WireSift
WireSift Research · Independent

Independent research that cuts through consensus.

Answers to the critical questions in finance, investing, and the economy. Every claim traced to a primary source.

Featured findings · auto-rotatingClick any to explore →
Dataset
AI Supplier Leaderboard · S&P 500 · Q1 2026 · n = 485
Google commands the hyperscalers. OpenAI and Anthropic neck-and-neck.

Across 485 S&P 500 earnings calls, Google leads at 42, over 2x Microsoft (21) and nearly 3x AWS (15). OpenAI (39) edges Anthropic (38) by one for the model-provider lead; NVIDIA (38) tied with Anthropic for third.

Google
Hyperscaler
42
9% of n
OpenAI
Model provider
39
8% of n
Anthropic
Model provider
38
8% of n
NVIDIA
Chip vendor
38
8% of n

+4 more on the full leaderboard →

WireSift Research · 485 of 503 S&P 500 reported · as of 2026-07-05
Subscribe

Independent research, direct to your inbox.

New findings, the data behind them, and the methodology behind every claim. Free.

By subscribing, you agree to our Privacy Policy.

About WireSift

Deep, evidence-based research, delivered as interactive data.

WireSift is independent research for sophisticated investors and finance professionals. We do the rigorous work consensus skips, and we render every finding as a click-through visual that traces back to the source evidence.

Deep research

Investigations into the questions consensus glosses over. Not takes, not summaries. Original analysis with primary sources.

Evidence-based

Every claim linked to a source. 10-K filings, earnings transcripts, regulatory documents. If we can't source it, we don't say it.

Data storytelling

Findings rendered as interactive visuals you can click through to the underlying quote, transcript, or filing. Not static charts. Auditable evidence.

Built for capital allocators

Sophisticated investors and finance professionals who make or advise on capital decisions and want better inputs, not retail tip-sheet content.

The methodology

How the research gets done.

Four principles that govern every project. Public, auditable, versioned.

Tier-rated sourcing

Every claim anchored to a primary source. Tier 1 is under-oath disclosure (10-K filings, earnings calls, regulatory documents). Lower-tier sources are explicitly labeled.

Dated, versioned data

Every dataset carries the date of the data behind it. Schemas are versioned and old records are kept, so a published finding can be reproduced.

Cross-validation

Quantitative work cross-validated against independent models. A stratified sample of every extraction is independently re-validated by a second frontier model.

Documented methodology

Source tiers, extraction schema, validation layers, and editorial choices are documented in public. Audit requests are welcome.

Stay informed

Get the next finding before it makes the headlines.

Free. New findings, the data behind them, and the methodology behind every claim. Independent research direct to your inbox.

By subscribing, you agree to our Privacy Policy.