Free cash flow tripled to $437M. Stock-based comp ate every dollar of it twice over.

Adjusted EBITDA tripled in two years, gross margin climbed to 59% in Q4 2025, and Q1 2026 dropped 75 cents of every incremental revenue dollar to EBITDA. Strip out the $1.02B Snap pays employees in equity each year and the picture inverts — owner free cash flow has been negative every year of Snap's public life. The next cycle has to take stock-based comp dollars below $1B and hold gross margin above 60% for both lines to point the same way.

Four years and $2.25B of repurchases failed to retire a single net share — period-end shares still rose ~6% across the program.

• The arithmetic. Across the four-year buyback program (FY2022–FY2025), $2.25B was spent on repurchases yet period-end shares still rose from 1,619M (end-2021, before the program began) to 1,712M (end-2025) — a 5.7% increase. Stretching to the six-year window (end-2019 to end-2025), period-end shares are up 21%, from 1,415M to 1,712M. Every buyback dollar offset stock-based-comp dilution; not a single share was net retired.
• The denominator is guided up. Management telegraphed FY2026 stock-based comp at $1.2B versus $1.02B actual in FY2025 — an 18% increase in the absolute dollars the buyback is asked to absorb, before any RSU mark-to-market on a stock-price recovery.
• The number that decides this. The FY2026 10-K cover-page share count, due Q1 2027, prints either below 1,690M — the first true contraction in Snap's public life — or above 1,712M, in which case the dilution offset continues for a fifth year and consensus EPS upgrades do not reach the owner.

In eight weeks an activist forced the first visible CEO response since the 2017 IPO.

Before: Evan Spiegel and Robert Murphy hold 99.5% of voting power through Class C super-voting stock; Class A public shareholders have zero votes. For nine years no outside investor had leverage on capital allocation — $2.25B of buybacks failed to shrink the share base, Spectacles consumed three billion of cumulative spend without scaling, and the $400M Perplexity partnership announced in November 2025 evaporated by Q1 2026.

Pivot: On March 31, 2026, Irenic Capital opened a public campaign arguing actionable steps could unlock roughly seven times the share price. Within 15 days Spiegel announced a 16% workforce cut and $500M of annualized cost takeout, attributing the move to AI productivity — 65% of new code, he said, is now AI-generated. Reuters tied the timing directly to Irenic pressure.

Today: The cost takeout bridges mechanically to GAAP profitability if Q1 2026's 75% EBITDA flow-through holds. The open question is whether outside pressure becomes a permanent feature of the governance setup — or whether the founders revert to the 2017–2024 pattern of buybacks that don't shrink shares once Irenic moves on.

Stories capped DAU growth in 2016. Reels capped engagement in 2020. Instants now targets the messaging core.

• The pattern. Meta launched 'Instants' on May 14, 2026 — a direct copy of Snapchat's ephemeral surface and the third Meta-clones-Snap cycle. The prior two compressed Snap's monetization ceiling rather than its user count; the stock fell on the day of the launch.
• Where the money actually lives. North America is 21% of daily users but the majority of revenue, with ARPU five-to-seven times Rest-of-World. NA daily users have printed −1% YoY at 99M for two consecutive quarters; each 1M permanent NA user loss removes roughly $30–40M of annual revenue at NA pricing.
• The bull's counter. Snap said in Q3 2025 it would deliberately let DAU decline in low-ARPU geographies; in Q4 2025 NA ARPU rose 5% YoY while DAU was flat, and the quarter was the first GAAP-positive operating print in four years. If that pivot holds, the Instants narrative loses force — but the proof requires the Q2 and Q3 prints to show NA users stable AND NA ARPU still climbing.

Snapchat+ at 24M subscribers and +71% YoY is buried in a 1.8× sales multiple — and consensus has not given it a separate number.

• The scale. Snapchat+, Lens+, and Platinum reached 24M paying subscribers by Q4 2025 — up 71% YoY — on roughly $700M of annualized revenue at 70–80% gross margin. No social peer has matched that paid-subscriber scale; Meta Verified and X Premium are a fraction of the size relative to their user bases.
• The mispricing. Other Revenue grew 87% YoY to $285M in Q1 2026 — embedded inside Snap's blended 1.8× EV/Sales ad-platform multiple. At a software peer multiple of 8–12× sales, the subscription line alone would be worth $5–10B of standalone equity value — half to all of the entire $9.7B market cap.
• The trigger. Snap has not yet disclosed subscription gross margin separately. A single line-item disclosure or a Q4 2026 milestone of 35M subscribers at $1.2B+ ARR would give the sell-side cover to model a sum-of-parts. Conversion is only 5% of daily users, so a competing $2–3 bundle from Meta or X is the obvious thing that could break the pricing arbitrage.