💳 San Francisco’s new hustle: corporate cards now swipe on Saturdays

Ramp analyzed transaction data from its corporate cards and spotted an unusual trend, San Francisco companies are increasingly active on Saturdays, signaling a cultural shift toward a 72-hour workweek in the city’s tech scene.

🖱 Spending activity has expanded to a “9 AM to 9 PM, six days a week” rhythm, according to Ramp’s analytics team.
🖱 The uptick in Saturday transactions is small, measured in tenths of a percent, but statistically new and consistent.
🖱 The shift is unique to San Francisco and wasn’t present in previous years.
🖱 The Saturday surge isn’t limited to AI startups, it’s showing up across sectors.
🖱 Ramp notes this could reflect the city’s post-AI boom intensity, where founders and employees are simply working longer.

A subtle but telling signal: San Francisco’s work culture may be evolving again, from “always online” to “always on.”

✔️Powered by Trade Watcher

🤖 xAI expands funding round to $20B with Nvidia’s backing

Elon Musk’s AI startup xAI is doubling its planned fundraising to $20 billion, with Nvidia joining as a key investor, according to Bloomberg. The deal mixes equity and debt through a special-purpose vehicle designed to buy GPUs for xAI’s massive Colossus 2 data center in Memphis.

🖱 Nvidia is contributing up to $2B to the equity portion, helping finance AI infrastructure while deepening ties with Musk.
🖱 The SPV will purchase Nvidia chips and rent them to xAI for five years, letting Wall Street investors recoup returns backed by the hardware rather than company equity.
🖱 The round includes roughly $7.5B in equity and $12.5B in debt, with Apollo Global Management and Diameter Capital among participants.
🖱 Valor Capital is leading the equity component, while Apollo is also investing.

This structure may set a precedent for GPU-backed financing across the AI sector - offering a new model for firms seeking capital-intensive compute without diluting ownership.

✔️Powered by Trade Watcher

💼 a16z denies reports of opening an India office, calls it “fake news”

Andreessen Horowitz (a16z) has publicly dismissed claims that it’s setting up operations in India, calling recent media reports “fake news.” The statement follows speculation that the Silicon Valley VC firm was preparing to launch a local office and hire a team on the ground.

🖱 Reports earlier suggested a16z was hiring for investment and operations roles in India as part of an Asia expansion strategy.
🖱 The firm quickly refuted the story, saying it has no immediate plans to establish a physical presence or team in India.
🖱 a16z remains focused on U.S.-based investments but continues to back startups globally through its existing funds.
🖱 The rumor comes as several U.S. venture firms, including Sequoia, Lightspeed, and Accel have deepened operations in India amid a rebound in tech deal flow.

While a16z says there’s “nothing to see here,” the timing of the rumor hints at growing interest from global funds eyeing India’s booming AI and fintech ecosystems.

✔️Powered by Trade Watcher

🛡️ Defence tech funding soars as startups reshape global security

Dealroom’s new State of Defence Tech 2025 report shows an explosive rise in private investment across military and dual-use technologies, from autonomous drones to cybersecurity and satellite intelligence. The charts tell one clear story: defence tech is officially a venture category.

🖱 Global defence tech funding has surged 7x since 2019, with over $120B in cumulative private investment.
🖱 The U.S. and Europe dominate, but Israel and South Korea are emerging as major innovation hubs for AI-enabled military systems.
🖱 Startups in autonomous systems, space, and cybersecurity attract the majority of capital driven by wars in Ukraine and the Middle East, and renewed NATO spending.
🖱 Corporate and sovereign funds are stepping in, with new specialist investors like Shield Capital, Stellar Ventures, and Narya leading rounds once considered too “sensitive” for traditional VC.

Defence tech has gone mainstream and the next frontier isn’t just AI for peace, but AI for deterrence.

✔️Powered by Trade Watcher

💥 Markets tumble as Trump reignites U.S. China trade war fears

The S&P 500 fell over 2% on Friday after Donald Trump threatened 100% tariffs on Chinese goods, escalating tensions just as Beijing tightened export controls on rare earth metals, materials critical to global tech and defense supply chains.

🖱 Trump said he sees “no reason” to meet with Xi Jinping and warned of massive new tariffs if China doesn’t reverse its restrictions.
🖱 Beijing’s new rules require export licenses for any goods containing Chinese-origin rare earths, rattling supply chains in EVs, semiconductors, and defense tech.
🖱 Investors fled tech and manufacturing stocks, sending the Nasdaq down 3.6%, its worst drop since April.
🖱 Analysts warn the move could revive 2018-style trade volatility, with rare earths emerging as a key strategic weapon in the U.S. China standoff.

If tariffs rise and supply chains tighten, the next global “chip war” may not be about silicon, but about the metals that power it.

✔️Powered by Trade Watcher

🤖 Reflection AI raises $2B to build America’s open frontier lab

Reflection AI has secured a massive $2 billion round at an $8 billion valuation, aiming to become the U.S. answer to China’s DeepSeek, an open frontier AI lab blending transparency, sovereignty, and scale. Founded by ex-DeepMind researchers, the startup wants to make frontier models publicly accessible while keeping data and safety controls intact.

🖱 Founded in 2024 by DeepMind veterans Misha Laskin and Ioannis Antonoglou, best known for work on AlphaGo and Gemini.
🖱 Plans to train a frontier-level MoE model on tens of trillions of tokens, releasing the model weights to promote open innovation.
🖱 Target customers: enterprises and governments seeking sovereign AI systems they can host and control independently of Big Tech.
🖱 Backers include Nvidia, Sequoia, Lightspeed, DST, and GIC, underscoring deep investor appetite for open, domestic AI infrastructure.

Reflection is positioning itself as America’s open frontier counterweight, betting that the future of AI power lies not in secrecy, but in sovereignty.

✔️Powered by Trade Watcher

⚙️ Qualcomm acquires Arduino to power the next wave of AI robotics

Qualcomm has acquired Italian open-source hardware pioneer Arduino, marking a bold step into the robotics and edge-AI market. The move gives Qualcomm access to Arduino’s 33-million-strong developer base and positions it at the heart of the global maker and embedded-AI community. Terms of the deal were not disclosed.

🖱 Arduino will retain its brand and open-source mission, continuing to support chips from multiple vendors, a key reassurance to its loyal developer community.
🖱 Qualcomm and Arduino unveiled the Arduino Uno Q, powered by Qualcomm’s new Dragonwing processor, blending microcontroller precision with on-device AI capability.
🖱 The board runs Linux Debian for high-level tasks and real-time control, and introduces App Lab, an integrated coding environment that bridges C/C++ robotics with AI scripting.
🖱 With a $44 entry price, the Uno Q targets AI hobbyists, robotics startups, and education labs seeking accessible edge-AI prototyping tools.

By merging Arduino’s grassroots innovation with Qualcomm’s chip muscle, the deal signals a new era where AI starts not in the cloud, but on the workbench.

✔️Powered by Trade Watcher

📊 Crypto Venture Weekly: October 6–10, 2025

$2.66B was raised across 25 projects this week, dominated by Polymarket’s $2B strategic round and Kalshi’s $300M Series D. Meanwhile, DDC Enterprise and Amdax allocated a combined $250M to Bitcoin reserves.

Here’s what the top 10 are building 👇

🖱 Polymarket ($2B, Intercontinental Exchange)
Onchain prediction market enabling real-world event trading across politics, economics, and sports.

🖱 Kalshi ($300M, a16z, Sequoia, Coinbase, Paradigm, CapitalG)
CFTC-regulated event market offering compliant access to prediction trading.

🖱 Meanwhile ($82M, Haun, Bain, Pantera, Apollo)
Life insurance platform fully denominated in Bitcoin.

🖱 Bee Maps ($32M, Pantera, Borderless, LDA Capital, Ajna)
Map data and fleet intelligence network powered by Hivemapper contributors.

🖱 Coinflow ($25M, Pantera, Coinbase, Jump, CMT Digital)
Global stablecoin payment infrastructure with AI-driven fraud prevention.

🖱 Anthea ($22M, Yunfeng Financial Group)
Crypto fintech group offering onchain life insurance and wealth products.

🖱 CipherOwl ($15M, General Catalyst, Flourish, Coinbase, OKX)
Compliance infrastructure platform for digital assets.

🖱 TransCrypts ($15M, Pantera, Techstars, Mark Cuban)
Blockchain-based digital document management and verification platform.

🖱 Fanable ($11.5M, Polygon, Borderless, Morningstar, Ripple Ventures)
Web3 collectibles marketplace for Pokémon, comics, and fan assets.

🖱 Block Street ($11.5M, Hack VC, Jane Street, Point72)
Execution and lending layer for tokenized assets.

Investor focus this week revolved around prediction markets, tokenized finance, compliance rails, and AI-integrated infrastructure — signaling a maturing intersection between crypto, regulation, and traditional capital markets.

✔️Powered by Trade Watcher

💸 The AI money loop: billions circulating inside Silicon Valley’s own ecosystem

Bloomberg uncovered how billions in the AI sector now circulate within a tight financial loop where the same players buy, sell, and invest in each other’s technology, reinforcing their dominance.

🖱 OpenAI spends billions on Nvidia GPUs to train its frontier models.
🖱 Nvidia then invests part of its profits back into OpenAI and its ecosystem startups.
🖱 Oracle buys Nvidia chips and simultaneously signs a $300B cloud deal to host OpenAI workloads.
🖱 Smaller startups like Anysphere purchase OpenAI tech while OpenAI itself invests in them.

The result is a closed capital circuit where a few companies OpenAI, Nvidia, Oracle shape the direction of the entire industry.

Altman’s real edge? Not engineering brilliance, but financial orchestration setting the tone for how AI power moves through money, not code.

✔️Powered by Trade Watcher

🎥 Top 7 YouTube Channels to Get Smarter in AI

If you want to truly understand how neural networks, LLMs, and generative AI work, these 7 channels will level up your AI knowledge fast:

🖱 3Blue1Brown — visual math that makes neural networks intuitive, not intimidating.
🖱 Two Minute Papers — cutting-edge research explained in bite-sized videos.
🖱 Yannic Kilcher — deep analysis of AI papers, trends, and model architectures.
🖱 Lex Fridman — long-form interviews with leading AI researchers and founders.
🖱 Sentdex — hands-on tutorials in Python, machine learning, and PyTorch.
🖱 Henry AI Labs — analytical breakdowns of new research and open models.
🖱 DeepLearningAI — Andrew Ng’s official channel for practical AI courses and case studies.

Each of these channels helps you stay sharp and fluent in the language of modern AI, from math to models to mindset.

✔️Powered by Trade Watcher

📈 VC funding jumps 38% as AI mega-deals dominate Q3 2025

Global venture funding surged to $97 billion in Q3 2025, up 38% year-over-year, with nearly half of all capital flowing into AI. According to Crunchbase, the quarter marked a sharp rebound in late-stage investing and an unprecedented concentration of money among frontier AI players.

🖱 AI led the charge, attracting around $45 billion, or 46% of total VC funding.
🖱 Anthropic alone captured almost 30% of all AI capital, reinforcing the dominance of a few large model developers.
🖱 Hardware and infrastructure saw $16B in deals, driven by GPU demand and AI data centers.
🖱 Late-stage rounds hit $58B, up 66% YoY, showing investors are doubling down on scale rather than seed.
🖱 The U.S. captured two-thirds of global funding, solidifying its lead in the AI and semiconductor race.

Venture money is no longer spreading across the ecosystem, it’s circulating between the same frontier players, fueling a tight capital loop between model labs, chipmakers, and cloud providers.

The new question isn’t who’s innovating fastest, it’s who controls the funding circuit of AI itself.

✔️Powered by Trade Watcher

🧠 Thinking Machines co-founder Andrew Tulloch departs to rejoin Meta’s AI team

Andrew Tulloch, co-founder of Thinking Machines Lab alongside former OpenAI CTO Mira Murati, has left the company to return to Meta, marking one of the first major leadership exits from the young frontier-AI startup.

🖱 Tulloch previously spent over a decade at Meta before joining OpenAI and later co-founding Thinking Machines in early 2025.
🖱 Meta reportedly offered a package worth up to $1.5 billion in stock and bonuses as part of its aggressive AI talent push, though the company denied that figure.
🖱 Thinking Machines, launched in February 2025, has raised $2 billion to build open-access frontier models and recently unveiled Tinker, its fine-tuning API.
🖱 The company remains led by Murati, who aims to position it as a transparent, sovereign alternative to closed-model labs like OpenAI and Anthropic.

Meta’s latest high-profile hire shows how the race for frontier-model talent is escalating, even open-AI labs aren’t immune to Big Tech gravity.

✔️Powered by Trade Watcher

⚖️ OpenAI risks billions as court weighs privilege in major copyright case

OpenAI faces potential multibillion-dollar exposure as a federal court considers whether it waived attorney-client privilege in a key copyright dispute. The outcome could determine whether plaintiffs gain access to internal communications about allegedly infringing datasets used to train its models.

🖱 The case centers on OpenAI’s deletion of a dataset sourced from pirated books, plaintiffs argue this was done to hide evidence of infringement.
🖱 OpenAI claimed privilege over its internal discussions, but the court is now weighing whether those protections were waived by partial disclosure.
🖱 If privilege is lifted, plaintiffs could access Slack logs, emails, and memos revealing how the company handled copyrighted material.
🖱 Legal experts warn that exposure could trigger enhanced damages, up to $150,000 per work, or sanctions for spoliation or bad-faith conduct.

The dispute highlights a growing legal risk for AI developers: the moment transparency in training data collides with the limits of legal privilege.

✔️Powered by Trade Watcher

🚀 International Venture Highlights; Week in Review

Venture activity surged across global markets this week, led by AI megadeals, new European funds, and rising investor optimism.

🖱 AI funding boom in Q3 2025
Global VC investment hit $97 billion, up 38% YoY, with nearly 46% of all capital flowing into AI startups.
Reuters and Crunchbase both note that growth is being driven by late-stage megarounds and AI infrastructure plays.

🖱 Iceye eyes a $2.5B valuation
Finnish satellite imaging firm Iceye is in talks for a new round valuing it around $2.5 billion. The Company backed by BlackRock, is riding the defense-tech wave as governments expand spending on real-time Earth observation.

🖱 Notion Capital launches $130M Growth Fund III
London’s Notion Capital raised $130 million to back growth-stage European startups in AI, defense, and logistics. The goal: bridge Europe’s funding gap between Series A and later rounds.

🖱 Concept Ventures unveils Europe’s largest pre-seed fund ($88M)
Closed in just three weeks, the new $88 million fund marks Europe’s largest dedicated pre-seed vehicle, signaling a surge of early-stage confidence.

🖱 Investor sentiment rebounds in Europe
A new EIF survey shows a sharp uptick in optimism among European VCs in Q3, especially toward deeptech and AI infrastructure plays.

The European ecosystem is heating up, from pre-seed to frontier defense tech, momentum is clearly shifting toward strategic, capital-intensive innovation.

✔️Powered by Trade Watcher

⚙️ California becomes first state to regulate AI companion chatbots

California has enacted SB 243, the first U.S. law specifically regulating “AI companion chatbots,” setting new rules on safety, transparency, and age protection after a series of incidents involving emotionally manipulative AI systems. The law takes effect January 1, 2026.

🖱 SB 243 requires clear disclosure that users are speaking with an AI, along with age verification to protect minors and restrictions on explicit or sexual content in AI chats.
🖱 AI companions are barred from posing as mental-health professionals, and must implement protocols for detecting and reporting self-harm or suicide-related content.
🖱 Developers must provide data to the state health department on flagged interactions and introduce break reminders for underage users.
🖱 Violations, including misuse of deepfake content, can trigger penalties up to $250,000 per offense.

California’s move sets a national precedent: the first legal acknowledgment that emotional AI isn’t just software, it’s a social actor that can harm, manipulate, or comfort millions.

✔️Powered by Trade Watcher

👰‍♀️ Phia raises $8M using Gen Z playbook to reinvent fashion search

Phoebe Gates and Sophia Kianni have raised $8 million for Phia, a fashion-tech startup that acts like “Google Flights for fashion,” letting users compare prices across online retailers. The seed round, led by Kleiner Perkins with investors like Kris Jenner, Hailey Bieber, and Sara Blakely, closed in just 3.5 weeks.

🖱 Phia launched in April 2025 and already claims ~500,000 users, driven by viral, founder-led marketing and transparent storytelling.
🖱 The founders relied on social-media outreach and community-driven hiring, openly building in public and treating users as collaborators.
🖱 Their podcast “The Burnouts” doubles as a growth funnel, building visibility and credibility among Gen Z audiences.
🖱 The team uses AI tools to ideate and scale content, blending authenticity with automation to sustain momentum.

Phia shows how the next generation of founders is merging influence, transparency, and AI-assisted storytelling to turn social capital into venture-scale traction.

✔️Powered by Trade Watcher

🚀 U2U Network — powering the DePIN wave across AI, devices, and real-world infra

U2U is a modular Layer-1 chain building the backbone for DePIN, connecting AI systems, IoT devices, and physical infrastructure through fast, scalable onchain architecture.

🖱 Market cap up +1581% since launch
🖱 +113% growth in daily transactions
🖱 +214% increase in verified contracts
🖱 40+ projects already powered by U2U
🖱 Over 1M active users across Southeast Asia
🖱 Recent Kraken listing fueling strong trading volume and a clear bullish setup on charts

With both technical and price momentum accelerating, U2U is emerging as one of the strongest DePIN plays to watch this cycle - bridging digital intelligence and real-world utility.

✔️Powered by Trade Watcher

🚤 Cornelius Vanderbilt built America’s infrastructure empire from a single boat

Cornelius Vanderbilt, nicknamed “The Commodore”, rose from ferrying passengers across New York Harbor to controlling America’s railroads. He didn’t invest in ideas; he owned the systems that made the economy run.

🖱 At 16, he borrowed $100 from his mother to buy a small boat, by 23, he ran a fleet and crushed rivals by slashing prices to zero.

🖱 He saw competition as war, bankrupting opponents through predatory pricing, then buying their businesses for pennies.

🖱 In the 1850s, he shifted from steamships to railways, founding New York Central Railroad and building Grand Central Terminal.

🖱 By his death in 1877, Vanderbilt was worth $100 million, about $300 billion today, making him America’s first self-made billionaire.

🖱 He had little formal education, struggled to write, but possessed a near-photographic memory for routes, numbers, and deals.

🖱 When Nicaraguan officials crossed his shipping line, he hired a private army and reclaimed control, a 19th-century corporate invasion.

🖱 He loathed bureaucracy, saying, “The law is too slow, I’ll ruin you myself.” He preferred money and manipulation over courts.

🖱 Vanderbilt never took partial stakes, he demanded control. For him, power over decisions mattered more than profit.

He didn’t follow trends, he owned the infrastructure that made them possible. His modern heirs: Stripe, AWS, OpenAI API, Notion, the invisible engines of today’s economy.

✔️Powered by Trade Watcher

⚡ Apple unveils M5, its biggest leap in AI performance yet

Apple has introduced the M5 chip, the next generation of Apple Silicon built for an on-device AI era. Designed on a new 3-nanometer process, M5 powers the latest MacBook Pro, iPad Pro, and Vision Pro, delivering major boosts in GPU, Neural Engine, and memory bandwidth performance.

🖱 The M5 introduces Neural Accelerators directly inside each GPU core, enabling up to 4× faster AI compute than M4 for tasks like local LLMs and image generation.
🖱 Built on third-generation 3 nm technology, it features a 10-core CPU and 10-core GPU, with a 16-core Neural Engine and 153 GB/s unified memory bandwidth — about 30% higher than before.
🖱 Apple’s Core ML and Metal frameworks automatically optimize for M5, allowing developers to run complex diffusion and transformer models fully on-device.
🖱 Vision Pro gains up to 10% higher pixel throughput and 120 Hz refresh rates, while laptops see better performance per watt for longer battery life.

Apple’s M5 marks a shift in its silicon strategy, from raw speed to AI-centric design, blurring the line between GPU and neural compute and tightening its grip on the local AI ecosystem.

✔️Powered by Trade Watcher

🤖 China’s factories run nonstop powered by 300,000 new robots

China may become the first nation where robots overtake humans in production. Facing a shrinking population, the country is rapidly automating its factories, in 2024 alone, 295,000 industrial robots were installed, accounting for over half of all new robots worldwide.

🖱 The total number of active industrial robots in China has reached 2.027 million, the highest in the world, spanning sectors from automotive to electronics.
🖱 Robots now handle most critical manufacturing tasks, welding car frames, assembling circuit boards, packing goods, and moving heavy materials.
🖱 Automation is no longer just about efficiency, it’s China’s answer to a structural labor shortage driven by demographics and rising wages.

If the trend continues, China could become the first fully robotized industrial economy, where growth no longer depends on human population size but on machine scalability.

✔️Powered by Trade Watcher

😱 Prisons, Torture, and Crypto Scams: The Fall of Cambodia’s “Billionaire” Chen Zhi

Chen Zhi, the flashy founder of Prince Holding Group, once hailed as Cambodia’s richest man, has been hit with a U.S. Department of Justice case after $15 billion in bitcoin was seized from him. Behind his conglomerate’s glossy facade of real estate, finance, and “social progress,” investigators allege a global web of fraud, torture, and human exploitation.

🖱 Prince Holding Group, active in 30 countries, allegedly ran private labor camps in Cambodia, fortified complexes where hundreds of people were detained and forced to operate crypto and investment scams targeting victims worldwide.

🖱 Victims were tricked through fake “profitable investment” pitches, while call centers with thousands of phones generated billions in stolen funds, later laundered through the Group’s legal businesses, including gambling and mining.

🖱 The U.S. seized 127,271 bitcoins linked to Chen Zhi, worth about $15 billion, from “non-custodial wallets” under his control. Documents found reportedly show Chen designed the money-laundering flows, kept photos of torture, and even beat detainees himself.

🖱 Ironically, just last year Chen Zhi was awarded “Entrepreneur of the Year” and “Visionary Leader in Social Progress.”

From polished billionaire to accused crime lord, Chen Zhi’s story reveals the dark evolution of some Southeast Asian conglomerates, where luxury, crypto, and coercion blend into a chilling new kind of global crime network.

✔️Powered by Trade Watcher