Talkshi Research
Talkshi Research publishes open, source-audited datasets on the AI-agent economy: a funding tracker covering 100 companies with $4.46 billion in disclosed 2026 rounds, and a placement-by-placement audit of a 219-entry agentic-commerce landscape map. Every dataset is downloadable as JSON, CSV, or Markdown and checked against first-party sources. It is operated by Raymond Xu.
What datasets does Talkshi Research publish?
AI Agent Funding Tracker 2026
100 dated funding rounds announced by AI-agent, payments, and identity companies between January 12 and July 8, 2026. 95 rounds disclose an amount, totaling $4.46 billion. Each record carries the announcement date, amount, round wording, investors, stated use of proceeds, official website, and a tier label: 46 companies build enabling agent infrastructure, 22 build direct agentic-commerce infrastructure, and 32 are economic participants or vertical agents. Every company links to its own source-audited analysis.
Agentic Commerce Landscape Audit
Every placement on a published agentic-commerce landscape map checked against current first-party sources: 219 placements and 209 distinct labels, with duplicates, non-companies, and one unresolvable logo called out individually. The full working record is the company-by-company ledger (Markdown).
How can this research be cited?
Cite the dataset name and Talkshi Research, and link the dataset page — for example: "AI Agent Funding Tracker 2026, Talkshi Research, talkshi.com/research". The machine-readable files carry the same records as the HTML pages, so an agent can cite a specific row by company and announcement date.
How the funding tracker is produced
- Start with a dated public company, investor, issuer-release, or reputable reporting source.
- Record the announcement date, amount, round wording, investors, stated use of proceeds, official website, and one additional funding-context fact.
- Keep totals-to-date, extensions, combined rounds, strategic financings, undisclosed amounts, and planned raises explicitly labeled.
- Map each company to a direct, enabling-infrastructure, or economic-participant tier without relabeling every AI company as a payment startup.
- Generate the HTML, Markdown, JSON, and CSV from that single structured record, then run checks for missing facts, duplicated language, links, metadata, and schema.
Automation disclosure
The collection is produced with software-assisted research and structured generation. Automation is used to keep repeated facts synchronized across formats; it is not permission to invent a fact. Missing disclosures are labeled as missing, and proposed Talkshi integrations are clearly separated from reported company claims.
Corrections and independence
Talkshi has no affiliation with a company in a research tracker unless the relevant page says otherwise. Funding data changes, and source pages can be corrected after publication. Send a correction and its supporting URL to [email protected].