Wall Street analysts from Citigroup predict that "Agentic Workflows" — AI-driven, usage-based cloud data platforms — will dominate enterprise software and cloud computing spending through 2026. Hyperscalers and "AI + cloud data platform" software firms like Oracle, Snowflake, MongoDB, and Elastic are poised to be primary beneficiaries.

Agentic Workflows enable AI agents to autonomously execute multi-step tasks (e.g., data extraction, analysis, and decision-making). Citi notes a global shift from "AI experimentation" to "early-stage规模化 deployment," projected to drive 20-35% growth in enterprise software and cloud spending by 2026, building on 2025’s 50% surge.

Key observations: - Snowflake’s AI modules (Intelligence/Cortex) adoption is accelerating, with单个AI use cases reportedly generating ~$200K incremental consumption. Large clients may deploy 10+ use cases, significantly boosting usage. - A major U.S. airline is testing Elastic’s AI cloud monitoring to replace Splunk, citing strong demand for Elastic’s "AI + cloud security" capabilities.

Recent earnings from cloud leaders (Alphabet, Meta, Salesforce) and AI specialists (AppLovin, Palantir) underscore robust demand for both AI infrastructure (e.g., NVIDIA GPUs) and enterprise-grade AI applications. Two trends dominate: 1. **Generative AI apps** (e.g., ChatGPT, Claude) expanding beyond chat interfaces. 2. **AI agents** evolving into productivity tools, poised to define the 2030s AI landscape.

**Google’s Gemini 3 Fuels Momentum** - Gemini 3’s launch triggered massive AI token demand, forcing Google to limit free access to Pro/Nano Banana Pro tiers. - OpenAI’s rumored GPT-5.2, reportedly outperforming Gemini 3 in benchmarks, signals intensifying competition.

**MongoDB’s Strategic Edge** - Deep integration with Google Cloud/Vertex AI positions MongoDB as a core beneficiary of推理算力 demand. Atlas + Vertex AI stacks are scaling rapidly, driven by enterprise/consumer demand for generative AI, search, and agentic workflows. - MongoDB’s recent earnings beat expectations, reflecting its pivotal role in Google’s AI ecosystem.

**Investment Shift: From Compute to Workflow Monetization** Post-2024’s "AI infrastructure phase," focus pivots to platforms monetizing usage (e.g., cloud data/observability tools). Citi advocates: - **Overweight**: Usage-based cloud data/security platforms (e.g., MongoDB, Snowflake). - **Caution**: Closed, expensive AI platforms (e.g., Palantir), prioritizing tangible metrics over hype.

**Outlook** With Gemini’s 650M+ monthly users and 20x annual token growth, Google’s ecosystem — and partners like MongoDB — exemplifies the "rising tide" of AI-driven cloud expansion.