The data indicates a $5.66 billion all-cash transaction. Salesforce acquires Braze at approximately 8.2x forward revenue against a FY2026 guidance midpoint of roughly $685 million. That multiple sits above the 2025 SaaS M&A median of 5-6x and well below the 2021 peak of 15-20x. The press release calls it an AI marketing play. The balance sheet calls it something less flattering: the purchase of a working architecture to replace one that does not work.
Context first. Braze was founded in 2011, listed in 2021, and built a cross-channel customer engagement platform spanning push, email, SMS, in-app messaging, WhatsApp, Content Cards, and webhooks. Salesforce Marketing Cloud is a stitched assembly of ExactTarget, acquired in 2013, plus Radian6 and Buddy Media. These are not two products in the same category. They are two different technology generations wearing the same category label. One runs a self-built realtime event pipeline feeding a low-latency decision engine. The other runs a legacy data model across multiple acquired codebases.
Here is the first finding, and it is the one the coverage missed. Code-level integration between Braze and Marketing Cloud is not feasible. The architectures share no common data model. What Salesforce can do is API-layer orchestration: keep both systems alive, bridge audience segments, and route execution through a common orchestration layer. That is a multi-year program, not a product launch. I have watched this pattern before. MuleSoft and Tableau both entered Salesforce in parallel-operations mode and stayed there. The integration did not converge. The maintenance burden doubled.
Salesforce's guidance implies Braze revenue decelerating from 26% growth in FY2025 to 15-16% in FY2026. Read that again. The acquirer is paying a premium multiple for an asset already in visible deceleration. Bulls will call this a strategic bargain. The strategic logic may hold. The valuation logic requires a growth reacceleration story that the numbers do not currently support. In the absence of data, opinion is just noise. The data here says deceleration.
Now the constraint almost entirely absent from the commentary: data processing agreements. Braze serves heavily regulated industries. Its customer DPAs, in standard form, do not permit using customer data to train vendor models. Salesforce's Agentforce pitch depends on an action layer fed by realtime behavioral data. Braze's event pipeline is exactly that action layer. But the moment the transaction closes, every enterprise customer's legal team reads the updated DPA and asks the same question: is our customer data now training Salesforce models? The answer must be negotiated customer by customer. The AI data flywheel cannot start on day one. It starts after a contract renegotiation cycle that, in my experience auditing post-merger data governance, runs twelve to eighteen months across a large enterprise base. The 'AI marketing synergy' narrative is a paper asset for year one.
Second technical finding. Braze's real moat is not the software. It is delivery operations. IP warming schedules, ISP relationship management, WhatsApp Business Solution Provider certification, country-by-country SMS sender ID registration. I have audited messaging infrastructure vendors, and this is the line item every financial model underestimates. Building that capability organically takes three to five years plus regulatory approval cycles in dozens of jurisdictions. Salesforce bought time. That is the honest description of this deal. Not synergy. Time.
The organizational risk deserves its own paragraph because it is structural, not cultural hand-waving. Braze operates on product-led growth: self-serve signup, sandbox environments, developer documentation as the primary acquisition channel, REST API and web SDK as first-class citizens. Salesforce operates on sales-led motion with AppExchange and MuleSoft as the integration center of gravity. These are not compatible operating systems. The historical Salesforce default is to converge acquired APIs onto its own standards. If Braze's REST surface gets flattened into Platform Events, the PLG engine stalls. The developers who adopted Braze because it was easy to integrate will notice. Latency in developer experience is a leading indicator, and it usually shows up in net revenue retention two quarters before management acknowledges it.
On the security and compliance side, the announcement-to-close window creates a friction period. Braze holds EU data residency infrastructure in Frankfurt. Merging legal entities triggers DPA update obligations across the enterprise base. Customer legal departments will queue questions during diligence. Deals slip. Renewals slip. This is not speculation; it is the standard choreography of regulated-industry M&A, and it is measurable in pipeline conversion rates during the quarter of close.
Let me be fair to the bulls, because they have one strong point. Salesforce's Marketing Cloud genuinely could not be rebuilt in place on a reasonable timeline. The ExactTarget architecture carries a data model that predates realtime expectations. Replatforming it organically is a three-to-five-year project with a high probability of failure, and Salesforce's own track record on marketing tech integration is its worst. Buying Braze converts an uncertain internal rebuild into a known integration cost. That is a defensible capital allocation decision. Paying 8.2x forward revenue for it is the debatable part, not the decision itself.
The second bull point: the OfferFit acquisition in 2024 gave Braze a bandit-based decision engine, which is precisely the execution surface Agentforce lacks. Agentforce can reason. It struggles to act across channels at millisecond latency. Braze can. If Salesforce grants Braze architectural independence, the combined action layer is genuinely differentiated. If it does not, the asset depreciates faster than the financial model assumes. That single governance decision, independent stack or absorbed stack, determines whether this deal works. Everything else is commentary.
There is a bug in the standard M&A narrative that I want to name explicitly. The narrative treats integration failure as the primary risk. It is not. The primary risk is integration taking too long. Dual product-line maintenance costs compound. Salesforce carries the Braze integration program plus the Marketing Cloud maintenance obligation simultaneously, and Marketing Cloud does not sunset. Its installed base generates cashflow that funds the transition, which means nobody has the incentive to kill it. Social Studio's chaotic 2024 shutdown is the precedent. Legacy lines enter maintenance mode and stay there for years, dragging gross margin.
My audit experience on post-merger data governance says the first-year synergy number in these deals is routinely overstated by a factor of two or more. The reason is always the same: DPA renegotiation and architecture convergence take longer than the deal model assumes. I would want to see, before believing the synergy case, the actual contract language on data usage rights, the engineering org chart post-close, and whether Braze's API surface survives the platform standardization review. Those three documents tell you more than any earnings call.
So what is the forward-looking judgment? Watch one signal above all others: whether Braze engineering reports into Salesforce platform leadership or retains an independent stack with its own release cadence. If the stack stays independent, this is a $5.66 billion purchase of time that was rationally priced. If the stack gets absorbed, the depreciation clock starts on day one, and the 8.2x multiple becomes a rounding error in a larger write-down conversation eighteen months out. The transaction closes. The architecture decision does not have a deadline. That is the problem. Silence in the ledger is loud.