TL;DR

xAI open-sources Grok LLMs; US takes 10% Intel stake; AI stocks hit 34% S&P 500.

Highlights

• xAI open-sources Grok 2 and 2.5 LLMs, with Grok 3 expected in six months; models require significant compute resources 1.
• Elon Musk projects xAI will deploy 50M Nvidia H100 GPUs in five years, identifying Chinese firms as main competition due to energy and hardware strengths 2.
• AI-related stocks now comprise 34% of S&P 500 market cap; Nvidia holds 92% of AI data center GPU share; Snowflake and Palo Alto upgraded 15.
• U.S. government converts $8.9B in CHIPS Act and Secure Enclave grants into a 10% equity stake in Intel , with potential for further warrants 4.
• Rocket Lab completes its 70th Electron launch and announces U.S. chip manufacturing expansion, supported by $23.9M in CHIPS Act funding 6.
• SpaceX prepares for its 10th Starship test flight, with a focus on reusability and Artemis program data 7.
• OpenAI and the UK sign a non-binding MoU to trial AI in public services; £2B ChatGPT Plus nationwide deal remains exploratory 5.
• Apple plans ultra-thin iPhone Air for 2025, foldable iPhone in 2026, and a 20th-anniversary model in 2027 12.
• Google confirms “Nano-Banana” generative image model for on-device editing in Pixel 10, advancing edge AI capabilities 8.
• Texas AG investigates Meta and Character AI over child-targeted mental health chatbots; regulatory scrutiny on AI safety intensifies 3.
• Microsoft AI chief warns of “AI psychosis” as bots mimic consciousness; calls for industry standards and guardrails 13.
• China tightens rare earth controls after a 53% export drop, but recent easing has alleviated electronics supply chain bottlenecks 14.
• Pump.fun regains Solana memecoin launchpad dominance, now holding up to 79% market share and $800M in cumulative revenue 9.
• White House launches National Design Studio led by Joe Gebbia to modernize 7,000+ government websites for improved UX and cost efficiency 10.
• Met Police to use live facial recognition at Notting Hill Carnival despite legal and safety concerns; UK remains an outlier in large-scale public deployment 11.

Commentary

AI infrastructure and open-source momentum continue to shape the competitive landscape. xAI’s release of Grok 2 and 2.5 LLM weights, along with Musk’s aggressive GPU deployment projections, highlight both the increasing accessibility of advanced models and the capital intensity required to compete at scale 12. The open-sourcing trend may lower barriers for startups and researchers, but the high compute needs still favor well-capitalized players and cloud providers. Meanwhile, AI equities remain a market driver, with Nvidia ’s entrenched position in AI GPUs and continued upgrades for software and security firms such as Snowflake and Palo Alto reflecting sustained investor appetite 15.

The U.S. government’s conversion of $8.9B in grants into a 10% equity stake in Intel marks a notable shift in how public funding intersects with private tech 4. This approach could set a precedent for future semiconductor or strategic tech investments, impacting cap tables and exit scenarios for late-stage or venture-backed hardware firms. Rocket Lab ’s CHIPS Act-backed expansion, alongside ongoing Starship testing by SpaceX , signals continued federal support for deeptech, space, and advanced manufacturing—areas where VC-backed companies may find increased non-dilutive funding and partnership opportunities 67.

On the consumer side, Apple ’s hardware roadmap and Google ’s move to on-device generative AI signal an ongoing push for differentiation through device innovation and edge AI 812. These developments may create new opportunities for startups in AI-powered consumer applications, edge computing, and supporting hardware. Pump.fun’s regained dominance in Solana memecoin launches and robust revenue also indicate that crypto infrastructure and community-driven platforms remain active, albeit volatile, areas for early-stage investment 9.

Regulatory scrutiny is rising, particularly around AI’s societal impact. Investigations into mental health chatbots involving Meta and warnings about “AI psychosis” from Microsoft underscore the need for compliance and risk management in sectors touching youth, healthcare, or public services 313. China’s rare earth policy remains a supply chain variable for hardware and electronics startups, though recent easing has alleviated some bottlenecks 14.

VCs should monitor open-source AI adoption, hardware supply chain shifts, regulatory developments in AI safety, and government-backed funding opportunities. These factors are likely to influence deal flow, diligence requirements, and exit strategies through the remainder of 2025.