About The Analysis
A public record of what companies said, what actually happened, and how we read the pattern. Built to be readable, cited, and correctable.
The disclaimer
The Analysis is a public record of what companies have said, what has actually happened, and jain.com's interpretation of the pattern. Every factual statement is cited to a primary source; every citation shows the original URL, our cached copy retrieved on a specific date, and (when available) an independent archive.org snapshot.
Analysis and commentary are labeled as such - they are our opinions, offered in good faith and open to correction. Nothing here is investment advice, legal advice, or a recommendation to trade any security. jain.com holds no advisory relationship with the companies covered.
Errors and disputes: contact us and we will investigate and, if warranted, correct or retract on the record.
How we distinguish fact, observation, and analysis
Every rendered block carries a tag naming its tier:
Verbatim quotes and paraphrases of a source document. Must link to at least one source with a specific quote span. A day after publication, and again the day after that, we re-fetch the cited source and check that the quote is still in it. A claim the source turns out not to support is withdrawn, and stays publicly visible as a tombstone with the reason attached, so a correction is as easy to find as the original claim.
Reproducible arithmetic on cited facts (e.g., "this milestone is 17 months past its originally stated date"). Names the computation used so a reader can reproduce the arithmetic.
Our interpretation of the pattern. Not fact-grounded to a source (there is none - it is our reading), but every analysis block cites the underlying claims it interprets. If any of those claims are later withdrawn or contradicted, the analysis block is automatically flagged for review, still rendered publicly with a banner noting the underlying revision.
How the source cache works
When we cite a source, we (a) fetch the live URL, (b) compute its SHA-256, (c) store the bytes on our server under /analysis/sources/<sha>/, and (d) push the URL to archive.org for an independent third-party snapshot. Every citation renders a tri-link (Live / Cached / Archive.org) so a scrubbed or moved original never silently breaks our record. Weekly, we re-check every cited URL: if the live content's hash no longer matches our cache, a small note appears on the source page.
Paywalled sources: we store the URL, our transcribed quote, and the archive.org link, but do not re-host the body.
Court records: any filing we purchase from PACER is fetched through the RECAP project, which automatically contributes a copy to the public RECAP Archive at the non-profit Free Law Project. Every document we buy becomes permanently free for everyone.
Where the photographs come from
Headshots on people pages are the official portraits published by the companies themselves - a leadership, management or board-of-directors page on a corporate site. Directors often sit on several boards, so a face missing from one company's site is frequently published on another we track; we search across every tracked company's pages, not only the ones where that person holds a role.
A photo is only published when we can tie it to the named filer. The strong case is an official page of a company where the person holds a role, with their name in the image's alt text or the surrounding card and corroborated by the image filename. Weaker matches go to a second check - the page context is read against the roles that person's SEC filings assert - and then, if that still does not settle it, to a search for an independent source placing that person in that role at that company. A source that contradicts the match rejects it. Anything left unresolved is held back for human review rather than published; we would rather show initials than the wrong face. Every attempt, successful or not, is logged.
We also drop any image that answers to more than one name. Company sites sometimes serve a single placeholder file for several people; the caption is right but the face is not, and an image that could be two people identifies neither.
As with documents, we cache the image: the bytes are stored under /analysis/sources/<sha>/ and the page serves our copy, so the portrait survives the company redesigning its site. The footnote beneath each photo names the page it came from, the date we retrieved it, and links to that cached copy. Rights remain with the original publisher; we display these portraits as identification in reporting on the people concerned. Any company or individual who wants a portrait removed only has to ask.
Market data
Share prices and market caps are read once per weekday after the US close, and never more than once a day. The group market-capitalisation figure at the top of the front page is the sum of that day's readings across every active public company we track - it is computed at render time from the stored daily snapshots, not maintained as a separate number. A day on which one company's reading did not arrive is labelled partial rather than shown as a fall. Analyst ratings, earnings surprises and company metrics refresh on the same weekday cadence, so a new rating is picked up within one trading day of publication.
Insider tables show two kinds of dollar figure. The Amount column is what the filing itself reported - actual sale proceeds, and $0 for a free grant. The @ last price columns restate the same share counts at the most recent price we read, so a grant that reported no cash value still shows what those shares are worth. They are a valuation, not a cash flow, and they move whenever the market does.
Also here
Patents · the global portfolio, each with the cached original PDF and a plain-English overview. Ventures · the companies and engineering projects behind this site, in one place.