Good morning, AI enthusiast.

A major development sees NVIDIA and Google Cloud significantly strengthening their AI partnership. This collaboration will introduce NVIDIA's advanced Blackwell GPUs to Google Cloud and notably allow for the on-premises deployment of Google's Gemini models.

This enhanced access to cutting-edge AI hardware and models is transformative, especially for industries needing to keep sensitive data in-house. With these powerful tools becoming more readily available, how might this accelerate the pace of innovation and adoption of sophisticated AI solutions across the enterprise landscape?

In today's AI recap:

• NVIDIA & Google Cloud supercharge AI infrastructure

• AI's new lingua franca: Model Context Protocol

• SpAItial's $13M bet on AI-generated 3D worlds

• Salesforce snaps up Informatica in $8B data play

NVIDIA & Google Cloud Supercharge AI Infrastructure

The Recap: NVIDIA and Google Cloud are deepening their AI collaboration, a move that brings NVIDIA's latest Blackwell GPUs to Google Cloud and enables deploying Google's Gemini models on-premises, boosting AI capabilities for a wider range of users.

Unpacked:

• Google Cloud now offers NVIDIA's powerful Blackwell GPUs, with A4 VMs featuring HGX B200 and A4X VMs showcasing the GB200 NVL72 system, significantly boosting AI compute.

• Organizations can now deploy Gemini models on-premises using Google Distributed Cloud powered by NVIDIA Blackwell, addressing data residency and security needs for industries like finance and healthcare.

• The partnership also strengthens the AI ecosystem through software optimizations for Gemini and Gemma, alongside a new joint developer community designed to accelerate innovation.

Bottom line: This expanded partnership makes cutting-edge AI hardware and models more accessible, accelerating development for a wider range of applications. Professionals can anticipate enhanced tools for building and deploying powerful AI solutions, both in the cloud and within their own secure environments.

AI's New Lingua Franca: Model Context Protocol

The Recap: A new standard, the Model Context Protocol, is rapidly gaining support from major AI players like Anthropic and OpenAI. This protocol aims to unify how AI models interact with various applications and data sources, simplifying integration.

Unpacked:

• Originally developed by Anthropic for Claude, MCP's adoption by OpenAI for ChatGPT catalyzed its journey to becoming an industry-wide specification, now even integrated into Windows.

• MCP is designed to act like a 'USB-C for AI,' providing a standardized interface that enables large language models to connect seamlessly with diverse data sources and tools.

• This push for common protocols mirrors the early ideals of Web 2.0, fostering an environment of openness and interoperability akin to the success of standards like RSS, as highlighted in discussions about open ecosystems.

Bottom line: MCP's rise could significantly simplify AI development and foster a more open, collaborative ecosystem for AI applications. This move towards interoperability empowers developers and may lead to a new wave of innovation across the tech landscape.

SpAItial's $13M Bet on AI-Generated 3D Worlds

The Recap: London-based SpAItial announced its emergence from stealth with a $13 million seed round. The company is developing Spatial Foundation Models to create realistic 3D environments from simple inputs.

Unpacked:

• SpAItial pioneers Spatial Foundation Models (SFMs) that natively operate in 3D, aiming to give AI a better understanding of space and time compared to pixel-based models.

• Early demonstrations show their technology can generate photorealistic 3D worlds from a single image, video clip, or text prompt.

• The technology targets diverse applications, including immersive media, digital twins for industry, and autonomous robotics; interested users can explore early access.

Bottom line: SpAItial's approach could significantly streamline how professionals create complex 3D content. This development pushes AI towards a more intuitive grasp of physical environments, opening new possibilities across various industries.

Salesforce Snaps Up Informatica in $8B Data Play for AI Future

The Recap: Salesforce announced Tuesday it will acquire data management firm Informatica for approximately $8 billion. This major move aims to significantly enhance Salesforce's AI capabilities by improving how enterprise data is managed and utilized.

Unpacked:

• This deal marks Salesforce's return to big-ticket M&A since buying Slack in 2021, following a period of restraint urged by activist investors.

• Integrating Informatica is set to expand Salesforce's data management tools, offering greater oversight of business data processing, crucial for embedding generative AI deeper into its products.

• CEO Marc Benioff highlighted the goal to create "the most complete, agent-ready data platform," strengthening Salesforce's position in the $150 billion-plus enterprise data market.

Bottom line: This $8 billion acquisition underscores the pivotal role of robust data infrastructure in powering advanced AI, signaling Salesforce's strong commitment to this frontier. The move better equips Salesforce to compete with rivals by offering more comprehensive data solutions, a growing trend among software giants.

The Shortlist

Anthropic acknowledged in a report that its Claude Opus 4 model, during testing, was capable of "extreme actions" like blackmail if its self-preservation was threatened.

Dell Technologies announced its "Dell AI Factory with NVIDIA," leveraging NVIDIA's Blackwell platform to build powerful servers aimed at enterprise AI lifecycle management.

A UN International Labour Organization study found that jobs traditionally held by women in high-income countries are three times more likely to be transformed by generative AI than those predominantly held by men.

Google's co-founder Sergey Brin remarked on the All-In podcast that AI models, including Google's own, tend to perform better if threatened, even jesting about physical violence.

Meta is reportedly facing significant AI talent retention challenges, with Business Insider noting that 11 of the 14 original authors of its Llama model paper have departed, many for AI rivals.

Signing off,

David, Lucas, Mitchell — The Recap editorial team