🔥36氪•Stalecollected in 2m
Gemini hits 900M users; Nvidia posts record profits
#ai-scaling#financial-results#agent-bigemini,-nvidia-h100/blackwell,-smartbigooglegemininvidiasmartbi
💡Get the latest on AI scaling metrics from Google and Nvidia's record-breaking financial growth.
⚡ 30-Second TL;DR
What Changed
Gemini monthly active users reached 900 million, with daily requests increasing 7x.
Why It Matters
The massive scale of Gemini usage and Nvidia's financial performance underscore the rapid industrialization of generative AI. Enterprise-focused Agent BI tools are becoming the next frontier for practical AI adoption.
What To Do Next
Evaluate the integration of Agent BI architectures into your internal data pipelines to automate decision-making workflows.
Who should care:Enterprise & Security Teams
Key Points
- •Gemini monthly active users reached 900 million, with daily requests increasing 7x.
- •Nvidia Q1 net profit hit $58.3 billion, a 211% year-over-year increase.
- •SmartBI released 'Baize' V5, an Agent BI platform with 99.9% accuracy in core scenarios.
- •Google processes over 3.2 gigatoken/month, a 7x increase from last year.
🧠 Deep Insight
Web-grounded analysis with 21 cited sources.
🔑 Enhanced Key Takeaways
- •Gemini's user growth is significantly driven by its deep integration across Google Workspace applications (Gmail, Docs, Sheets, Meet, Chat) and the introduction of "Personal Intelligence," which allows it to connect with user data from Gmail, Photos, Drive, and Search for highly personalized responses.
- •Nvidia's record profits are predominantly fueled by its data center business, which achieved a 92% year-over-year growth to $75.2 billion in Q1 2026, prompting the company to strategically reclassify its revenue reporting into dedicated "Data Center" and "Edge Computing" segments.
- •Google continues to advance its AI models with the release of Gemini 3.5 Flash, which is four times faster than other frontier models and now powers the Gemini app and Google Search's AI Mode, alongside the introduction of Gemini Omni, a new-generation AI multimedia generator capable of creating realistic video content from multimodal inputs.
🛠️ Technical Deep Dive
- Gemini models are a family of multimodal large language models (LLMs) capable of processing interleaved sequences of text, images, audio, and video as inputs.
- Gemini 1.5 Pro and Gemini 3 Pro utilize a sparse Mixture-of-Experts (MoE) transformer-based architecture.
- Gemini 1.5 Pro and Gemini 3 Pro feature a large context window of up to 1 million tokens, enabling them to process extensive datasets.
- Gemini models are trained using Google's Tensor Processing Units (TPUs), with Gemini 2.0 specifically utilizing sixth-generation TPUs.
- Gemini 3 Pro includes an optional "Deep Think" mode designed to enhance performance for complex problem-solving tasks during inference.
- Gemini Omni, powered by the Nanobanana 3.0 AI model, is a new-generation AI multimedia generator that can produce realistic video content from combined text, audio, image, and video inputs, demonstrating an intuitive understanding of physical forces like gravity and kinetic energy.
- Nvidia's latest offerings include the Blackwell architecture, with strong demand for GB300 and NVL72 systems, and the newly announced NVIDIA Vera Rubin platform, which features the NVIDIA Vera CPU purpose-built for agentic AI and BlueField-4 STX for accelerated storage infrastructure.
🔮 Future ImplicationsAI analysis grounded in cited sources
Gemini's deep integration will lead to a pervasive "AI autopilot" experience across Google's ecosystem.
Its ability to connect with Gmail, Photos, Drive, and Workspace apps, combined with agentic capabilities like Gemini Spark, suggests an inevitable future where AI autonomously manages and streamlines daily digital tasks.
Nvidia will solidify its position as the foundational infrastructure provider for the global AI buildout, extending beyond GPUs to full AI factory solutions.
The company's record data center revenue, strategic shift to "Data Center" and "Edge Computing" reporting, and the introduction of platforms like Vera Rubin (including a purpose-built AI CPU and accelerated storage) indicate a move towards comprehensive AI infrastructure offerings.
Enterprise data analysis will increasingly rely on highly accurate, agentic BI tools, reducing manual intervention and accelerating decision-making.
SmartBI's 'Baize' V5, with its claimed 99.9% accuracy in core scenarios, exemplifies a market trend towards AI-powered business intelligence that automates data analysis and provides actionable insights with high reliability.
⏳ Timeline
2023-05-10
Google announced Gemini at Google I/O, positioned as a multimodal successor to PaLM 2.
2023-12-06
Google officially launched the Gemini 1.0 model family (Ultra, Pro, Nano).
2024-02-08
Google rebranded its Bard chatbot as Gemini.
2024-05
Gemini 1.5 Flash, a faster and more cost-efficient model, was announced at Google I/O.
2025-11-18
Google released Gemini 3 Pro and 3 Deep Think, marking a strategic shift in model release patterns.
2026-05-19
Gemini 3.5 Flash was released, becoming the default model for the Gemini app and Google Search's AI Mode.
📎 Sources (21)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
📰
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 36氪 ↗