Venture Capital

🎯 Michael Burry just reignited a showdown with Nvidia and markets may be entering a turning point According to TechCrunch, Burry is mounting a full-blown assault on Nvidia’s AI-driven valuation and accounting practices and this Thanksgiving drama could ripple across the entire AI industry. 🖱 Burry claims Nvidia’s stock-based compensation has cost shareholders roughly $112.5 billion, effectively slashing owner earnings by about 50%. 🖱 He argues many AI firms especially those using heavy GPU infrastructure are “cooking the books” by stretching depreciation on hardware that loses value fast, overstating profitability in the process. 🖱 In response, Nvidia fired back with a detailed memo to Wall Street analysts, stating Burry’s calculations are flawed: real buyback totals are closer to $91 billion, not $112.5 billion, and its compensation practices are “in line with peers.” 🖱 Burry says he isn’t calling Nvidia a fraud but likens it to Cisco during the dot-com bubble: overbuilt infrastructure, overhyped expectations, and a potential crash when reality sets in. 🖱 Beyond just Nvidia, his critique targets the broader AI infrastructure boom, a wave of optimism grounded more in accounting assumptions than hard cash flow.

Burry’s campaign isn’t just criticism, it’s a warning bell. If enough investors start to scrutinize AI companies’ accounting and cost structures, the next leg of the AI wave could be built not on hype, but on hard financials.

✔️Powered by Trade Watcher

📌 Trump just launched “Genesis Mission,” a national AI push to turbocharge U.S. scientific discovery President Trump signed a major executive order creating Genesis Mission, a federal program that unifies government data, national-lab supercomputers, and AI systems to accelerate breakthroughs in energy, biotech, materials science, and national security. 🖱 Genesis Mission’s core goal is to supercharge scientific discovery using large-scale AI and federal datasets. 🖱 The program directs the Department of Energy and national labs to build a unified AI platform that trains and deploys scientific models. 🖱 It aims to turn America’s huge stockpile of government data into a strategic AI advantage. 🖱 The initiative brings together federal agencies, universities, and private tech companies in one research pipeline. 🖱 Supporters frame it as the biggest U.S. science mobilization since the Apollo era, shifting focus from regulation to full-speed innovation.

Genesis Mission is a bet that the next wave of national power won’t come from new laws or new hardware but from AI-driven breakthroughs the government can produce at scale.

✔️Powered by Trade Watcher

🥲 NVIDIA posts a “congratulatory” tweet to Google with a knife hidden behind the smile NVIDIA suddenly dropped a very polite-but-not-really tweet today, congratulating Google on its AI progress but the subtext was impossible to miss. This comes right after the bombshell that Meta will buy Google TPUs for the first time, a move that hit NVIDIA’s stock by 6%. So NVIDIA decided to remind everyone who still runs the AI world. Here’s the literal meaning behind their velvet-toxic message: 🖱 “We’re happy for Google’s success… and we remain their supplier.” Translation: They still buy our stuff. 🖱 “NVIDIA is at least one generation ahead of the entire industry.” Translation: Your TPUs are cute, but we build the real frontier. 🖱 “We’re the only platform that can run any model anywhere compute happens.” Translation: Everyone else is specialized. We’re universal. 🖱 “NVIDIA offers higher performance, flexibility, and interchangeability than ASIC chips.” Translation: Google’s TPUs? Yeah they’re just ASICs. Limited. Narrow. Not NVIDIA-level.

Meta’s defection stung, and NVIDIA responded the way only a trillion-dollar titan does with a smile sharp enough to draw blood.

✔️Powered by Trade Watcher

🧠 Key takeaways from Ilya Sutskever’s new interview with Dwarkesh Patel 🖱 The scaling era is ending: Ilya argues that “just add more compute” no longer produces major breakthroughs, the next leaps require new scientific discoveries, not bigger clusters. 🖱 Generalization is the real bottleneck: Current models look impressive on benchmarks, but much of that progress is misleading. Models overfit to evals, especially after RL, while researchers unintentionally reinforce the loop by creating similar benchmarks and training setups. 🖱 He knows what’s missing but won’t say: Ilya hints that today’s models lack a core ingredient. While he keeps the details private, the interview suggests SSI is exploring Continual Learning and dynamic reward systems as possible directions. 🖱 AGI timeline: wide uncertainty: Could be 5 years, could be 20. But when it arrives, Ilya believes the safest way for humans to stay useful and stay in control is to merge with AI (a subtle wink toward Neuralink-style interfaces).

A rare interview where Sutskever openly says the frontier has shifted: brute-force scaling is no longer enough.

✔️Powered by Trade Watcher

🔔 Robinhood’s CEO just raised $120M to build “error-free AI" Harmonic, the AI startup co-founded by Robinhood CEO Vlad Tenev, secured $120M at a $1.45B valuation to tackle one problem everyone in AI feels: models still make too many mistakes. 🖱 Harmonic’s core mission is to reduce hallucinations and make AI outputs consistently accurate. 🖱 The company is building an “accuracy layer” that sits on top of existing models to catch and correct errors. 🖱 Investors see reliability as the next frontier, driving major interest in companies focused on precision. 🖱 The product targets enterprises that need trustworthy AI for finance, operations, and regulated sectors. 🖱 Tenev’s involvement signals a push toward safety, guardrails, and real-world dependability not just speed.

Harmonic is betting that the winners in the AI race won’t be the flashiest models, but the ones users can actually trust.

✔️Powered by Trade Watcher

📱 Altman says OpenAI’s new device will feel “more peaceful and calm” than the iPhone Sam Altman offered new hints about OpenAI’s upcoming hardware, describing it as a calmer, less stressful alternative to today’s smartphones. Speaking with TechCrunch, he emphasized that the goal is to rethink personal computing around attention, not distraction. 🖱 The device won’t mimic a phone, Altman says the world doesn’t need another rectangle with apps. 🖱 OpenAI is targeting a “peaceful, ambient” interaction style where the AI handles tasks proactively. 🖱 Jony Ive’s design influence is clear: the hardware is meant to feel physical, elegant, and minimal. 🖱 Altman hinted the device may prioritize voice, sensors, and context over screens. 🖱 He claims the product represents a shift toward “calm technology” computing that fades into the background.

OpenAI wants to build the first AI-native gadget that replaces, rather than competes with, the smartphone’s attention economy.

✔️Powered by Trade Watcher

📊 TAM / SAM / SOM: the three circles of hell for every pitch A simple, no-nonsense breakdown of how to calculate TAM / SAM / SOM without embarrassing yourself in front of investors: TAM — Total Addressable Market The whole universe. The market if you captured 100% of global demand. How to calculate: Top-down: — Use credible market reports — Select the segment tied to your category — Users × average spend Bottom-up (preferred): — Count how many users could use the product — Multiply by your real ARPU/ACV Formula: TAM = potential customers × annual revenue per customer Mistakes: — “global tech market $4T” nonsense — Random blogs with made-up numbers SAM — Serviceable Available Market The part of the market you can actually serve with your geography, niche, channels, model. How to calculate: Cut off everything out of reach: — Geography — Audience definition — Distribution limits — Regulation — Data/API/partner constraints Formula: SAM = reachable customers × your price Example: TAM = all online stores worldwide SAM = online stores in CIS that can plug in your product SOM — Serviceable Obtainable Market The realistic slice you can win in the next 12–36 months. Not “in theory.” With your team, budget, funnel, and sales speed. How to calculate: Start with SAM → apply a believable share (0.5–5%). Justify with your funnel: — Leads you can bring — Conversion logic — Retention Formula: SOM = realistic share × SAM Rule: If you claim more than 1–3% early on, nobody believes you. Bottom-up example (B2B SaaS for cafés): — 60,000 cafés in Russia → TAM = 60,000 × $800 = $48M — 15,000 cafés with 3M+ RUB turnover → SAM = 15,000 × $800 = $12M — You capture 2% in 2 years → 300 clients → SOM = 300 × $800 = $240k/year The trio in short: TAM — the idea’s theoretical ceiling SAM — the business’s real opportunity SOM — the team’s reality check

Real growth starts the moment the fantasies stop.

✔️Powered by Trade Watcher

🔥 Stay connected anywhere — without roaming. Teleleo gives you full remote control of your SIM card, so your number works for you no matter where you live or travel. ✔️ With Teleleo platform you can: ● Receive and make calls without international charges ● Receive & send SMS remotely ● Forward SMS to email or messengers ● Plan SMS campaigns ● Set up auto-reply & auto dialer ● Manage multiple devices ● Use USSD commands remotely ● View full logs of calls and SMS ● Connect to the Internet via your own private VPN ● Use Wi-Fi mode

🎁 Your device unlocks 30 days of free full access. Try it first, decide later.

🌐 Your number. Your SIM. Worldwide.

⚡️ Google’s rebound shows Big Tech’s AI cycle is far from settled Google has quietly rebuilt market confidence after being prematurely written off in the early ChatGPT era and its stock momentum now reflects a much broader shift inside the company. 🖱 The launch of Gemini 3 and Google’s new TPU generation signal that the company is once again competitive at the frontier of AI performance. 🖱 Record search revenues show the core business is not only intact but accelerating despite years of disruption narratives. 🖱 Waymo’s expansion and improving utilization highlight Google’s lead in autonomous driving as rivals scale back. 🖱 Major Google Cloud AI deals with both startups and enterprises are pushing the division toward stronger margins. 🖱 YouTube’s dominance on Smart TV cements Google’s grip on the fastest-growing video consumption channel.

The message to markets: Google wasn’t falling behind, it was regrouping, and now the fundamentals are compounding again.

✔️Powered by Trade Watcher

📊 2025 cheat sheet: how much startups raise and how much equity they give up Here’s the quick snapshot of what “normal” looks like for U.S. fundraising in 2025, heavily skewed by SF pricing and big AI rounds. 🖱 Pre-seed: $500K–$2M rounds, typically 15–25% dilution. 🖱 Seed: $2M–$6M (AI teams often higher), around 18–22% dilution. 🖱 Series A: $8M–$20M, usually 15–20% sold. 🖱 Series B: $20M–$60M, about 10–15% dilution. 🖱 Series C: $40M–$150M, generally 5–10% sold as price per point jumps.

In San Francisco, these numbers hold; elsewhere in the U.S., expect lower valuations and slightly higher dilution for the same traction.

✔️Powered by Trade Watcher

⚠️ Court filings claim Meta buried internal evidence of social-media harms New U.S. court filings allege that Meta withheld internal research showing Facebook and Instagram could negatively affect teens’ mental health, a claim the company denies. 🖱 A 2020 internal study (“Project Mercury”) allegedly found that users who deactivated Facebook for a week reported lower depression, anxiety, loneliness, and social comparison. 🖱 Plaintiffs say Meta halted or buried the project due to concerns about “media narrative,” rather than methodological issues. 🖱 The lawsuit, filed by U.S. school districts, claims Meta prioritized teen engagement even when internal teams warned it increased exposure to harmful content. 🖱 According to filings, Meta told Congress it lacked clear evidence of harm despite allegedly having research pointing to meaningful negative effects on teenage girls. 🖱 Meta disputes the allegations, calling the study flawed and arguing that documents cited by plaintiffs were improperly filed.

These filings raise the stakes for Meta and could trigger a new wave of regulatory, legal, and public-policy pressure across the entire social-media industry.

✔️Powered by Trade Watcher

⚡️ Waymo accelerates its robotaxi rollout with 3 new city launches Waymo is expanding its commercial robotaxi network to Detroit, Las Vegas, and San Diego its fastest multi-city rollout yet as it pushes toward 1 million weekly rides by 2026. 🖱 The company will deploy mixed fleets of Jaguar I-Pace and Zeekr RT vehicles, starting with mapping and safety-operator phases before shifting to fully driverless service. 🖱 Detroit marks Waymo’s first major cold-weather market, leaning on years of winter-testing data to prove reliability in snow and low-visibility conditions. 🖱 Las Vegas and San Diego are strategic high-ridership targets, giving Waymo dense, high-traffic environments to scale autonomous operations. 🖱 The aggressive expansion reflects Waymo’s transition from R&D-heavy moonshot to a commercial mobility operator building a national footprint. 🖱 Leadership says the company is pacing toward 1 million autonomous trips per week by late 2026, a signal of confidence in both demand and operational maturity.

Waymo isn’t just adding cities, it’s laying the groundwork for a nationwide autonomous ride-hailing network.

✔️Powered by Trade Watcher

👤 Google hires ex–Boston Dynamics CTO to build “Android for robots” Google has brought on former Boston Dynamics CTO Aaron Sanders as VP of Hardware, openly signaling its plan to turn Gemini into a universal control layer for robots the same way Android became the standard for smartphones. 🖱 Google wants Gemini to function as open, modular robot firmware that developers worldwide can adopt across form factors echoing Android’s winning recipe of openness, scalability, and hardware flexibility. 🖱 With Sanders steering hardware strategy, Google is positioning itself to unify robotics software and hardware under one ecosystem instead of fragmented, proprietary stacks. 🖱 Competition is intense: Tesla and Figure push closed, vertically integrated systems, while Nvidia champions open tooling but lacks Google’s distribution and Gemini 3 performance edge. 🖱 Google’s last major VLA, Gemini Robotics 1.5, showed strong potential but leadership now hints at a much broader portfolio of robotics-tuned models. 🖱 If Google truly delivers a powerful open VLA adaptable to diverse robots, it could become the de facto platform layer for the robotics industry.

The race is simple: whoever controls the default OS for robots will control the robot economy.

✔️Powered by Trade Watcher

🔔 OpenAI braces for “rough vibes” as Google surges back A leaked internal memo from Sam Altman shows OpenAI warning employees about temporary economic headwinds and intensifying pressure from Google’s Gemini 3.0 rollout across the Google ecosystem. 🖱 Altman told staff to expect “rough vibes” as OpenAI faces slower revenue momentum and tougher competitive dynamics after Google’s strong model releases. 🖱 Google’s advantage is distribution: Gemini 3.0 is shipping into Search, Android, and Workspace, giving Google an ecosystem-wide moat that a standalone model provider can’t easily match. 🖱 OpenAI is shifting into a “wartime footing,” prioritizing product speed and efficiency as enterprise buyers test alternatives and model performance gaps narrow. 🖱 Internal concerns point to potential near-term economic softness, slower growth, pricing pressure, and rising compute costs as hyperscalers tighten partnerships. 🖱 The biggest risk is strategic: if Google’s ecosystem pull becomes the default interface for AI, OpenAI may need to rethink how it captures distribution, not just build better models.

The fight now is less about raw model quality and more about who owns the user’s daily workflow.

✔️Powered by Trade Watcher

⚡️ Investors pile into data centers despite uncertain exits A new PitchBook analysis shows investors are pouring record amounts into data-center development even as traditional exit paths remain weak or nonexistent. 🖱 Capital is shifting toward ground-up builds, with investors chasing AI-ready, high-density facilities instead of buying existing centers. 🖱 Demand from hyperscalers and AI labs is driving the surge long-term leases with Big Tech are now seen as the strongest de-risking mechanism. 🖱 Despite a frozen secondary market, developers are relying on structured debt (ABS/CMBS) and recapitalizations to extract liquidity. 🖱 Some funds expect a future exit wave in 2027–2028, once today’s projects stabilize, though buyers remain scarce for now. 🖱 Power constraints, regulatory bottlenecks, and bond-like yields on stabilized assets remain the biggest risks for investors piling in.

The bet is clear: compute demand will keep compounding and liquidity will eventually catch up.

✔️Powered by Trade Watcher

🔔 Genspark hits unicorn status with massive $275M series B raise Genspark, a Palo Alto–based AI startup, has raised $275 million in a Series B round, catapulting its valuation to $1.25 billion. 🖱 The funding was backed by major players: Emergence Capital Partners, SBI Investment, LG Technology Ventures, Pavilion Capital, and UpHonest Capital. 🖱 Genspark’s platform automates routine knowledge-work tasks from building slide decks and taking meeting notes to doing research and compiling reports. 🖱 Just five months after launching its AI agent suite, the company hit $50 million in annualized revenue. 🖱 The startup aims to compete with Microsoft 365 Copilot and Google Gemini by offering a unified workspace of specialized AI agents. 🖱 Co-founders include Eric Jing, a former Microsoft exec, and Wen Sang, an MIT PhD who previously built a SaaS business.

Genspark’s rapid growth and deep investor support mark it as a major contender in the enterprise AI agent space.

✔️Powered by Trade Watcher

🧮 A simple framework to allocate costs and calculate true P&L for each client or business line Entrepreneurs often struggle to understand whether a side service or specific client is actually profitable mostly because costs aren’t allocated correctly. A framework from the Looking for Leverage blog offers a clean way to break this down. 🖱 Direct costs are assigned 1:1, anything tied explicitly to a client or product (labor hours, materials, usage-based fees) should be allocated directly with no averaging. 🖱 Shared operational costs can be allocated by usage or complexity, for example: time spent, number of tickets handled, amount of compute used, or customer-specific workflows. 🖱 Fixed overhead (rent, admin, leadership) should be allocated using a stable driver, usually revenue share, headcount share, or workload share across business lines. 🖱 Variable costs should follow activity, allocate them based on units processed, volume handled, or the intensity of service required for each client. 🖱 Strategic costs (R&D, long-term bets) require judgment, either leave them unallocated or assign them only to the product lines that directly benefit.

This structure can help you see which clients and services truly make money and which ones quietly drain margin.

✔️Powered by Trade Watcher

📈 Startup M&A is heating up again, Carta reports a record 200 acquisitions in Q3 2025 Carta just logged the highest acquisition volume in its history, with 200 startup M&A deals in the third quarter, a clear sign the market has snapped back from the 2023 freeze. 🖱 Deal-making is alive again, with buyers ranging from other startups to Big Tech and private funds. 🖱 Early-stage dominates ~50% of all acquisitions are pre-seed/seed, and another 25–30% land at Series A. 🖱 Series B hits an all-time high, with 26 acquisitions in Q3 alone, real strategic buys, not acquihires. 🖱 Exits aren’t guaranteed paydays, many deals clear below liquidation prefs, leaving some shareholders with nothing. 🖱 AI leads the wave, accounting for at least 30% of all acquired startups this year, with the share rising each quarter.

With volumes surging and buyers returning, 2025 is shaping up to be one of the best moments in years for founders looking to sell.

✔️Powered by Trade Watcher

🔔 Adobe is buying Semrush in a $1.9B push into AI-first marketing Adobe has announced a $1.9 billion acquisition of Semrush, marking its biggest deal since the failed Figma takeover and signaling a major move to dominate AI-driven search and brand visibility. 🖱 Semrush brings both classic SEO and next-gen GEO (Generative Engine Optimization), giving Adobe a direct pipeline into how brands appear across LLMs, AI assistants, and traditional search. 🖱 Adobe will fold Semrush into Experience Cloud, integrating GEO data into AEM, Analytics, and its new Brand Concierge to power smarter AI-native marketing workflows. 🖱 The acquisition also reflects Adobe’s return to large-scale M&A, with the company betting that AI-mediated discovery will become as important as traditional search ranking. 🖱 With Semrush’s massive dataset, Adobe gains the ability to optimize how brands surface inside AI chatbots, answer engines, and agentic shopping flows, a channel growing 1,200% YoY.

Adobe isn’t just buying an SEO tool, it’s securing the discovery stack for an AI-first internet.

✔️Powered by Trade Watcher

💰 Anthropic lands a massive $45B deal with Microsoft and Nvidia Anthropic has signed a record-scale investment and infrastructure agreement with Microsoft and Nvidia worth $45 billion, instantly pushing the company’s valuation to $350B up from $183B just two months ago. 🖱 Anthropic will commit $30B to buying Azure compute roughly 1GW of capacity locking in long-term infrastructure for Claude’s next-generation models. 🖱 Nvidia is investing $10B and Microsoft $5B, giving both deeper strategic ties to one of the fastest-growing AI labs in the world. 🖱 Nvidia and Anthropic will co-develop new optimization methods for both models and chips, aiming to boost performance and efficiency across the stack. 🖱 Jensen Huang called the partnership a “dream come true,” underscoring Nvidia’s plan to shape the architecture behind frontier AI systems.

With this structure, Anthropic isn’t just raising capital, it’s locking in the full-stack infrastructure to compete at the absolute top of the AI race.

✔️Powered by Trade Watcher