Venture Capital

⚖️ Amazon sues Perplexity over AI agent that shops on its own Amazon has filed a lawsuit against Perplexity AI, claiming the startup’s browser agent “Comet” made unauthorized purchases on Amazon’s marketplace violating the platform’s terms of service and U.S. computer-fraud laws. 🖱 According to Amazon, Comet allows users to buy items without disclosing that an AI, not a person, is completing the transaction. 🖱 The case was filed in federal court in San Francisco and demands Perplexity immediately halt all such automated purchases. 🖱 Amazon argues the practice undermines marketplace integrity and poses risks for fraud and misrepresentation. 🖱 Perplexity, known for its fast-growing AI search and agent tools, has not yet issued a detailed public response.

The dispute highlights a growing legal gray zone: how far autonomous AI agents can go in acting for humans in online commerce. With Amazon defending platform control and Perplexity pushing agentic innovation, this could set a landmark precedent for AI behavior in digital marketplaces.

✔️Powered by Trade Watcher

🚀 Google to launch solar-powered space data centers with TPUs by 2027 Google has unveiled Project Suncatcher, an ambitious plan to deploy data centers in orbit, powered entirely by solar energy and equipped with its proprietary TPU chips for AI processing. 🖱 The first two satellites, each carrying four TPUs, are scheduled for launch in early 2027 to test full operational stability in orbit. 🖱 At an altitude of 650 km, the satellites will continuously harvest solar power unaffected by weather, time of day, or seasons. 🖱 Vacuum-based cooling could reduce maintenance costs by up to 40%, a major efficiency gain over Earth-based data centers. 🖱 Google says its TPUs have passed radiation resistance tests, performing 15× above the mission’s minimum threshold a strong sign of hardware readiness.

While the hardware works, key challenges remain from data transmission latency to orbital maintenance. And with startups like Starcloud aiming to have a full space data center running by 2027, Google’s race to scale AI infrastructure beyond Earth is officially on.

✔️Powered by Trade Watcher

🎲 Coinbase CEO’s word stunt exposes cracks in prediction markets Coinbase CEO Brian Armstrong deliberately dropped a string of buzzwords, “Bitcoin, Ethereum, blockchain, staking, Web3” during the company’s earnings call after learning that $84,000 was wagered on whether he’d say them. 🖱 Armstrong acknowledged tracking the bets in real time, then said the words live, instantly determining the outcome of active prediction-market contracts. 🖱 The incident revealed how “mention markets,” where users bet on specific phrases in public events, can be easily manipulated when the subject knows the wager. 🖱 Platforms like Kalshi and Polymarket now face scrutiny over whether such bets qualify as legitimate prediction tools or border on market manipulation. 🖱 Coinbase’s otherwise strong earnings were overshadowed by the episode, raising governance and regulatory questions around CEO conduct.

Prediction markets promise transparency but when the speaker controls the outcome, the odds no longer mean anything.

✔️Powered by Trade Watcher

🤖 Ex-consultants are now training AI models to do their old jobs Former McKinsey, Bain, and BCG consultants are being hired to train AI systems designed to replicate the work of junior consultants from research to client slide decks. 🖱 Around 150 ex-consultants are working on a project called Argentum, run by a data-labelling firm developing models that can perform entry-level consulting tasks. 🖱 The AI is being trained to structure problems, analyze data, and produce client-ready materials, the same outputs that typically occupy analysts and associates. 🖱 By using real consultants to label and refine data, the system aims to capture not just workflows but the reasoning patterns behind professional consulting work. 🖱 The move points to a future where consulting’s “junior labor layer” could be automated, cutting both costs and entry-level opportunities.

The consulting pyramid may be flattening not by design, but by code.

✔️Powered by Trade Watcher

💼 Director pay in startups shows how risk scales faster than reward Directors at startups valued between $25 million and $250 million see only modest pay growth as valuations climb but their equity stakes shrink dramatically. 🖱 A tenfold increase in valuation doesn’t bring a tenfold rise in salary; compensation grows only slightly as companies mature. 🖱 Stock grants in larger startups are about 75% smaller than those in smaller ones, reflecting lower dilution tolerance and more stable structures. 🖱 Equity in later-stage startups is generally more likely to be worth something, while early-stage shares can easily end up worthless. 🖱 The trade-off highlights how financial upside narrows as companies derisk and valuations soar.

The startup pay curve reveals a clear truth, the bigger the company, the smaller the gamble, but also the smaller the win.

✔️Powered by Trade Watcher

🍏 Apple targets nearly $140B holiday quarter with new iOS and iPad lineup Apple is gearing up for a record ~$140 billion holiday quarter, driven by fresh software updates and a wave of hardware refreshes, as the company approaches its 50th anniversary. 🖱 Software push: The rollout of iOS 26.1 is imminent, with iOS 26.2 already in early beta, signaling Apple’s continued focus on ecosystem upgrades. 🖱 Hardware refresh: The MacBook Air, iPad Air, and iPad mini are all getting OLED displays, marking a major leap in screen quality across Apple’s mid-range lineup. 🖱 Retail revamp: Updated in-store product displays and experiences will launch starting November 12, timed with the holiday shopping surge. 🖱 Revenue momentum: The company’s near-record projection reflects strong pricing power and resilient demand across its product ecosystem.

Apple’s $140 billion target highlights how the company’s real advantage lies not just in individual devices, but in its tightly integrated ecosystem that turns every update into a catalyst for growth.

✔️Powered by Trade Watcher

💰 Amazon signs $38B deal to supply OpenAI with Nvidia-powered compute Amazon Web Services (AWS) has struck a seven-year, $38 billion agreement to provide OpenAI with access to hundreds of thousands of Nvidia GPUs, one of the largest infrastructure deals in AI history. 🖱 The deal secures OpenAI long-term access to Nvidia’s high-end chips through AWS data centers, ensuring stable compute supply for training and inference at scale. 🖱 It marks a deeper strategic partnership between Amazon and OpenAI, strengthening AWS’s position in the cloud race against Microsoft Azure and Google Cloud. 🖱 Nvidia, though not a signatory, remains the biggest beneficiary as its GPUs continue to underpin nearly every large-scale AI deployment. 🖱 The agreement underscores both the scale of OpenAI’s compute demand and its growing dependency on AWS infrastructure.

The scramble for AI compute is accelerating, and Amazon’s deal with OpenAI signals a power shift, where control over infrastructure, not model innovation, is becoming the real currency of dominance in the AI era.

✔️Powered by Trade Watcher

🧠 OpenAI founder speaks out on talks with Anthropic and internal conflicts In a deposition related to the company’s November 2023 leadership shake-up, co-founder Ilya Sutskever shared new details around merger discussions with Anthropic and internal tensions at OpenAI. 🖱 A merger proposition: Sutskever said that following the removal of Sam Altman, OpenAI’s board approached Dario Amodei of Anthropic about a potential deal. According to Sutskever, Anthropic was initially “excited” by the possibility. 🖱 Internal memo: Sutskever recorded a memo accusing Altman of “a consistent pattern of lying, undermining his execs, and pitting his execs against one another.” 🖱 Strategic fork: The merger discussions apparently stalled as Anthropic opted to chart its own path rather than combine with OpenAI. 🖱 Showdown for control: The deposition portrays the unfolding power struggle, not just with external rivals, but internally over mission, safety, transparency and leadership style.

AI companies continue to scale rapidly, these revelations shine a light on the growing pains and governance risks behind the scenes particularly the tension between commercial drive, safety oversight and founder-control.

✔️Powered by Trade Watcher

🚀 SpaceX marks 100th Starlink launch of 2025, pushes satellite network past 8,800 in orbit SpaceX has completed its 100th Starlink mission of the year, reinforcing its lead in global satellite broadband and showcasing the company’s unmatched launch cadence. 🖱 A Falcon 9 rocket lifted off from Vandenberg Space Force Base carrying 28 Starlink V2 Mini satellites into low-Earth orbit at 1:41 p.m. PDT. 🖱 The veteran booster B1063 first launched in 2020 successfully landed on the drone ship Of Course I Still Love You. 🖱 This mission brings total Starlink satellites launched in 2025 to more than 2,500, with over 8,800 now operating in orbit. 🖱 The company has averaged nearly 10 launches per month, highlighting how reusability is driving unprecedented efficiency.

As SpaceX turns its satellite network into a full-scale telecom platform, the milestone underscores a bigger shift from a launch company to a global internet infrastructure provider racing ahead of rivals.

✔️Powered by Trade Watcher

🖌 Ballpoint pen turns 80, the invention that rewrote everyday writing The world’s most widely used writing tool just marked its 80th anniversary. Invented by Hungarian journalist László Bíró, the ballpoint pen transformed how people wrote replacing messy fountain pens with a simple, reliable mechanism still used today. 🖱 Bíró was frustrated by ink smudges and slow drying times, so with his brother György, he created a pen that used a tiny rolling ball to spread quick-drying ink evenly on paper. 🖱 The design, patented in 1938, proved invaluable during World War II when the British Air Force adopted it, fountain pens failed at high altitudes. 🖱 In 1945, the first commercial pens launched in the U.S. and U.K., selling for $12.50 each, a luxury item at the time. 🖱 Later, entrepreneur Marcel Bich licensed the design and mass-produced the plastic “Bic,” turning it into a cheap global staple.

Eight decades later, the pen’s core design remains almost unchanged a reminder that sometimes, the most enduring innovations are the simplest solutions to everyday problems.

✔️Powered by Trade Watcher

📊 Crypto Venture Weekly: October 27–31, 2025 $349M was raised across 26 projects this week, led by Bron Labs ($15M) and ZAR ($12.9M). Meanwhile, OceanPal allocated $120M to NEAR for its digital asset treasury. Gate Ventures, Redacted, ANT.FUN, and Lolli all executed strategic M&A deals, highlighting increasing consolidation across fintech and DePIN. Here’s what the top 8 are building 👇 🖱 Bron Labs ($15M, GSR, LocalGlobe, Fasanara) Self-custodial platform offering institutional-grade wallet security for everyday users. 🖱 ZAR ($12.9M, a16z, Coinbase, Dragonfly, VanEck) Fintech bringing dollar-backed stablecoins to consumers in Pakistan and other emerging markets. 🖱 KapKap ($10M, Animoca, Shima, Mechanism, Big Brain) AI-driven Web3 platform powering the attention economy and value distribution. 🖱 Standard Economics ($9M, Paradigm, Lightspeed) Building Uno — a mobile app for cross-border payments, remittances, and USD stablecoins. 🖱 Intuition ($8.5M, Polygon, ConsenSys, CoinList, F-Prime) Developing a composable trust layer for the internet and AI systems. 🖱 Capybobo ($8M, Pluto.Vision, Animoca, HashKey, Mirana) Web3-native art toy IP bridging physical collectibles with digital identity. 🖱 Accountable ($7.5M, Pantera, OKX Ventures, Karpatkey) Real-time financial verification platform enabling private proof of assets and liabilities. 🖱 DeepSafe ($3M, Spark Digital, Gate Ventures, ViaBTC) AI network for trustless verification of both onchain and offchain communications.

Investor focus this week centered on self-custody, stablecoin access in emerging markets, and verifiable AI infrastructure — with early signals that DePIN and programmable finance may drive the next leg of onchain adoption.

✔️Powered by Trade Watcher

⚖️ ChatGPT banned from acting as a doctor or lawyer OpenAI has updated its usage rules to forbid ChatGPT from providing professional consultations that require a license such as legal or medical advice unless a certified expert is directly involved. 🖱 The restriction applies to how ChatGPT is used, not to the model itself. 🖱 GPT-5 can still discuss medical or legal topics in general or educational terms. 🖱 The change targets misuse where AI could replace licensed professionals. 🖱 OpenAI clarified that the rule is meant to protect users, not limit open discussion.

By separating professional use from open consultation, OpenAI is drawing a clear boundary between AI assistance and licensed expertise.

✔️Powered by Trade Watcher

🔔 Jensen Huang celebrates $5T milestone with Samsung and Hyundai chiefs in Seoul diner Nothing fancy, just NVIDIA’s Jensen Huang marking a $5 trillion market cap at a local diner in Seoul, joined by the leaders of Samsung and Hyundai. After the meal, Huang reportedly picked up the tab for every table in the restaurant. He then ordered extra fried chicken and shared it with fans waiting outside for photos.

A casual moment and a reminder that even trillion-dollar celebrations can happen over street food with friends.

✔️Powered by Trade Watcher

🎵 Udio disables downloads after copyright threats, partners with Universal Music Udio has shut off downloads of AI-generated music following a surge of copyright complaints and warnings of billion-dollar penalties from major labels. 🖱 The startup has signed a new partnership with Universal Music Group to pivot toward generating licensed remixes of existing songs. 🖱 Users can no longer download tracks created on the platform. 🖱 As compensation, Udio has issued free credits to affected users. 🖱 The move follows increasing pressure on AI music tools accused of replicating copyrighted material without consent.

By aligning with UMG, Udio is signaling a shift from the free-for-all AI music scene toward a legally compliant, industry-backed model.

✔️Powered by Trade Watcher

📊 The Venture Capital Risk/Return Spectrum, how risk and reward evolve across different investment stages. Early-stage investors (Seed, Series A) commit smaller amounts but take on higher risk, expecting outsized returns, sometimes 10–100x if a company succeeds. 🖱 As startups mature (Series B to D and beyond), investment size increases but risk drops, leading to more moderate returns, typically 3–5x. 🖱 By the time companies reach the public markets or bonds, they can absorb large sums safely, yet investors trade off high growth potential for stable, lower returns.

The earlier you invest, the higher the risk, but also the greater the possible reward.

✔️Powered by Trade Watcher

⚖️ Meta denies using adult films for AI training after lawsuit Meta is pushing back against a lawsuit alleging it downloaded thousands of copyrighted adult videos to train its AI models, insisting the activity came from personal use by individuals, not company research. 🖱 The suit, filed by Strike 3 Holdings, claims over 2,300 videos were downloaded from IP addresses linked to Meta since 2018. 🖱 Meta says the downloads were sporadic about 22 titles per year and likely tied to employee misuse, not data collection for AI. 🖱 The company emphasized that its AI policies prohibit adult content and that no such data was used in model training. 🖱 Strike 3 is seeking over $350 million in damages, citing copyright infringement and potential misuse of its content for AI.

As legal scrutiny over AI data sourcing intensifies, Meta’s defense could set a precedent for how courts view the difference between corporate intent and individual activity inside major tech firms.

✔️Powered by Trade Watcher

📈 OpenAI prepares for landmark IPO, eyes $1 trillion valuation OpenAI is laying the groundwork for a blockbuster initial public offering that could value the company at up to $1 trillion, according to sources cited by Reuters. 🖱 The company may file with regulators as early as the second half of 2026, with a goal of raising $60 billion or more. 🖱 Its recent restructuring gives the OpenAI Foundation oversight power over the for-profit entity while maintaining investor confidence. 🖱 OpenAI is on track to hit an annualized revenue run rate of $20 billion by year-end, though it still operates at a loss. 🖱 Microsoft, which owns about 27%, is expected to remain a key strategic partner after the listing.

If successful, the offering would be one of the largest in history, marking the moment AI companies move from private labs to the core of global capital markets.

✔️Powered by Trade Watcher

Nvidia became the first company in the world with a market capitalization of 5 trillion dollars. ✔️Powered by Trade Watcher

💼 NVIDIA invests $1B in Nokia to build AI-powered networks NVIDIA is taking a 2.9% stake in Nokia through a $1 billion equity deal, marking a deep move into telecom infrastructure and AI networking. 🖱 The two companies will co-develop AI-native 5G and 6G networks, integrating NVIDIA’s accelerated computing platforms into Nokia’s wireless systems. 🖱 Nokia will deploy NVIDIA’s chips across its RAN and data-center products to improve network speed, efficiency, and edge AI capabilities. 🖱 Initial trials of the joint platform begin in 2026, with commercial rollouts expected soon after. 🖱 The investment strengthens Nokia’s pivot toward software and AI infrastructure, while expanding NVIDIA’s footprint beyond data centers into global connectivity.

By merging compute and network layers, NVIDIA is betting that the next era of AI won’t just run on networks, it will be built into them.

✔️Powered by Trade Watcher

⚡️ Qualcomm takes aim at Nvidia with new AI data-center chips Qualcomm is accelerating its push into the data-center market with two new AI chips, the AI200 and AI250 set to launch in 2026 and 2027, expanding beyond its smartphone roots. 🖱 The AI200 and AI250 are designed for AI inference workloads, offering greater memory and efficiency for running large models in real time. 🖱 Qualcomm will also sell complete rack systems built around the chips, marking a shift from chip supplier to full-stack infrastructure provider. 🖱 A 200-megawatt deployment deal with Saudi-backed startup Humain will roll out these systems starting in 2026. 🖱 Qualcomm shares rose 20% after the announcement, as investors bet on its long-term data-center ambitions.

With Nvidia’s dominance showing cracks amid rising demand and cost pressures, Qualcomm is positioning itself as a new player in the AI compute stack, one that could offer hyperscalers and sovereign AI projects a credible alternative.

✔️Powered by Trade Watcher