In partnership with

Good morning. It’s Wednesday, January 22nd.

Did you know: On this day in 2000, the domain twitter.com was registered?

In today’s email:

• Project Stargate Promises $500B AI Investment

• Anthropic Set To Launch Voice Feature

• AI Drug Trials from Deepmind

• Grok 3 Tease

• 5 New AI Tools

• Latest AI Research Papers

You read. We listen. Let us know what you think by replying to this email.

US goes all-in on AI future with $100B from OpenAI and others, plans $500B total funding

OpenAI, Oracle, and SoftBank have set the stage for a $500 billion AI infrastructure overhaul with the Stargate Project. The partnership, revealed by President Donald Trump and tech leaders Sam Altman, Masayoshi Son, and Larry Ellison, includes plans to establish extensive data centers, enhance chip production, and expand energy resources.

Anchored by an initial $100 billion investment, it aims to support job creation and the development of a Texas-based 1-million-square-foot data center. SoftBank will manage funding, with operational oversight by OpenAI. Industry insiders frame the project as essential to outpacing China in the escalating AI arms race. While proponents highlight economic opportunities, critics, like Rob Enderle, caution about its broader socioeconomic ripple effects.

Alongside this announcement, Microsoft has lost its exclusive status as OpenAI's cloud provider. As part of the new agreement, Microsoft will retain the "right of first refusal" for any additional capacity requests, but OpenAI is now free to tap other cloud providers. While Microsoft continues to support OpenAI through its Azure platform, the collaboration has grown increasingly competitive, with OpenAI also integrating Oracle’s cloud infrastructure and other third-party services. Read more.

Anthropic Secures $1B from Google, Set to Launch Claude’s Two-Way Voice Feature

Anthropic plans to release a 'two-way' voice mode for Claude Anthropic plans to introduce a "two-way" voice mode for its chatbot, Claude, along with a memory feature that enables the model to recall previous interactions with users. CEO Dario Amodei, speaking at the World Economic Forum in Davos, noted the company’s growth driven by a surge in demand over the past year, though it faces challenges in meeting the required compute power.

In a new development that will strengthen its position in the competitive AI landscape, Anthropic has secured an additional $1 billion from Google, following its previous $2 billion investment. Read more.

DeepMind’s Hassabis Eyes AI-Designed Drug Trials by Year-End

Isomorphic Labs, the DeepMind spinoff under Google’s umbrella, sets its sights on launching AI-designed drug trials by the close of 2025, as outlined by CEO Demis Hassabis. The company is compressing the typical drug discovery timeline, transforming years of research into mere months. With strategic partnerships with pharma giants like Eli Lilly and Novartis, Isomorphic is positioning itself to exploit AI’s potential for rapid breakthroughs.

However, the pharmaceutical world is treading carefully, with early-stage trial data remaining inconsistent. Hassabis, while bullish on progress, noted that artificial general intelligence (AGI) is still a long way off—five years, perhaps, depending on pivotal advances. Founded in 2021, Isomorphic’s innovations are rooted in DeepMind’s AlphaFold, refining molecular structure predictions and offering promising solutions to complex biological puzzles. Read more. (Paywalled) Alternative Link.

Musk Teases Grok 3 Int4 Testing, Signals Imminent xAI Chatbot Launch

Elon Musk hinted at the forthcoming release of xAI's Grok 3 chatbot, announcing on X that he was "testing Grok 3 int4 inference." The next-generation chatbot, reportedly trained on the Colossus supercomputer powered by 100,000 Nvidia H100 GPUs, is expected to eclipse its predecessors in both computational efficiency and multimodal capabilities.

While prior iterations introduced image generation, Grok 3’s architecture is engineered to refine text synthesis, data interpretation, and multimodal reasoning. Read more.

• Perplexity Goes Head-to-Head with Google and OpenAI with the Market’s Lowest-Cost Search API, Sonar

• Meta's Hardware Ambitions Include Oakley Glasses and AR Tech to Rival Apple

• Tencent's Hunyuan3D 2.0 Pushes Boundaries in 3D Design Efficiency

• Moonshot AI unveils Kimi k1.5, China's next o1 competitor

• Foxconn is now using robots instead of people to make iPhones

• Deloitte: 74% of enterprises have already met or exceeded gen AI initiatives (but challenges remain) 

• Kling AI’s “Elements” to combine images into cohesive videos

• Trump’s Repeal on Biden’s EO Pits AI Innovation Against Oversight

• DeepSeek's latest R1-Zero model matches OpenAI's o1 in reasoning benchmarks

• Huawei seeks to grab market share in AI chips from Nvidia in China

• World's first chatbot, ELIZA, resurrected from 60-year-old computer code

• Abu Dhabi's $330-billion sovereign wealth fund says no one appreciates the level of disruption that AI is about to unleash

• UK to unveil 'Humphrey' assistant for civil servants with other AI plans to cut bureaucracy

• Gartner sees 10% IT spending jump in 2025, but don't get too excited

• New smartglasses use insect vision tech to restore full 3D visual experience

• KPMG: Artificial intelligence in the financial sector

• Next-gen AI device utilizes ion-controlled spin wave interference in magnetic materials

📄 VideoWorld: Exploring Knowledge Learning from Unlabeled Videos

📄 Multiple Choice Questions: Reasoning Makes Large Language Models (LLMs) More Self-Confident Even When They Are Wrong ​​

📄 GameFactory: Creating New Games with Generative Interactive Videos

📄 Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training

📄 MMVU: Measuring Expert-Level Multi-Discipline Video Understanding

Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X!