Baselight (@BaselightDB)

Baselight Weekly Update: 80K+ datasets, nearly 500K tables, and now DuckLake support. Baselight keeps growing - both in scale and interoperability. This week, we crossed 80,000 indexed datasets, moved closer to 500,000 tables, and integrated DuckLake with the Baselight catalog. With this milestone, Baselight now supports 3 lakehouse formats: - Baselight's native format - Apache Iceberg - DuckLake Current catalog scale: 512,193,557,002 rows (up 3B this week) 488,617 tables (up 4K this week) 80,296 datasets (up 1K this week) A broader catalog, more lakehouse compatibility, and one simple interface to query the world's structured data.

We’re tracking 767 AI models across every major lab. What the data says right now: - Top Intelligence Score: Claude Opus 4.8 (61.4) & GPT-5.5 (60.2) - Best value: Gemini 3.5 Flash - 55.3 score at $1.50 prompt / $9 completion (1M tokens) The gap between “best” and “best value” is closing fast. This week we added AI Models Intelligence to Baselight: normalized model pricing, benchmarks, capabilities, endpoints, and performance data from sources including OpenRouter and Artificial Analysis. We also added NOAA NCEI datasets with climate and environmental data from 1763 to present. Baselight now has: - 509,001,889,281 rows - 484,031 tables - 78,773 datasets That’s +4B rows, +7K tables, and +2K datasets this week. Structured data is becoming the intelligence layer for everything - from climate history to the AI model economy.

What does the world look like in 2026 - by the numbers? The GDELT Events Dataset has been tracking every major news event globally, every 15 minutes. Here's what 2026 looks like so far: 15.7 million events recorded in just 5 months 62% of events are cooperative - but tone is still negative March 2026 was the most conflict-heavy month of the year 🇺🇸 The US is the most-covered country - 1 in 3 global events 🇮🇷🇮🇱 Iran & Israel both crack the global top 5 Some curiosities regarding Top 10 GDP Nations: 🇺🇸 USA scores the most negative tone in the dataset - every single month. Not one big crisis. Just relentless volume. 🇯🇵 Japan is the only country to go positive - thanks entirely to a single Toyota CEO story in February. 🇫🇷 France improved in May - but only because previous months had kidnappings, firefighter strikes, and a judge scolding Shia LaBeouf. The bar was low. This isn't just data. It's the pulse of the planet. Explore it on Baselight: @gdelt.events_v2

Everyone saw the upset after the final whistle. Baselight had already spotted something unusual days earlier. Baselight Daily Insights are produced by fully autonomous agents that query our verified structured data every day to find anomalies, outliers, inflection points, and signals worth human attention. Today they flagged Torreense’s shock Taça de Portugal final win over Sporting CP as a major outlier: a second-division side winning 2–1 after extra time, despite Sporting being priced around 1.16 and Torreense around 14.5. But the more interesting signal came days earlier. On May 22, Baselight flagged unusual odds divergence around Torreense: Betano priced the win at 32.0, while market consensus was around 17.19 - an 86%+ divergence. Baselight did not “predict the upset”. It surfaced a market anomaly that deserved human review: a possible stale price, model disagreement, or risk miscalculation. That is the goal: autonomous agents turning verified structured data into explainable, auditable signals before they become obvious. Link to the May 22 insight in the comments.

Cybersecurity data is accelerating fast. Baselight now includes NIST NVD data, and the latest numbers are striking: The week of May 11, 2026 saw 1,889 CVEs published in a single week - the highest 7-day total ever recorded in the NIST National Vulnerability Database. And we all know what is driving part of this: GenAI is getting very good at finding vulnerabilities. 📊 Jan avg: ~1,094/week 📊 Feb avg: ~1,198/week 📊 Mar avg: ~1,435/week 📊 Apr avg: ~1,344/week 📊 May so far: ~1,749/week With NVD now in Baselight, CVEs become structured, queryable data that can be correlated with almost 500K tables and 500B+ rows across all knowledge domains. That is the real power: not just tracking vulnerabilities, but connecting them to the wider world of structured data.