Tech News

Story of the Day

Nvidia has partnered with six Wall Street giants, including BlackRock and Goldman Sachs, to establish a monumental $500 billion financing program to fund AI infrastructure buildouts. The move marks a crucial macro-economic shift toward treating “AI factories” as a standardized, investable infrastructure asset class, with Nvidia putting up to 25% of its own capital on the line to backstop the risk.

Top Stories

Wall Street Mobilizes $500 Billion for Nvidia-Backed AI Infrastructure · New York Times Six major investment firms, including Apollo, Blackstone, BlackRock, Brookfield, Goldman Sachs, and KKR, have launched an independent financing platform to deploy massive institutional capital into AI data centers. Under the partnership, Nvidia-based “AI factories” will be treated as repeatable, productive infrastructure assets rather than risky, project-by-project tech ventures. To reassure safety-seeking capital pools like pensions and insurance floats, Nvidia is backstopping the deals by providing up to 25% residual-value financing—essentially putting its own massive profits on the line to keep the hardware engine running.

Meta and TikTok Denied Section 230 Immunity in $1.4 Trillion Youth Addiction Lawsuit · Ars Technica The US Court of Appeals for the 9th Circuit has rejected claims from Meta and TikTok that they are legally immune to state-led youth addiction lawsuits under Section 230 of the Communications Decency Act. The landmark 3-0 ruling clears the way for a massive trial brought by a bipartisan coalition of state attorneys general seeking an astronomical $1.4 trillion in damages. Jury selection is scheduled to begin immediately, with the full trial starting on August 19 to determine if the platforms’ underlying recommendation algorithms were intentionally designed to hook minor users.

Coinkite Coldcard Breach Drains $130 Million in Bitcoin, Shaking Self-Custody Believers · Bloomberg A devastating security breach targeting Coinkite’s “cold” Bitcoin wallets has resulted in the theft of approximately $130 million in cryptocurrency across 7,300 wallet addresses. Security experts suggest the exploit may have been executed or assisted by advanced AI hacking models, dealing a major blow to long-held industry confidence in self-custody hardware. The incident is expected to shake investor confidence and potentially accelerate capital outflows from private hardware wallets into regulated spot Bitcoin ETFs.

Anthropic Inks Colossal $9.1 Billion Compute Deal with Riot Platforms · Bloomberg Claude developer Anthropic has struck a massive $9.1 billion agreement with Riot Platforms, a prominent Bitcoin mining company that has aggressively pivoted to selling AI data center capacity. The 20-year contract secures 191 megawatts of power capacity from Riot’s campus in Rockdale, Texas, to help fuel Anthropic’s rapidly expanding frontier model training and inference requirements. The astronomical outlay highlights how critical locking down raw power and physical infrastructure has become as frontier labs struggle to secure enough compute to train next-generation models.

Two-Month-Old River AI Secures $1.1 Billion in Funding Led by General Catalyst · TechCrunch River AI, a personal agent startup founded just two months ago by former xAI co-founder Igor Babuschkin, has closed a massive $1.1 billion funding round. Co-led by General Catalyst, the astronomical valuation demonstrates the venture capital market’s insatiable and aggressive appetite for agentic AI technology. Babuschkin’s new venture aims to build highly customizable, open-source AI personal agents that run outside the control of consolidated tech conglomerates.

OpenAI Chief Operating Officer Brad Lightcap Steps Down to Launch New Venture · The Verge Brad Lightcap, OpenAI’s former COO and special projects lead, has announced his departure from the company after an eight-year tenure to start “something new”. In an internal memo, Lightcap expressed excitement to continue advancing the lab’s mission from an external, different vantage point. His exit adds to a broader wave of executive and scientific departures across frontier labs, following major reshufflings at Google DeepMind with Demis Hassabis and Jeff Dean stepping away from day-to-day operations.

Spotify Introduces ‘AI Persona’ Labels and Excludes Fake Artists from Recommendations · The Verge Spotify is rolling out an “AI Persona” profile badge for artists who do not represent a real human, excluding their music from algorithmic, personalized, and editorial recommendations by default. The streaming giant will allow creators to self-identify but will also employ human reviews and AI detection tools to scan for photorealistic AI-generated identities. The defensive policy represents a major move to protect human artists’ careers and control the tidal wave of low-effort AI-generated tracks flooding the service.

Bumble Scraps Core Rule Requiring Women to Make the First Move · The Verge In a major strategic reversal, Bumble has eliminated the requirement that only women can send the first message in heterosexual matches—a core mechanic of the app since its 2014 launch. The app is also extending its response deadline from 24 to 72 hours in a bid to court Gen Z users and reverse stagnant growth. Executives emphasized that while the original rule was revolutionary, a modern “women-first” design requires giving users more flexible ways to connect.

Also Worth Knowing

• Claude Will Apply Invisible Watermarks to AI-Generated Text and Images (The Verge): Anthropic committed to marking Claude-generated outputs with digital watermarks and signed provenance metadata to comply with incoming European AI Act transparency rules.
• AI-Discovered Zoom Vulnerability Allowed Device Hijacking on Calls (Wired): Researchers used a public AI model with under 20 prompts to find a zero-day screen-sharing and real-time annotation flaw (now patched) that allowed silent takeover of Zoom call participants’ devices.
• DeepSeek Overtakes Google in Token Volume as Model Mix Shifts (Vercel): Vercel’s latest AI Gateway Production Index reveals that DeepSeek has surpassed Google to become the second-largest lab by token volume, driven by the massive popularity of its ultra-cheap V4 Flash model.
• Chrome Adopts Device-Bound Session Credentials to Combat Cookie Theft (Ars Technica): Google Chrome added a new security feature (DBSC) that cryptographically binds session cookies to a device’s hardware-isolated TPM or secure enclave, preventing attackers from hijacking authenticated sessions even if a cookie is stolen.
• Desktop Linux Market Share Briefly Spikes to 22% on Workday (Slashdot): Cloudflare data shows a post-holiday workday surge in North American Linux desktop usage to 22%, fueled by developers and automated AI agents running resource-intensive local environments on the OS.

📊 I can compile these major venture funding rounds and hardware deals into a structured spreadsheet to help you track capital flow across the AI supply chain.