Google DeepMind

Google DeepMind is Alphabet’s unified AI research laboratory, formed in 2023 from the merger of Google Brain and DeepMind under CEO Demis Hassabis. As part of Alphabet with a market cap over $2 trillion, it develops multimodal AI models including the Gemini family, the new Gemini Intelligence consumer surface (May 12, 2026), the Googlebook ambient-reading device (May 12, 2026), NotebookLM, Veo 3.1 video, and the Google Stitch design tool.[https://deepmind.google/]

What They Do

Google DeepMind focuses on advancing AI reasoning and multimodal capabilities through models trained on proprietary TPU infrastructure. The Gemini family powers integrations across Google products like Search, Gmail, and Android, leveraging massive distribution to billions of users.[https://blog.google/technology/ai/]

The lab emphasizes agentic AI for complex tasks in science, engineering, and research. Products like NotebookLM provide document-grounded analysis, while tools such as Google Stitch enable design prototyping from prompts.[https://deepmind.google/models/]

Compute scale and data access from Google’s ecosystem provide cost advantages, especially for low-latency inference in models like Gemini Flash variants.[https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/]

Current Flagship Products

• Gemini LLM family. Gemini 3.1 Pro (released April 16, 2026) leads on agentic reasoning and scored 77.1% on ARC-AGI-2, roughly double Gemini 3.0 Pro. Gemini 3.1 Flash-Lite anchors low-cost inference. Gemini 3 Deep Think remains the technical-problem mode for subscribers.[https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/][https://deepmind.google/models/gemini/]
• Gemini Intelligence (announced May 12, 2026): a new consumer assistant layer that fuses Gemini 3.1 Pro reasoning with Workspace, Search, Android, and the Googlebook device. Positions the Gemini brand as the cross-device intelligence layer, not just a chatbot.
• Googlebook (announced May 12, 2026): an ambient-reading hardware device tightly integrated with Gemini Intelligence for long-form reading, dictation, and citation workflows. First Alphabet hardware launch built end-to-end on Gemini.
• NotebookLM: AI research assistant with document-grounded responses and Audio Overviews; integrates Gemini 3.1 Pro.
• Google Stitch: AI design tool for UI mockups from text or images; powered by Gemini models.
• Veo 3.1 (released April 28, 2026): text-to-video model with longer clip duration, improved physics, and tighter cross-frame consistency than Veo 3; directly contests Runway Gen-5 expectations and Kling 3.0.
• AlphaFold 3: Protein structure prediction model expanding beyond proteins.
• Imagen 3: Text-to-image model via Vertex AI.

Recent News

• May 12, 2026: Announced Gemini Intelligence, a consumer-facing assistant layer that fuses Gemini 3.1 Pro reasoning across Search, Workspace, Android, and the new Googlebook reader. Also unveiled the Googlebook hardware device, Alphabet’s first end-to-end Gemini-native product.
• May 12, 2026: Isomorphic Labs (Alphabet drug-discovery spinout from DeepMind) closed a $2.1 billion Series C, sharpening the AI-for-science story alongside AlphaFold 3.
• April 28, 2026: Released Veo 3.1 with longer clips, stronger physics, and improved continuity, sharpening Google’s position in the text-to-video race against Runway Gen-5 and Kling 3.0.
• April 16, 2026: Released Gemini 3.1 Pro-2 (roughly double Gemini 3.0 Pro); rolled out to Gemini API, Vertex AI, and NotebookLM.[https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/]
• Early 2026: Launched the Gemini 3 series with Deep Think mode for subscribers and Gemma 4 open models derived from Gemini 3 research.[https://deepmind.google/models/][https://blog.mean.ceo/new-ai-model-releases-news-april-2026/]
• No new external funding reported for DeepMind itself; the lab remains Alphabet-funded. Isomorphic’s $2.1B Series C is separate.
• June 12, 2026: Re-verified against official sources; no new flagship model or hardware announcements since the May 12 Gemini Intelligence and Googlebook launches. Gemini 3.1 Pro, Veo 3.1, NotebookLM, and Google Stitch remain DeepMind’s current lineup.

Strategic Position

Google DeepMind leads in compute via TPUs and distribution across Google products, enabling low-cost scaling unmatched by startups. Gemini 3.1 Pro is competitive on reasoning benchmarks, and the May 12 Gemini Intelligence and Googlebook launches make Alphabet’s plan explicit: own the consumer “intelligence layer” across phone, browser, Workspace, and hardware. Veo 3.1 keeps Google in the cinematic-video race.

Threats include continued model velocity from Anthropic (Opus 4.8, Claude Mythos research preview) and OpenAI (GPT-5.5, Daybreak). Opportunities lie in enterprise via Vertex AI, Isomorphic Labs’ science track funded by the May 12 $2.1B Series C, and Gemini Intelligence converting Google’s distribution into recurring assistant usage.