-time high, valuation versus measured TVL and revenue, catalyst momentum, exchange accessibility and tradable liquidity. Headroom is gated by fundamental quality, so a weak micro cap cannot outrank a strong small cap, and mega caps carry diminishing upside because each further multiple demands far more capital. Verified on-chain value, community size, competitive moat, a proven record of shipping product, real earnings relative to size and low unlock pressure can each lift a score but never lower it. A team of elite profile — founders whose record rivals the builders of Sui, Aptos or Monad, as judged by an AI review of public reporting — earns an extra bonus, and projects focused on today’s leading sectors (AI, DePIN, real-world assets, Layer 2 and zero-knowledge) earn a smaller one. Memecoins and tokenised shares never qualify for the sector bonus.

• Extreme Upside90+ · very large market-cap headroom
• Very High Upside80+ · large market-cap headroom
• High Upside70+ · above-average headroom
• Moderate Upside60+ · moderate headroom
• Limited Upside0+ · limited headroom

These ranges are analytical estimates derived from measured data, not guaranteed price predictions.

The six quality criteria

How the score is calculated

The Gem Engine™ blends the six quality readings with the upside reading using a proprietary weighting. The exact formula is not published, but it is applied identically to every asset, so scores stay consistent and comparable. A major established asset can hold a very high quality reading yet a moderate Gem Score, because its remaining market-cap headroom is smaller. That is intentional.

Rating bands

• ELITE GEM95–100
• STRONG GEM90–94
• HIGH POTENTIAL85–89
• PROMISING80–84
• ON WATCH70–79
• SPECULATIVE60–69
• HIGH RISK0–59

Confidence and data gaps

Each analysis carries a confidence level from 0–100 reflecting how much verifiable evidence the model found. When information is missing — an anonymous team, no audited contracts, undisclosed funding — the model records the gap explicitly and lowers both the relevant criterion score and confidence rather than guessing. Every coin page lists the data gaps behind its score.

Data sources and refresh cadence

Market data (price, market cap, volume, supply, all-time highs) comes from public market data providers and refreshes continuously. Qualitative research draws on project documentation, developer activity, community metrics and exchange listings available at analysis time. Scores are re-generated on a rolling schedule and stored append-only, so the full score history for every asset is preserved.