Tom Lee

Where the scored predictions come from

Sources searched

• Fundstrat Direct research archive (his own notes)

  Tom Lee's own research notes on Fundstrat's retail site, published as FS Insight and now as Fundstrat Direct. Everything under his byline was listed out through the site's own machine-readable content index: 1,936 dated notes plus 296 short intraday alerts, running from October 2018 to the end of July 2026. This is the densest primary record of his forecasting, because his daily 'First Word' note, his 'Macro Minute' video write-ups and his year-ahead outlooks state his S&P 500 year-end target, his bitcoin and ether targets and his near-term market calls in his own words, and the full text of the longer notes is publicly readable. His research from before October 2018 was never published openly, so that earlier stretch of his record has to come from television and press coverage instead.

  How many there are

  2232

  How many we read

  2,232

  Time span

  2018-10-02..2026-07-30

  How we sampled

  Every one of the 2,232 items was screened by title and web address against a broad forecast-word list: commitment words (will, won't, going to, set to, expect, predict, forecast, bet, we think, we believe, our view, target), timing words (by a named year, by year-end, within, soon, imminent, this year, next year, coming, decade), and outcome words (rally, surge, plunge, crash, collapse, correction, recession, bear market, bull market, melt-up, peak, bottom, record high, upside, downside, buy, sell), together with price levels, dollar figures and percentages and the names of the things he forecasts (the S&P 500, the Nasdaq, bitcoin, ether, the Fed, inflation). That list is a memory aid, not a gate that decides what counts. It flagged 2,072 items and left a remainder of 160, almost all of them very short intraday alerts whose titles are cut off after about seventy characters. Thirty of those 160 were drawn at random and read end to end: fifteen of the thirty contained at least one dated, checkable market call that the truncated title had hidden, a miss rate of roughly one in two. Because that remainder is small and its miss rate is that high, the honest rule is to read all 160 rather than trust the word list on them. Sixty-seven flagged notes, chosen to spread across every year and to favour explicit price targets and year-ahead outlooks, plus all thirty randomly-read items, are listed for detailed extraction.

• Fundstrat institutional client portal (his macro notes)

  The same firm's institutional site carries a parallel stream of Tom Lee's writing: 867 pieces under his byline from September 2022 to the end of July 2026, mostly 'Flash' commentaries and the written accompaniment to his Macro Minute videos. It was listed out in full through the site's own content index and the notes are readable. Most of it restates what he publishes on the retail side within a day, so it serves mainly as a second copy for checking exact wording and dates, and only the pieces with no retail counterpart add genuinely new forecasts.

  How many there are

  867

  How many we read

  867

  Time span

  2022-09-18..2026-07-30

  How we sampled

  All 867 notes under his byline were listed with their dates and sections and screened against the same forecast-word list used on the retail archive, then matched against the retail notes to identify the genuine duplicates. Because the overlap is heavy and the writing is the same, no separate random audit of the non-flagged remainder was run here; the audit done on the retail archive covers the same material. A small number of items with no retail twin are listed for extraction, along with the author index page so the rest can be worked through.

• CNBC written coverage

  Tom Lee is a CNBC contributor and has been quoted or written about there continuously since his years as J.P. Morgan's chief U.S. equity strategist. The network's own site search was worked through page by page under several phrasings of his name and his firm, producing 1,278 items that genuinely concern him; 384 of those are written articles and blog posts, running from March 2010 to June 2026. These carry the earliest part of his public record, including the year-end index targets he issued before Fundstrat existed, and they usually quote him directly rather than paraphrasing.

  How many there are

  384

  How many we read

  384

  Time span

  2010-03-18..2026-06-25

  How we sampled

  All 1,278 CNBC items were listed with title, date, summary and web address, then separated into written pieces (384) and television clips. Screening the written ones against the same broad forecast-word list flagged effectively every one of them, because a CNBC headline about Tom Lee almost always contains a number, a direction or a forecast verb; with nothing left over there was no non-flagged remainder to audit at random here, and the honest reading is that the word list simply cannot discriminate on this material. Selection was therefore done by hand instead: roughly three or four items per calendar year from 2010 to 2026, chosen for explicit dated claims such as index targets, bitcoin price forecasts and recession calls, plus every item that names a specific number in the headline.

• CNBC televised appearances

  Television is where Tom Lee makes most of his calls. 894 CNBC clips featuring him, from December 2012 to July 2026, were listed through the network's own site search. They are segments from Halftime Report, Closing Bell and Closing Bell Overtime, Squawk on the Street, Fast Money, Power Lunch, Money Movers, Last Call and TechCheck, and each clip page carries a dated summary of what he argued. His actual words have to come from the clip itself; where the audio cannot be reached, the dated summary and a reputable outlet's direct quotation of the same remark are the fallbacks, and anything still relayed at second hand is held back from counting until it is traced to what he actually said. This is the single largest body of his forecasts and the one most thinly sampled so far.

  How many there are

  894

  How many we read

  894

  Time span

  2012-12-06..2026-07-27

  How we sampled

  All 894 clips were listed with their dates and summaries and screened against the same forecast-word list, which flags nearly all of them for the same reason as the written coverage: a clip about Tom Lee is nearly always headlined with the call he is making, so the list cannot discriminate and there is no meaningful non-flagged remainder to audit at random. Selection was therefore by hand across the whole span, roughly three or four per year from 2012 to 2026, favouring clips whose summary states a specific number and a deadline (index targets, bitcoin price forecasts, named-month calls) and the full-length sit-down interviews over the thirty-second reaction hits.

• Fundstrat YouTube channel

  Fundstrat's YouTube channel, at about 119,000 subscribers, is Tom Lee's own video outlet. Its back catalogue was listed exhaustively: 632 items on the main uploads list plus 10 more on the separate live-streams list that the uploads list does not show, 642 in total. Of those, 203 are identifiably his by title; the rest belong to colleagues who cover technical analysis, digital assets and Washington policy, and are out of scope for him. Titles on this channel are unusually explicit about the call being made, often naming the index level or price outright, and the spoken words come from each video's captions.

  How many there are

  642

  How many we read

  642

  Time span

  2019-12-05..2026-07-30

  How we sampled

  Both the main uploads list and the separate live-streams list were pulled in full, 642 videos in total, and every title was screened for whether the speaker is Tom Lee rather than a colleague, leaving 203 that are his. Those 203 were then screened against the same forecast-word list, which flagged 142; the 61 that did not flag are mostly interview trailers, channel explainers and topic primers rather than call-making videos, so no separate random audit was run on them. Forty-eight videos are listed for extraction, weighted towards the last two years and towards titles naming a specific index level, price or deadline, together with the live-streams list and the channel's own interview playlist so the older material can be worked through.

• Podcast and long-form interview appearances

  Tom Lee is a frequent podcast guest, and podcasts are where he explains the reasoning behind a target at length. Podcast directories were searched under several phrasings of his name, his firm and BitMine and cross-checked against individual show archives, producing at least 71 episodes in which he actually appears rather than merely being discussed, from October 2019 to May 2026. Recurring hosts include The Compound and Friends, Odd Lots, Unchained's Chopping Block, RiskReversal, Bankless, Wealthion, Anthony Pompliano's shows, ARK Invest's podcast, The Meb Faber Show, The Scoop, ETF Prime and CNBC's own show feeds. This count is a floor, not a complete census, because no single directory indexes every show. Some episodes have published transcripts, which is the best evidence of his exact words; where none exists the audio itself or a reputable outlet's direct quotation is used, and a host's third-person recap of what he said is not treated as his wording.

  How many there are

  71

  How many we read

  71

  Time span

  2019-10-01..2026-05-15

  How we sampled

  Podcast directories were searched under four phrasings of his name, his firm and BitMine, returning 151 episodes that mention him. Those were separated by hand into episodes where he is the guest and episodes where other people simply discuss him or his firm's ether holdings, which left at least 71 genuine appearances. Named episodes found through general web searching but absent from the directory results, including shows by Meb Faber, Anthony Pompliano and ARK Invest, were added on top. Because this list is assembled from search rather than from any single complete index it is a floor, and there is no defined non-flagged remainder here to audit at random. Forty-nine episodes are listed for extraction, weighted towards long-form interviews where he states numbers and deadlines and towards the episodes that have a published transcript.

• Published books

  Tom Lee has not written a book. Book retailers, his firm's own biography of him, his encyclopaedia entry and general web searching turn up no title he authored; the only books associated with his name are ones he has recommended to readers. There is therefore no book record to mine and nothing missing from this side of his output.

  How many there are

  0

  How many we read

  0

  Time span

  not applicable

  How we sampled

  Searched the major book retailers, his firm's own biography page, his encyclopaedia entry and the open web for any title written by him. The only results are reading recommendations he has made to subscribers, not works of his own, so there was nothing to sample and nothing to leave out.

Extra web searches

Known gaps

Three gaps stand out. First, the years before October 2018. His firm's site does not publish his notes from earlier than that and his J.P. Morgan research was never public, so everything from 2010 to 2018 rests on television clips and press coverage. That captures his headline index targets but not the reasoning around them or the smaller calls he made in between. Second, television. Nearly nine hundred CNBC segments feature him and that is where a great many of his calls are actually made, but only a sample of them has been worked through so far, so his broadcast record is represented rather than complete; his short reaction hits, where a single sentence often contains a dated call, are the biggest part of what is still untouched. Third, his short intraday alerts. Their titles are cut off mid-sentence, which hides the forecast, and reading thirty of them at random found that about half contained a dated call the title never revealed, so the small unflagged remainder of that stream needs reading in full rather than filtering by wording. Two smaller gaps. His posts on X, where he publishes brisk one-line market calls under the handle @fundstrat, cannot be listed out without access to the platform's own archive, so they are absent entirely. And his firm's newsletter feed cannot be used to gather his writing automatically: the feed the site offers has not been refreshed since spring 2025, carries only opening extracts rather than full notes, and mixes his work with colleagues' under a single heading, so his archive has to be collected piece by piece instead. He has written no books, so nothing is missing on that side.

What we didn’t score & why

If-then call whose condition never happened

22

Implied call, no deadline stated

355

Out of scope for scoring

286

Outcome not yet known

66

Quote status reporter paraphrase

9

Secondhand quote, not traced to the original yet

1

Set aside for human review

126

Verification unverifiable

9