Zoom Communications Inc. is betting its future on a "federated AI" strategy that blends third-party models like Anthropic's Claude with its own proprietary small language models. The approach, backed by an early $51 million stake in Anthropic, is now worth billions and helped drive the company's fastest enterprise growth in three years, according to its fiscal second-quarter earnings report.
CEO Eric Yuan told analysts that the multi-model architecture gives Zoom a distinct edge across core products including Zoom Contact Center and Zoom Virtual Agent. "A federated AI approach put us in a unique position," Yuan said, citing superior speech quality, lower latency, and enhanced customer trust as key differentiators.
How the federated model works
Rather than relying on a single AI provider, Zoom routes each task to the best model for the job. The company pairs Anthropic's large language models with its own smaller, specialized models, directing traffic between them based on cost and performance needs.
CFO Michelle Chang said this flexibility protects margins even as AI adoption accelerates. Non-GAAP gross margin reached 79.1% during the quarter. Chang attributed confidence in long-term margins to "our federated approach in AI," which allows Zoom to "be able to take the best model for the right time at the right cost and be able to fluidly direct traffic in between that."
The strategy is central to Zoom's evolution into what executives call an "AI-first system of action." For leaders tracking how AI investments translate into enterprise value, the model offers a concrete example of cost discipline applied to AI infrastructure - a key concern for operations-focused executives weighing similar bets.
Financial results and outlook
Second-quarter revenue reached $1.28 billion, beating estimates. Enterprise revenue rose 7.8% year-over-year to $787.5 million, and monthly active users of Zoom Workplace AI features jumped 125%. The company raised its full-year guidance to between $5.085 billion and $5.095 billion.
Analysts at BTIG and Wedbush view the Anthropic stake as a core asset protecting Zoom from competitive disruption while supporting platform monetization. The stake's value has grown substantially ahead of anticipated Anthropic IPO activity.
Shares fell 5.77% in premarket trading Wednesday after the company reported flat online revenue expectations, though the stock remains up 16.95% year-to-date. For executives evaluating AI strategy, the market reaction suggests investors are weighing near-term revenue growth against long-term positioning - a dynamic that applies well beyond Zoom.
The company's approach offers a working model for how enterprises can adopt Generative AI and LLM technologies without locking into a single vendor. By maintaining flexibility across models, Zoom keeps cost control and performance in its own hands rather than ceding them to a dominant AI provider.
Why this matters for executives and strategy
Zoom's federated approach demonstrates a practical alternative to the "pick one AI vendor" playbook. For executives, the key takeaway is architectural: building systems that can route work across multiple models - rather than committing to a single provider - preserves bargaining power and cost flexibility as AI pricing evolves. The company's margin performance during a period of heavy AI adoption suggests this discipline can pay off financially, not just technologically.
The strategy also signals a shift in how AI investments create value. Zoom's Anthropic stake, initially a modest strategic bet, has become a financial asset worth billions. That dual benefit - operational AI capability plus equity upside - is worth considering for AI for Executives & Strategy planning, particularly as major AI providers approach public markets and early investors see outsized returns.
Your membership also unlocks:
