🧠 Netflix’s rise from DVD startup to global streaming powerhouse

Netflix began in the late 1990s as a DVD-by-mail rental service, a business that looked odd compared to dominant players like Blockbuster, which had thousands of physical stores and seemed untouchable. Netflix avoided brick-and-mortar costs, focused on online ordering, eliminated late fees, and introduced a subscription model with free trial periods, moves that were unconventional at the time but made the service more user-friendly.

🖱 Disruption over imitation: Netflix’s early model didn’t chase the latest movies or mimic Blockbuster’s rental strategy. Instead, it reimagined convenience and pricing for customers, which helped it build a loyal base even as traditional video rental revenues collapsed.

🖱 Blockbuster missed its pivot: Around 2000, Netflix offered to partner with Blockbuster, but Blockbuster’s leadership rejected the opportunity, underestimating online DVD subscriptions. Blockbuster later launched initiatives to counter Netflix, but those moves blew through cash and failed to stop the disruption, ultimately leading to bankruptcy.

🖱 From mail to streaming: Netflix successfully transitioned from mailing discs to internet streaming, anticipating where customer demand and technology were headed. That pivot helped it become a dominant global OTT platform with original content and tens of millions of subscribers.

🖱 Disruption mindset: Netflix’s journey illustrates how a startup can upend an established industry by challenging entrenched business models and leveraging new delivery methods, classic disruption theory in action.

Netflix didn’t just grow, it reshaped the market by betting early on online convenience and subscription pricing. Blockbuster’s failure to adapt highlights how innovation can overturn dominant players when customer habits and technology shift.

✔️Powered by Trade Watcher

64GB of RAM in the USA cost more than a semi-automatic AR-15 rifle

Amazing timeline…

✔️Powered by Trade Watcher

📉 OpenAI loses half of its enterprise LLM market share in two years

Menlo’s latest annual AI market report points to a sharp power shift in the enterprise LLM API market, as OpenAI’s early dominance continues to erode.

🖱 Since 2023, OpenAI’s share of the enterprise LLM API market has fallen from roughly 50% to 27%, a dramatic decline over just two years.

🖱 Anthropic is now the clear leader, controlling about 40% of the market nearly triple its share in 2023 driven by strong enterprise adoption of Claude.

🖱 Google sits in third place with ~21%, benefiting from tight integration with its cloud, data, and enterprise distribution channels.

🖱 The market is increasingly multi-vendor, as enterprises hedge risk, avoid lock-in, and route workloads to different models based on cost, safety, or context length.

🖱 OpenAI remains a major player, but no longer the default choice, as competition on reliability, governance, and enterprise fit intensifies.

The LLM API market has shifted from winner-takes-all to fragmented infrastructure and OpenAI is now competing on equal footing, not from a monopoly position.

✔️Powered by Trade Watcher

🇨🇳 AI exam for Apple in China

China is formalizing one of the world’s strictest AI approval regimes and Apple Intelligence is now directly in its scope.

🖱 According to The Wall Street Journal, every chatbot entering the Chinese market must pass a 2,000-question government exam designed to surface responses on prohibited topics.

🖱 To pass, a model must refuse to answer at least 95% of politically or socially sensitive prompts. The question bank is updated monthly, forcing continuous re-certification.

🖱 Enforcement is massive: in just three months, regulators removed 960,000 units of “illegal” AI content and shut down 3,500 AI products.

🖱 AI is now officially framed as a national security risk, on par with earthquakes and epidemics. Xi Jinping has called it an “unprecedented risk,” while a senior aide likened unregulated AI to “driving on a highway without brakes.”

🖱 This directly impacts Apple Intelligence. Outside China, Siri escalates complex queries to ChatGPT, and some workloads rely on Gemini on Apple’s servers but in China, Apple is effectively forced to use a local partner.

🖱 That partner is Alibaba, with Apple relying on the Qwen3 model to comply with local regulations.

🖱 A notable side effect: researchers at the Carnegie Endowment find Chinese models score better on some safety metrics (less violence, pornography, and self-harm), but are easier to jailbreak using English prompts.

China isn’t just regulating AI, it’s stress-testing whether global AI platforms can exist as fully localized, state-compatible systems without breaking their core product experience.

✔️Powered by Trade Watcher

🚕 Closing a door in a Waymo robotaxi isn’t free

Waymo’s driverless taxi rollout across the US hides a growing operational cost: human intervention for edge cases autonomy still can’t handle.

🖱 According to The Washington Post, Waymo relies on an on-demand network of human “rescue workers” to assist robotaxis when they become immobilized.

🖱 One of the most common failures is mundane: if a passenger exits without fully closing the door, the vehicle refuses to move and must be manually assisted.

🖱 These helpers are dispatched via Honk, a roadside-assistance marketplace founded in 2014 that operates like Uber for car help and has raised $33 million.

🖱 Rescue workers report being paid $22–24 just to close a door on a stalled Waymo vehicle.

🖱 More complex interventions pay more: $60–80 to tow a robotaxi that ran out of battery or failed to reach a charging station.

🖱 The process is inefficient at times workers say they’re often given imprecise locations, spending up to an hour searching city streets for the vehicle.

Waymo removed the driver, not the humans autonomy still depends on a hidden workforce handling the last, expensive edge cases.

✔️Powered by Trade Watcher

📈 Elon Musk predicts double-digit GDP growth and triple-digit AI upside

Elon Musk says global GDP could grow 10%+ in the next 12–18 months, and that if applied AI becomes a proxy for economic output, 100%+ annual growth could be possible within five years.

🖱 Musk argues applied AI is no longer theoretical, it is directly improving productivity in software, manufacturing, logistics, and science.

🖱 If AI output scales faster than labor or capital constraints, traditional GDP models may understate real economic expansion.

🖱 Musk frames AI as a general productivity multiplier, similar to electricity or the internet but with faster compounding.

🖱 Critics note that GDP growth is constrained by energy, regulation, and real-world deployment speed, even if AI capability advances rapidly.

🖱 In response, a public challenge emerged: name at least one Musk prediction or promise that actually came true.

🖱 Common answers include reusable rockets (SpaceX), private human spaceflight, EVs outperforming ICE cars, and global satellite internet via Starlink.

Musk’s forecasts remain extreme but history shows he has repeatedly been early rather than wrong, which is why markets still pay attention.

✔️Powered by Trade Watcher

📊 India startup funding hits $11B in 2025 as investors grow more selective

India’s startup ecosystem raised nearly $11 billion in 2025, but the pace and structure of deals shifted significantly compared with previous years signaling a maturing market and more cautious investor behavior.

🖱 Total funding slid modestly relative to 2024, while the number of funding rounds fell sharply (~39%), underscoring greater selectivity among investors.

🖱 Seed-stage funding dropped around 30%, reflecting pullback on early, high-risk bets, while late-stage funding also cooled amid stricter scrutiny on scale and profitability.

🖱 Early-stage funding grew (~7%), indicating investor interest in startups with strong product–market fit, revenue visibility, and unit economics.

🖱 AI funding in India grew modestly (~4%) and was mostly directed at application-led startups, contrasting sharply with the U.S. AI boom where funding surged far more and was dominated by late-stage rounds.

🖱 Activity became more concentrated: a smaller group of lead investors drove most rounds, and domestic capital played a larger role as some global investors pulled back.

🖱 Government initiatives including a $1.15 B Fund of Funds and a ₹1 trillion R&D & innovation scheme helped catalyze funding, especially for deep tech sectors like energy transition, quantum, robotics, biotech, and space.

India’s startup funding in 2025 remained robust in absolute terms, but the ecosystem is shifting toward more disciplined, quality-focused capital deployment, with emphasis on early-stage performance, pragmatic AI bets, and deeper involvement from domestic investors and public sector support.

✔️Powered by Trade Watcher

📉 Inflated startup valuations are setting up painful corrections and AI looks next

A Crunchbase analysis warns that hype-driven valuations, not fundamentals, are once again shaping startup funding with AI showing many of the same warning signs seen in past bubbles.

🖱 In previous cycles, startups raised at extreme revenue multiples without solid unit economics, leaving them vulnerable once markets tightened.

🖱 When follow-on rounds failed to materialize at higher prices, many companies were forced into down rounds, shutdowns, or fire-sale acquisitions.

🖱 Inflated valuations distorted decision-making, encouraging aggressive burn rates under the assumption that capital would always be available.

🖱 The article argues that early-stage AI startups are now raising on narrative and assumed future dominance, often before product–market fit or durable revenue.

🖱 As hype fades, only AI companies with strong margins, real customers, and disciplined spending are likely to survive the correction.

🖱 Founders are urged to prioritize runway, fundamentals, and realistic pricing over headline valuations.

The lesson from past cycles is clear: valuation is not success and in AI, today’s inflated prices could become tomorrow’s biggest liability.

✔️Powered by Trade Watcher

🔔 OpenAI hires a Head of Preparedness as AI risks shift from theory to reality

OpenAI is introducing a new senior role, Head of Preparedness, focused not on mitigating model flaws, but on anticipating what happens after powerful AI systems are released into the world.

🖱 Unlike traditional safety roles, this position is about modeling second- and third-order effects that can’t be reliably tested in labs or benchmarks.

🖱 The mandate includes identifying non-obvious danger zones: long-term societal risks, emergent misuse patterns, and failure modes that only appear at scale.

🖱 Sam Altman says 2025 brought the first serious warning signal: measurable negative impacts on mental health, including reports linking ChatGPT use to suicides.

🖱 At the same time, strong coding capabilities elevate risks in cybercrime, automation of attacks, and lowering the barrier for sophisticated exploitation.

🖱 The role exists because model capabilities are now high enough that downstream consequences matter as much as model alignment itself.

🖱 Compensation is up to $555k including bonuses, serious money, though modest by frontier-AI executive standards.

AI labs are beginning to treat post-deployment societal impact as a core strategic risk, the question is whether this becomes real governance or polished reassurance.

✔️Powered by Trade Watcher

📉 Nvidia finalizes $5B strategic stake in Intel, a major shakeup in the chip world

Nvidia has completed a $5 billion purchase of Intel shares under a private placement agreement first announced in September 2025, buying about 214.7 million shares at $23.28 each.

🖱 The deal was executed after U.S. antitrust clearance and is viewed as a significant financial lifeline for Intel, which has struggled with expensive production expansions and recent strategic setbacks.

🖱 Nvidia’s shares dipped modestly in pre-market trading after the disclosure, while Intel’s stock was relatively unchanged.

🖱 The investment gives Nvidia a notable minority position in Intel and marks one of the most unusual equity moves between two major U.S. chipmakers in years.

🖱 Analysts see it as a vote of confidence in Intel’s turnaround and a way for Intel to raise cash without debt, even as it expands its foundry and AI-focused chip efforts.

🖱 Beyond finance, the agreement signals closer strategic ties between the companies going into 2026 especially in AI infrastructure and advanced processor development.

A $5 billion vote of confidence from Nvidia could reshape competitive dynamics in AI chips, bolster Intel’s balance sheet, and deepen collaboration at a pivotal moment for the semiconductor industry.

✔️Powered by Trade Watcher

⚠️Nvidia quietly exits the cloud business and steps back from competing with AWS

Nvidia has effectively abandoned plans to run its own cloud service, folding DGX Cloud back into an internal-only operation and redeploying most of the team into R&D.

🖱 DGX Cloud technically still exists as an org chart item, but it now serves only Nvidia’s internal compute needs, not external customers.

🖱 Demand for the service was minimal, which appears to be the real reason for the shutdown, the market showed little interest in Nvidia as a cloud provider.

🖱 The move signals a clear retreat from competing with hyperscalers, especially AWS, rather than a strategic pause.

🖱 Notably, AWS previously refused to participate in the DGX Cloud program, drawing a hard line against Nvidia encroaching on its core business.

🖱 By shutting DGX Cloud down, Jensen Huang publicly reassures Nvidia’s biggest customers that it will remain a supplier, not a rival.

🖱 Resources shifting to R&D reinforces Nvidia’s core strategy: sell picks and shovels for AI, not run the gold mine itself.

Nvidia chose ecosystem dominance over platform ambition and decided that angering AWS, Azure, and Google was far riskier than abandoning its own cloud dreams.

✔️Powered by Trade Watcher

🔔Meta buys Manus to jump-start its AI agent strategy

Meta has acquired Manus, the Singapore-based AI startup, in a deal reportedly valued at ~US$2B, betting on agentic AI that can do work, not just answer questions.

🖱 Manus builds general-purpose AI agents that autonomously execute tasks, research, document analysis, hiring workflows, coding, and business automation.

🖱 Unlike most AI labs, Manus already had real revenue: over US$100M ARR within months of launch, proving demand beyond demos.

🖱 The startup moved its HQ to Singapore in 2025, distancing itself from China exposure and making it acquisition-friendly for US Big Tech.

🖱 Meta plans to keep Manus as a standalone product while integrating its agent tech into Meta AI, WhatsApp, Instagram, and enterprise tools.

🖱 Strategically, this helps Meta close the gap with OpenAI, Microsoft, and Google, who are all racing to own the “AI that actually works for you” layer.

🖱 The deal signals Meta’s shift from pure model competition to applied, monetizable AI systems that justify its massive AI capex.

🖱 Buying Manus is faster than building: Meta gets product-market fit, paying customers, and a team that already cracked agent execution.

Meta isn’t just chasing smarter models anymore, it’s buying its way into the operational AI era, where agents replace workflows, not chats.

✔️Powered by Trade Watcher

✨ Another New Year’s promise from Elon Musk: xAI will surpass everyone in compute

Elon Musk says xAI will have more computing power than all other companies combined within 5 years an aggressive claim even by his standards.

🖱 The bet hinges on Colossus 2, xAI’s massive data center under construction in Memphis.

🖱 Colossus 2 already exceeds 400 MW of power capacity, with an ambitious target of 2 GW.

🖱 If completed as planned, it would become the world’s first gigawatt-scale data center, far beyond today’s hyperscaler norms.

🖱 To fuel this expansion, xAI is raising $20B to buy additional GPUs at unprecedented scale.

🖱 The strategy is brute force: overwhelm competitors not with better algorithms, but with sheer compute dominance.

🖱 The risk is equally massive, power availability, grid constraints, GPU supply, and returns on capital at this scale are all unproven.

Musk is betting that in AI, the company with the most electricity wins and he’s willing to build power plants disguised as data centers to prove it.

✔️Powered by Trade Watcher

⚠️ Bad New Year’s news: ads are coming to ChatGPT and soon

It’s not exactly the update anyone wanted to start the year with, but here we are: ChatGPT will definitely have ads, and the launch looks imminent.

🖱 The report comes from The Information and when they cite insiders, they’re usually right.

🖱 Earlier, OpenAI delayed ads because of a declared “red code”, triggered by intense competitive pressure from Google.

🖱 That emergency mode is now over, and ads are back at the top of the priority list.

🖱 Advertising layouts for multiple formats are already prepared, which strongly suggests testing and rollout are close.

🖱 From OpenAI’s perspective, this was inevitable: ads inside ChatGPT are a gold mine, especially given the company’s massive burn rate.

🖱 ChatGPT already has ~900 million users, and projections point to 2.6 billion users by 2030.

🖱 Add to that: ChatGPT knows what users like, what they ask, what tools they use, and how they think, perfect signal for hyper-targeted ads.

🖱 Taken together, this could become the largest advertising platform ever built, hiding inside a “helpful assistant.”

🖱 The small consolation: ads are planned only for the free tier, at least for now.

It’s disappointing, sure, but this was always a question of when, not if.

✔️Powered by Trade Watcher

Yesterday was Warren Buffett’s final day as CEO of Berkshire Hathaway.

Here’s his first interview from 63 years ago.

Thank you, Warren Buffett, the Oracle of Omaha.

✔️Subscribe on @venture

Peter Thiel predicting Bitcoin in 1999.

✔️Subscribe on @venture

Former Google CEO Eric Schimdt drops a chilling warning on AI's future.

✔️Subscribe on @venture

This is why it's impossible to tax billionaires.

✔️Subscribe on @venture

Growth is no longer guaranteed at Tesla

The company wrapped up Q4 2025 with results that undercut expectations, confirming two consecutive years of shrinking annual volume, a major shift from Tesla’s long-held growth narrative.

🖱 End-of-year shipments came in well below forecasts, down in the mid-teens versus the prior year.

🖱 Vehicle output also moved lower, pointing to structural softness rather than a one-off demand pause.

🖱 Total 2025 volumes fell by roughly 8–9%, locking in a second annual contraction.

🖱 BYD now dominates global EV scale, increasing the distance between itself and Tesla.

🖱 Higher-priced and experimental vehicles (S, X, Cybertruck) saw the sharpest drop, exposing weak pull outside the core lineup.

🖱 More cars left factories than reached customers in Q4, signaling inventory pressure heading into 2026.

🖱 Incentive pull-forwards and intensifying competition blunted the impact of price cuts.

Tesla’s valuation still leans on autonomy and AI optionality but the numbers show the auto business is no longer compounding, and the execution gap is becoming harder to ignore.

✔️Subscribe on @venture

🗣️ Naval Ravikant on Elon Musk: “The great entrepreneurs are willing to start over”

Naval argues that pride is the enemy of learning:

“When I look at my friends and colleagues, the ones who are still stuck in the past and have grown the least are the ones who were the proudest because they feel they already had the answers and don’t want to correct themselves publicly… Pride prevents you from saying, ‘I’m wrong. The problem with pride, Naval explains, is that it prevents you from saying “I’m wrong. When you don’t admit that you were wrong, you get stuck in it and you get trapped in a local maxima, as opposed to going back down and climbing up the mountain again… The great artists always have this ability to start over whether it’s Paul Simon, Madonna, or U2.”

And Naval argues that the best entrepreneurs are always willing to start over too:

“I’m always struck by the Elon Musk story where he did PayPal… And he said something along the lines of: ‘I made $200 million from the sale of PayPal. I put $100 million into SpaceX, $80 million into Tesla, $20 million into Solar City, and I had to borrow money for rent.’ This guy is a perennial risk taker. He’s always willing to start over. He doesn’t have any pride about being seen as successful or being seen as a failure. He’s willing to put it all on the line to back himself again. But the key thing is he’s always willing to start over… It’s a willingness to look like a fool and a willingness to start over.”

He continues:

“A lot of people just don’t have that. They become successful or rich or famous and that’s it. They’re stuck. They don’t want to go back to zero, but creating anything great requires going from zero to one, and that means you go back to zero. And that’s a really painful and hard thing to do.”

✔️Subscribe on @venture