GTM Stack Movements: September–October 2026
GTM Stack Movements: September–October 2026
Seven weeks, one cheque worth writing about, and then — in the space of forty-eight hours — three platforms announcing the same product. Between September 1 and October 19 the venture market stayed quiet and the incumbents got loud, and what they all announced was a version of the same thing: an agent builder grounded in their own data. The most revealing detail is what they charged for it.
The round that reset the ceiling
Clay raised $115M on September 9 at a $7.1B valuation, led by Wellington Management with Sequoia, StepStone, a16z Perennial, CapitalG and DST participating. That is more than double the $3.1B its August 2025 Series C carried, and it makes Clay worth roughly twenty-two times the capital it has raised — about $319M by our arithmetic, since the company published neither a lifetime total nor whether $7.1B is a pre- or post-money mark. Clay says more than 17,000 teams build on it (company-stated). Full record on the Clay profile.
The investor matters as much as the number. Wellington is crossover money — the kind that prices a company against where it could list, not against the last private round. For a company that sells to GTM engineers rather than CROs, being valued like late-stage infrastructure is the strongest signal yet that the GTM engineering motion is now a category rather than a job title. It was also the only AI-GTM round in the window that cleared our verification bar.
Three platforms, two days, one announcement
On September 30, Gong used its Celebrate ’26 event to announce Mission Callisto, whose centrepiece is Gong Enrich — third-party account and contact data piped into the Gong Revenue Graph — alongside a Deep Mode in Gong Assistant and an upgraded Agent Builder. Gong’s own help centre still lists Enrich as rolling out through September, so treat availability as staged rather than shipped.
The same day, Apollo used its ApolloNEXT keynote to unveil an Intelligence Layer, Builder Studio and Messaging OS. Read the release carefully: only the Intelligence Layer is available now. Builder Studio and Messaging OS are “coming soon,” with no dates and no pricing. Several write-ups have described all three as shipped; they are not, and the difference matters if you are building a 2027 plan around them.
Then on October 1, ZoomInfo launched Agent Teams inside GTM Studio — multi-agent plays that run across a company’s own systems and signals. The notable part is commercial, not technical: it is available now with no new SKU and no new contract. A vendor midway through moving its customers from seats to consumption credits just gave its flagship agent capability away as an entitlement. That is either confidence that usage will follow, or an admission that agent features are now table stakes rather than an upsell. We will find out which when Q3 earnings land in early November.
One more thing worth noticing in Gong Enrich’s launch partner list: Apollo and ZoomInfo are both on it, next to Findymail, Firmable, Kernel, LeadIQ, Lusha, RocketReach, Wiza and ZeroBounce. The same two companies that announced competing agent platforms within twenty-four hours are suppliers inside Gong’s. The data layer is consolidating and commoditising at the same time, and vendors have stopped pretending those are different trends.
One day from acquisition to product
ZoomInfo announced the acquisition of DoubleO.ai, a multi-agent orchestration platform, on September 30 — terms undisclosed — and shipped Agent Teams on its architecture the next day. A one-day gap between announcement and productisation tells you the integration work was already done, which in turn tells you this was a technology-and-team purchase rather than a revenue one. It is the same shape as the five signal-layer tuck-ins we tracked in Q3, and it is now on the funding tracker.
In the same stretch, ZoomInfo folded Workbounce — acquired back in November 2024 — into its GTM Context Graph, adding internal unstructured material like call transcripts and decks to the external verified data. Between DoubleO.ai and Workbounce, the strategy we described in The Context-Layer War has become explicit: own the market data, own the customer’s own knowledge, and rent both to whatever agent the customer already uses.
The consolidation kept closing
SoundHound AI completed its acquisition of LivePerson on September 4 — announced April 21, closed for roughly $43M in equity value, which implies about $250M of enterprise value once LivePerson’s discounted debt is accounted for. A once-$4B conversational-commerce company exited for less than the price of a Series B. If you want a single data point on what happened to pre-LLM conversational AI, that is it.
Further out but closing this quarter: Publicis agreed in May to buy LiveRamp for about $2.5B ($38.50 per share, expected to complete by the end of 2026). The identity layer underneath a great deal of B2B targeting is becoming part of an advertising holding company. Nobody in GTM asked for that, and everyone in GTM will live with it.
Quick hits
Clari + Salesloft announced a partnership with allGood on August 5, giving its legacy email-marketing customers a migration path to an AI-native tool — notable because the merged entity chose to rent that capability rather than build it, a useful input for our December twelve-month re-score of the merger. Relay.app‘s paid accounts ended September 14, completing the category’s only genuine shutdown rather than tuck-in. Stuut raised a $52.5M Series B on October 7 (Insight Partners, with a16z and M12) for order-to-cash agents — adjacent to GTM rather than in it, but the clearest sign yet that the agent thesis is moving downstream of the close. And Claudeforce, the Salesforce–Anthropic partnership announced August 26, was slated to reach open beta in September; we have found no confirmation that it did. If you were counting on it for a Q4 pilot, verify before you plan.
The scoreboard
Q3 2026 closed with $275M across five venture rounds — Aligned $60M (July 1), Alta $25M (July 8), Sable $45M (July 16), Encore AI $30M (July 29), Clay $115M (September 9) — and seven recorded acquisitions: Common Room, Ciro, Inconvo, Seam AI, Agency, Highspot and DoubleO.ai. Live view on the funding tracker. Both figures are higher than the ones in our August edition, because the Seismic–Highspot close and the two September events have since been recorded. Q4 opens with nothing on the board nineteen days in. One company absorbed the quarter’s venture capital and three absorbed its attention, and on current evidence the interesting question for Q4 is not who raises next — it is whether anyone can charge for an agent.