Meta Is Up 6.1% Since Friday. Monday’s Leg Was Over Before the AI Launch Landed
Meta stock closed 6.10% higher across two US sessions, and the one-minute bars show Monday’s part of it was over before the announcement post went live. Meta set its Monday high within three minutes of the post, which published at 1:08 p.m. ET, and the last bar wholly before the post had the stock 3.81% above its October 9 close, higher than where the session finished. Whether a jump this size says anything about the weeks ahead is a separate question, and the record of 71 past runs of 6% or more does not answer it.
Meta published “Introducing Muse Wonder and Muse Wonder Full: Frontier Models for Agents” on its own newsroom, dated October 12, 2026. Meta claims “state-of-the-art performance on personal and professional agentic tasks” in domains including law, finance, health and coding, with no third-party benchmark published alongside the claim. No pricing was disclosed.
What the one-minute bars show about Monday
Monday opened at $728.50, 1.37% above the October 9 close of $718.67, so the advance began as an overnight gap. Five minutes in, the stock was at $734.08, 2.14% above that close, and it kept building through the morning. The session high, $752.24, came in the bar covering 1:05 to 1:06 p.m. ET, within three minutes of the post, 4.67% above that close. The last reading wholly before the post, the bar covering 1:07 to 1:08 p.m. ET, was $746.02, 3.81% above that close. Monday closed at $742.96, 0.41% below that last pre-post reading and 1.23% below the high. Tuesday opened at $736.09, 0.92% below Monday’s close, hit its session high of $771.27 in the 9:59 to 10:00 a.m. ET bar, and closed below that at $762.50, an advance complete inside the first half hour. Monday traded 24,882,400 shares against a five-session mean of 13,934,200, about 1.8 times, and Tuesday 22,627,792. One-minute and daily bars and volumes from Yahoo Finance, pulled October 13, 2026.
None of that makes the launch unimportant. It does mean the two-session gain cannot be read as the post’s effect: about three fifths of it, 62%, was in the price before the post was public, and the rest came a session later from a lower open. What the bars time is the newsroom post rather than the moment the market learned, and the models were available to developers when it published.
The move, and the index that went the other way
Meta closed at $762.50 on October 13, 2.63% above Monday’s close and 6.10% above the October 9 close. Prices from Yahoo Finance daily bars, data as of October 13, 2026.
The index did not follow. Over the same two sessions the Nasdaq Composite fell 0.59% to 27,204.38, and the S&P 500 finished 0.08% lower at 7,805.45. AMD was down 0.42% and Nvidia 0.33% over the pair, while Microsoft added 1.57% and Apple 1.38%. Both chipmakers finished Tuesday itself higher, AMD up 1.90% and Nvidia 0.97%, so Tuesday was not a session where Meta rose and the AI complex fell. What sets Meta apart is the size of its move, not the direction of everything around it: 6.10% against the Nasdaq’s 0.59% loss is a gap of roughly 6.7 percentage points, several times any neighbour’s move.
This is not a high for the past year either. On our reading of Yahoo’s daily closes, Meta’s highest close over the prior 252 sessions is $777.59, from September 24, 2026, and at $762.50 the stock sits 1.94% below it and 15.5% above its December 31, 2025 close of $660.09. The chart for both sessions is on the TradingView Meta symbol page.
Meta’s own record will not settle it either way
We took every two-session gain of 6% or more in Meta’s full daily history, 3,621 sessions from May 18, 2012 to October 13, 2026. That turns up 132 qualifying days, 129 with a full 20 sessions of forward data, and those arrive in clusters, so we kept only events at least 10 trading sessions apart, leaving 71 de-clustered events measured close to close against a control of every day in the same history. Today’s 6.10% held to the close and qualifies, and it carries no forward return yet, so the distribution below is the 71 events that do. Seventy-one events across 14.4 years is about five a year, and 2026 holds four.
| Horizon | After the run (n=71) | Any day |
|---|---|---|
| 5 sessions | median -0.17%, mean +0.51%, 47.9% positive | median +0.54%, mean +0.59%, 55.8% positive |
| 10 sessions | median +0.45%, mean +0.27%, 50.7% positive | median +1.19%, mean +1.19%, 57.8% positive |
| 20 sessions | median +2.35%, mean +2.79%, 59.2% positive | median +2.42%, mean +2.38%, 61.4% positive |
On medians the conditional column trails at all three horizons, by 0.74 points at ten sessions, more than half the control median. On means that flips at 20 sessions, where +2.79% sits above the control’s +2.38%.
None of the gaps is statistically significant: the hit-rate differences run 1.34, 1.19 and 0.39 standard errors across five, ten and twenty sessions, and the mean differences give t-statistics of minus 0.08, minus 0.84 and plus 0.31. Read those as directional rather than exact: 20-session windows overlap even after de-clustering, so the observations are not independent.
The events are not spread evenly either: 2022 holds 10, 2020 holds 9 and 2012 holds 8, so 38% of the sample sits in three years. Drop those years, recut the control to the same surviving years, and 44 events remain. They still trail at five and ten sessions, medians of +0.04% against +0.65% and +0.59% against +1.43%. At 20 sessions they are ambiguous, a higher median of +4.89% against +2.82% with a hit rate of 63.6% against 64.2%.
This is a distribution, not a forecast, and it is wide: over 20 sessions the best outcome after one of these events was plus 26.30% and the worst minus 35.04%. The sample also runs bigger than today’s move: a median two-session gain of 7.66%, a mean of 8.89%, a largest of 31.50%, and only 4 of the 71 smaller than 6.10%. It counts any 6% two-day run whenever it occurred, post-earnings gaps included.
Over any 20 sessions in its history, whatever preceded them, Meta has finished higher about 61% of the time with a median gain of 2.42%. A 6% run in front does not measurably change those odds. The data cannot tell the two apart at any of the three horizons.
Three ways a Canadian can own it, and what each one costs
The first is listed in Canadian dollars. A CAD-hedged Meta CDR trades on the TSX under the ticker META, carrying a ratio of 0.0370958 against the Nasdaq-listed shares, per CIBC’s own CDR data. One receipt is about a twenty-seventh of a Meta share, which at 1.4222 Canadian dollars per US dollar puts it near $40 rather than $762. The ticker is identical on both exchanges, so an order has to specify which one. The TSX was closed Monday for Thanksgiving, so the receipt has had one session covering both US ones: $38.04 at the October 9 close, $40.33 at Tuesday’s close, up 6.02%. The Nasdaq-listed shares returned 6.10% over the same span, so the receipt tracked them to within a tenth of a percentage point.
It is not free. Our own measurement across 64 US CDRs with a complete two-year record put the all-in cost at 2.11% to 2.85% a year, most of it covered-interest carry rather than a CIBC fee. The same forward-selling cost sits inside anything that hedges US dollar exposure back to Canadian dollars. That is an annual rate, so it bears on a holding measured in months. Our guide to Canadian Depositary Receipts works through the mechanics.
Across the two sessions the US dollar fell 0.34% against the Canadian dollar, 1.4271 to 1.4222 on the Bank of Canada daily exchange rates, so an unhedged Canadian in the Nasdaq-listed shares made 5.73% in Canadian dollars rather than 6.10%, against the receipt’s 6.02%. The hedge was worth about 0.29 percentage points there. Four days of carry is a few hundredths of a point, lost inside tracking noise. One two-day currency move is not evidence the hedge pays: over months the carry compounds, while hedging rests on a view that the currency risk is not worth carrying.
Meta also comes bundled inside a Canadian-dollar S&P 500 index fund, a different decision from buying the single name on a two-day move. Our ranking of Canadian ETFs names the Canadian-listed S&P 500 funds.
The Nasdaq-listed shares swap the hedge cost for a conversion plus the broker’s US-dollar trade fee, set out in our guide to trading fees in Canada.
The next scheduled test is the October 21 earnings report
Meta reports Q3 2026 after the close on Wednesday October 21, 2026, call at 4:30 p.m. ET, eight days from today. That is company-confirmed from Meta’s own investor advisory dated October 12, 2026. Earnings land on the sixth trading session after October 13, inside the table’s ten and twenty-session windows but not its five-session one. For today’s event those two rows are substantially a bet on one earnings report, not on the run.
Meta runs a calendar fiscal year, so its Q3 ended September 30 and the launch cannot appear in those numbers. The announcement attached no price to the models and no cost, neither capex nor operating spend, so October 21 is the next scheduled occasion on which either could get a number. Management is likely to be asked about the launch there, and 14 years of its own trading will not tell you what it is worth.
Disclaimer: The content on bestcanadianstocks.ca is for informational and entertainment purposes only and does not constitute financial advice. Past performance is not indicative of future results. Always consult a qualified financial advisor before making investment decisions. Share prices, index levels, volumes and the Meta CDR price are from Yahoo Finance, read from daily bars and, for the announcement window and each session’s high, one-minute bars stamped with their start time. Figures for October 13 are closing prices. The USD/CAD rates are the Bank of Canada daily exchange rate. The historical study of past two-session runs, including the control and the era-matched comparison, is our own arithmetic on Yahoo’s daily closes. CDR terms are from CIBC’s own CDR data, and the hedge-cost measurement is our own, published in our guide to Canadian Depositary Receipts. Meta’s fiscal year end is from SEC submissions metadata, CIK 0001326801. The product announcement, its publication time and what it did and did not disclose are from Meta’s own newsroom, and the earnings date from Meta’s own investor advisory.
