Hey {{first_name|TEDchats Fam}}!

Happy Thursday! The White House is no longer waiting for AI regulations, it may already be deciding when the most powerful models can launch.

Explore all the AI resources you’ll ever need

Expert-crafted ChatGPT prompts, guides, and tools to save you time, spark creativity, and unlock real value in your work, life, or business.
→ See What’s Inside

📣 TOP AI NEWS

The Trump administration directly asked OpenAI to stagger its upcoming model release rather than launching publicly, prompting the company to roll out a restricted preview to select partners instead. This marks a sharp departure from previous US AI governance, which relied on voluntary commitments and broad executive orders rather than targeted White House intervention in specific product timelines. The pressure came roughly two weeks after rival Anthropic faced similar treatment, suggesting a pattern of hands-on administration engagement with frontier AI releases that left competitors guessing where capability thresholds trigger government response. OpenAI's compliance, while avoiding direct confrontation, creates tension with its typical rapid-launch culture and arrives alongside separate reports of its IPO delay, compounding a week of strategic constraint. The approach fills a legislative vacuum, Congress having stalled on comprehensive AI regulation, but the ad hoc nature raises questions about consistency and due process that formal rulemaking would address. For Google, Microsoft, Meta and xAI, the unclear standards create planning hazards as they scale toward frontier capabilities without knowing what triggers intervention (read more).

Meta's $60 billion infrastructure binge finally has a revenue thesis. The company is building a cloud business to sell excess AI compute and Llama model access, following SpaceX's xAI playbook. The stock jumped on the news as investors, long skeptical of Meta's AI spending, glimpsed a path to direct returns beyond advertising. AWS, Google Cloud, and Microsoft Azure control over 60% of enterprise cloud, but Meta's infrastructure was built for AI training clusters, not generic hosting, giving it potentially unique economics for AI-native workloads. The catch: enterprise cloud requires billing systems, compliance certifications, and sales teams that look nothing like Meta's consumer app DNA. The initial target is likely AI startups and researchers who'll tolerate rough edges for cutting-edge access (read more).

SoftBank Group shares cratered 12-13% on June 26, their steepest drop since August 2024, after reports that OpenAI is leaning toward delaying its IPO until 2027. The selloff erased most gains from the previous session's record Nikkei close, with the broader Japanese market falling 4.15% as investors absorbed the implications. OpenAI advisors reportedly presented management with a choice between a quick listing at lower valuation or patience until 2027 for better market conditions, and the company chose the latter, a rational move for a firm valued above $850 billion but one that squeezes early backers like SoftBank who had baked near-term liquidity into their investment theses. The whipsaw was particularly painful given that SoftBank shares had surged over 14% just weeks earlier on IPO anticipation, making this week's collapse a stark illustration of how concentrated thematic bets can vaporize. Masayoshi Son's AI-centric portfolio strategy now faces a two-year wait before its cornerstone holding faces public market validation (read more).

🛠️ THIS WEEK’S AI COMBO: 1 PROMPT + 1 TOOL

⚡️ MIND SPARK

🔬 Claude Science Goes Local: Anthropic turned Claude into a local “science workbench” for macOS and Linux, letting pharma teams run AI research without sending sensitive data to the cloud. Beyond tooling, it’s also seeding adoption with 50 research grants at $30K in credits each, effectively buying early biotech use cases while embedding Claude into workflows via Benchling (see MIT review).

🍌 Google’s Ultra-Cheap Image Engine Targets Scale Over Perfection: DeepMind’s Nano Banana 2 Lite pushes image generation to ~4 seconds and fractions of a cent per output, dramatically undercutting rivals like Midjourney and Stability AI. The trade-off is quality, positioning it for fast ideation rather than final assets, while integrations into Adobe Firefly and Gemini tools tighten Google’s grip on the creative pipeline.

📱 Google Finance Finally Gets an App After 20 Years: Google is bringing Finance to mobile with a redesigned Android app featuring AI “Key Moments,” Deep Search, and earnings call summaries layered over live market data. The rollout signals a shift from static charts to AI-driven interpretation of financial news, with iOS still queued for later this year (see Google blog).

📉 SpaceX IPO Mix-Up Leaves Korean Broker Empty-Handed: South Korea’s Mirae Asset Securities ended up with zero SpaceX IPO allocations after a breakdown with lead underwriters, despite having taken investor deposits and promoted access. Retail investors were forced into the open market instead, paying a sharp premium as SpaceX rallied ~19% on debut, while regulators now probe disclosure failures.

🍏 Apple Hardware Exit Deepens as Vision Pro Chief Joins OpenAI: Apple’s VP behind Vision Pro and its smart glasses push has reportedly defected to OpenAI’s hardware team, stripping key leadership just as Apple pivots toward lighter AR form factors. The move underscores rising talent flow from Apple into AI-native hardware efforts, as OpenAI signals it wants to control devices, not just models (read more).

🎨 AI ART

Want More AI News?
Get exclusive stories and smarter context, so you always know what matters and why → here