The Engine
How Sourcery researches for you
01 · Understand
Break Down The Question
Sourcery first understands what is actually being asked, identifying hidden assumptions, ambiguities, and key dimensions worth investigating.
02 · Plan
Build Research Strategy
It maps sub-questions, evidence requirements, likely controversies, and the types of sources needed for rigorous analysis.
03 · Generate
Create Search Angles
Multiple focused search queries are generated — designed for breadth, balance, and discovery of both primary and critical sources.
04 · Search
Discover Sources
The web is searched across multiple angles to uncover documents, datasets, filings, papers, reports, articles, and technical material.
05 · Clean
Remove Noise
Pages are cleaned locally — stripping ads, navigation, cookie banners, boilerplate, and irrelevant markup before analysis begins.
06 · Extract
Capture Evidence
Important facts, numbers, dates, entities, assumptions, and uncertainty signals are extracted into structured research notes.
07 · Expand
Go Deeper
Related URLs, missing questions, cited papers, and unexplored directions are recursively followed to configured depth.
08 · Cross-check
Find Contradictions
Claims are compared across sources to surface conflicting numbers, disputed interpretations, and logical inconsistencies.
09 · Remember
Persist Research Memory
Sources, claims, entities, summaries, and checkpoints are stored for auditability, reuse, and resumable research.
10 · Synthesize
Build Draft Insights
Evidence is synthesized in parallel chunks so important details are preserved even across large research runs.
11 · Verify
Protect Citation Integrity
References are rebuilt deterministically so every citation points back to a real, traceable source.
12 · Deliver
Final Report
The final output is a fully referenced report with linked citations, contradictions preserved, and every claim traceable.
The Output
Published Research Reports
The Hidden Financial Bubble in AI Infrastructure
A deep dive into debt-funded AI infrastructure, GPU-collateralized lending, hyperscaler CapEx, CoreWeave’s leverage, NVIDIA’s circular financing model, data center overbuild risk, and telecom-bubble parallels — with scenario analysis and open questions around utilization, debt maturity, and systemic contagion.
How Vulnerable Is Asia to a Gulf Energy Shock?
India, China, Japan, and South Korea assessed against the 2026 Iran-Israel war and the largest supply disruption in IEA history. Reserve levels, Hormuz exposure, IEA membership, ISPRL financial health, and three-scenario outlook — all citation-grounded across 17 sources.
The Coming Power Wars Between Humans and Datacenters
AI-driven data centers assessed against rising U.S. electricity prices, grid bottlenecks, capacity-market shocks, fossil-fuel lock-in, water stress, subsidies, and the unresolved question of who gets power priority during shortages. Citation-grounded across 54 sources.
The Hidden Infrastructure of Financial Markets
A deep dive into the clearing houses, default waterfalls, margin systems, settlement gaps, and systemic risks that quietly keep financial markets running. Covers CCP fragility, the Einar Aas default, COVID-era stress tests, cross-CCP contagion, blockchain experiments, and why payment rails remain one of the least visible layers of market infrastructure.
What Happens If Compute Becomes a Sovereign Reserve Asset?
A deep analysis of whether nations could treat AI compute like strategic reserves: sovereign wealth funds buying data centers, national compute reserves, compute-backed collateral, and the possibility of future compute sanctions — grounded across 38 sources.
The Financialization of Compute Futures
How GPU-hours are becoming tradable, securitized, and potentially hedgeable assets. Covers compute spot markets, GPU-backed debt, futures exchanges, fixed-price inference contracts, energy hedging, tokenized compute, and the structural risks behind turning AI infrastructure into a financial market.
Next report in progress…
Sourcery is crawling, extracting, and synthesizing. Check back soon for the next deep-research dispatch.
Another topic under research…
More reports will appear here as Sourcery completes each deep-research run.
Why Sourcery
Built differently from the start
Evidence before conclusions
Sourcery does not start by writing. It starts by collecting evidence. Every conclusion in the final report is backed by source material discovered, fetched, cleaned, and analyzed during that exact research run.
Research that thinks before it searches
Instead of firing one search query and hoping for the best, Sourcery first maps sub-questions, risks, disputed areas, and evidence gaps — then investigates from multiple angles like a careful analyst would.
Contradictions are surfaced, not hidden
When sources disagree, Sourcery does not smooth it over with confident prose. Conflicting numbers, disputed claims, and competing interpretations are preserved and explicitly carried into the final report.
Every citation is traceable
Inline citations are not decorative. Every numbered reference maps back to a real source that was actually fetched, parsed, and included in the same research session — with references rebuilt deterministically.
Built for depth, not summaries
Sourcery recursively follows evidence, cited papers, source links, datasets, and unanswered questions — digging deeper where the strongest signals exist instead of stopping at the first page of search results.
Research gets smarter over time
Previously analyzed sources, extracted evidence, discovered entities, and validated findings are remembered across runs — making follow-up research faster, cheaper, and progressively more informed.