TBLA · Taboola.com Ltd. · Stock research

Current Thesis

The leg this name was being bought for — open-web adtech repricing as a toll booth on AI answer surfaces — met its number on 2026-08-05 and missed it. Q2 revenue was $476.8M, +2.4% YoY, against a $499.4M consensus. FY26 revenue guidance was cut to $1.930–1.956B from $2.006–2.062B, and Q3 was guided to $460–473M against a $518.1M consensus. Shares went from a 2026-08-04 close of $5.29 to a $3.71 premarket quote the next morning, and closed 2026-08-14 at $4.05, 27.4% under the $5.58 52-week high.

What the quarter did deliver was profit mix: ex-TAC gross profit $192.4M (+11.8% YoY), adjusted EBITDA $55.5M (+22.8%), net income $4.3M against a $4.3M loss a year earlier, and an FY26 adjusted EBITDA range raised to $228–240M from the prior $222–240M. Management attributed the revenue shortfall to a deliberate cleanup of low-quality publisher supply and the removal of Explore More after a Google policy change. That is a coherent account of shedding gross-profit-poor revenue. It is also a different story from the one that carried the stock through June and July, and the tape has not repriced onto it yet.

The 2026-08-05 release contained no disclosed DeeperDive revenue contribution and no third-party customer count — the exact secondary condition this coverage set ahead of the print.

Bullish and bearish views on Taboola.com Ltd.

The model's bull view on Taboola.com Ltd. (TBLA), in brief: Gross profit grew roughly five times faster than revenue. The bear view: The Q3 guide is not a trim. $460–473M against a $518.1M consensus is a gap that forces FY27 models to be rebuilt, not adjusted. Estimate cuts of that width typically arrive in more than one wave. The FY26 revenue midpoint dropped from roughly $2.03B to roughly $1.94B on the… Both cases follow in full.

Bull Case

• Gross profit grew roughly five times faster than revenue. Q2 ex-TAC gross profit $192.4M (+11.8% YoY) against revenue $476.8M (+2.4%). If the cleanup narrative holds, the deleted revenue was carrying almost no margin.
• Guidance moved in opposite directions on purpose. FY26 revenue cut to $1.930–1.956B while adjusted EBITDA was raised to $228–240M and ex-TAC gross profit set at $772–783M. Q3 adjusted EBITDA is guided $51.5–56.5M on $460–473M of revenue — a margin structure the 2025 business did not have.
• Balance sheet moved the right way. Cash and equivalents $133.1M at 2026-06-30; revolver balance $72.0M against $102.3M at year-end 2025. Net income turned positive at $4.3M versus a $(4.3)M loss in the year-ago quarter.
• Sell-side cut targets without cutting ratings. Benchmark maintained Buy and moved its target to $5.50 on 2026-08-06; TD Cowen maintained Buy and cut from $6.00 to $5.00 on 2026-08-10. Both sit above the 2026-08-14 close of $4.05.
• Distribution keeps adding logos. Fox News was named in the 2026-08-05 release alongside Realize momentum; the DeeperDive publisher roster built through the prior leg (USA TODAY, Ynet, Ouest-France, India Today, plus Wikitree from 2026-07-01) is intact.

Bear Case

• The Q3 guide is not a trim. $460–473M against a $518.1M consensus is a gap that forces FY27 models to be rebuilt, not adjusted. Estimate cuts of that width typically arrive in more than one wave.
• The FY26 revenue midpoint dropped from roughly $2.03B to roughly $1.94B on the 2026-08-05 release as reported by wire coverage of the print — an entire quarter's worth of prior optimism removed in one session.
• DeeperDive is still an unpriced line. No revenue contribution, no third-party AI customer count disclosed as of 2026-08-05, four-plus months after the 2026-06-16 ad-network opening that drove the June re-rating. A multiple was expanding on a number that has never been printed.
• Free cash flow halved. $17.3M in Q2 2026 against $34.2M in the year-ago quarter. The repurchase programme that removed roughly 19% of float on a 2025-YTD cumulative basis is funded from that line, and the company also used cash to pay down the revolver.
• Google can delete a product. Explore More was removed after a policy change, and management flagged the loss came faster than expected. That dependency is structural and applies to any future Taboola surface that sits downstream of Google distribution.
• Litigation-solicitation headlines followed the gap. Holzer & Holzer announced an investigation on 2026-08-05. Announcements of this type are routine after a >25% single-day decline and are not themselves a filed claim, but they generate recurring headline flow.

Setup & Price Structure

Reference close 2026-08-14: $4.05. 52-week high $5.58, so price sits 27.4% below it. Three-month return −18.8%. RSI(14) 34.1 — pressed, not yet washed out.

The structure that defined the June–July leg is gone. It was built on the 2026-06-16 DeeperDive ad-network announcement and extended on the 2026-07-01 Wikitree win, which moved the stock 4.7% on a single small logo. The 2026-08-05 gap erased it in one session: from $5.29 the prior close to $3.71 quoted premarket, with wire coverage putting the post-print decline in the 26–30% band depending on the measurement point. The $4.80 weekly-close level that this coverage previously named as the leg's floor was lost with no defence. Overhead supply now runs from roughly $4.80 up through the pre-print $5.29.

The narrative is dead for the AI-answer-monetization re-rating leg, dated to 2026-08-05. The narrative was accelerating between 2026-06-16 and 2026-07-17; the print removed both the revenue trajectory it was extrapolated from and the disclosure that would have quantified it. What exists at $4.05 is a fresh, unformed base on a different argument (ex-TAC gross profit and EBITDA compounding on a shrinking top line), and that argument has one quarter of evidence behind it.

Positioning and crowding observables, stated without a verdict: two brokers cut targets inside five sessions of the print while keeping Buy ratings, which leaves published targets 23–36% above the last close and means the sell-side estimate reset is visibly incomplete. There is no earnings date inside 30 days to force resolution. The 2026-06-26 Russell 2000/3000 inclusion flow was executed and is spent. Free cash flow of $17.3M in Q2 against $34.2M a year earlier reduces the mechanical repurchase bid that supported the shares through the first half.

Catalyst Calendar (next 30 days)

• No publicly dated company event falls inside 2026-08-15 → 2026-09-14. The gap between the 2026-08-05 print and the next scheduled disclosure is the defining feature of the current window.
• ~2026-11-04 (est.) — Q3 2026 results. Q1 printed 2026-05-07 and Q2 printed 2026-08-05; the estimate follows that cadence and is unconfirmed. This is where the $460–473M revenue guide, the $184–190M ex-TAC gross profit guide and the $51.5–56.5M adjusted EBITDA guide get marked.
• Ongoing, undated — any 6-K/8-K disclosing a DeeperDive commercial metric, or any further Google policy action affecting Taboola-distributed products. Neither is scheduled.

What Would Change Our Mind

The datapoint that would flip the read constructive is a printed DeeperDive number: a disclosed revenue contribution or third-party AI customer count at the ~2026-11-04 Q3 release. Absent that, the AI-monetization premium has nothing to attach to, and the name is a low-single-digit-growth ad network trading on a margin-mix argument with one quarter of history.

On the tape, a weekly close below $3.71 — the post-print quote low from 2026-08-05 — would show the reset is not finding a floor and would open fresh 52-week-low territory. On the other side, reclaiming $4.80 on a weekly close would put price back inside the June–July shelf and force a re-examination of the dead label.

The fundamental condition that would break the surviving bull argument: ex-TAC gross profit growth decelerating toward revenue growth once the publisher cleanup laps, which would identify the +11.8% as a subtraction artifact rather than a mix improvement. A further FY26 revenue cut at the Q3 print would say the cleanup is larger than management scoped on 2026-08-05.

Correlation Notes

• Trades with the open-web adtech complex — Criteo, Outbrain/Teads, Perion, Magnite — which shares the same exposure to platform policy changes and to AI answer surfaces absorbing referral traffic that publishers monetize.
• Direct dependency on Google product policy: the Explore More removal was disclosed as a policy-driven revenue loss on the 2026-08-05 call, so Google Search/AI Overviews product announcements are a live input to this name specifically.
• Small-cap beta: Russell 2000/3000 inclusion executed at the 2026-06-26 close, so index-flow support is behind it and the name now moves with small-cap risk appetite rather than with mechanical demand.
• Performance-advertising demand cycle: revenue is CPC-billed against roughly 20,000 performance advertisers, tying it to direct-response ad budgets rather than brand budgets.