Skip to content
Emerging.

Method

How fields are staged, how momentum is computed, and where every row comes from. So far: 17 fields, 362 companies, 427 events, 485 papers; growing daily.

Stages

Each field sits in one of four stages, set by hand from a dated, sourced fact shown on its page (the Changes page keeps every earlier stage):

  • Research: Lab results; no working prototype of a product yet.
  • Prototype: Working prototypes or demonstrators; not yet in customer use.
  • Pilot: First customers or pilots paying or deploying; not yet routine.
  • Early commercial: Sold and in routine use by early adopters; far from mainstream.

Momentum

Momentum answers "how quickly is this field moving right now, compared with the others". It is built from three counts and nothing else:

  • Paper growth: journal articles, reviews and conference papers in the field's OpenAlex topics (Crossref-indexed; preprint duplicates left out) published in the 12 months ending two months ago, against the 12 months before. The two most recent months are skipped because journals are indexed with a lag; they are drawn lighter on the charts. January bars run high because papers dated only by year are filed under January 1; comparing whole 12-month windows cancels that out.
  • Funding: the US dollars of funding rounds tracked here in the last 365 days, as the sources report them.
  • Events: dated events tracked here in the last 90 days (researched and scanned).

Each count is ranked across the tracked fields (0 for the lowest, 100 for the highest; ties share a rank). Momentum is the mean of the three ranks, so it is relative: it compares fields with each other this week. A count with no data ranks lowest; nothing is estimated. Funding and events only count what this site has found, so a field with little coverage can look slower than it is. Every week a snapshot of these numbers is stored per field (/api/snapshots), so the history compounds.

Where rows come from

  • Researched by hand (verified): each field's definition, sub-areas and stage evidence; companies and their role; labs; dated events and funding rounds. Every row links the page it was checked against: official newsrooms and agencies first, then papers and SEC filings, reputable press only when nothing official exists.
  • Daily scan (Unverified until checked): official RSS feeds (listed below), tied to fields by keyword rules; arXiv's daily listings for new papers; SEC EDGAR 8-K filings of the public companies tracked here.
  • Weekly run: OpenAlex monthly paper counts, most-cited papers, and the researchers and institutions publishing most per field (institutions that are companies become candidate companies); rounds on funding.fru.dev that fit a field; companies on companies.fru.dev whose description fits a field, enrolled as candidates.
  • Optional model note: when a model key is set, one call a week writes a one-line "this week" note per field from that week's event titles only. Without a key nothing changes.

Every row carries its source URL, source type, first seen and last seen dates, a verified flag and the run that found it. Nothing is deleted; rejected finds are hidden and kept.

Feeds read daily

How each field is measured

Gaps

  • Patents: there is no free patent API that works without a registered key (PatentsView now needs one and its search host did not answer; USPTO's Open Data Portal, EPO OPS and Lens need keys too). Patents appear only where someone added one by hand from its Google Patents page; there is no automated patent scan.
  • Funding coverage: rounds come from hand research and funding.fru.dev, which follows data and AI companies first; frontier hardware and biology rounds are under-counted until added.
  • Keyword matching: the scan ties news to a field by words in the title and summary, so some finds are off-topic. They stay Unverified until checked.

About this data

Facts as stated by the sources linked, on the dates shown. Events, papers, patents and funding from agency and company newsrooms, SEC EDGAR, arXiv, OpenAlex and funding.fru.dev, scanned every day and counted every Tuesday. What the scan finds is checked by hand; unverified items are marked. The list grows as new work appears. Stages and momentum are summaries for orientation, not forecasts or investment advice. Photos from Wikimedia Commons and official company pages, with credit. Logos via logo.dev; trademarks belong to their owners.

Data

JSON at /api/technologies, /api/events, /api/companies (the companies.fru.dev feed), /api/rounds, /api/papers, /api/researchers, /api/snapshots and /api/changes; plain text at /llms-full.txt. For AI agents: /agents and /openapi.json.

Frontier tech by email

Tuesdays, only in weeks with new milestones, rounds or stage changes in the fields tracked here.

Double opt-in. Unsubscribe any time.