Trending Thursday #34

In partnership with

Hi !

An open-source AI agent went viral and sparked a Mac Mini buying frenzy; earnings season hit like a freight train; Tesla is killing two iconic models to build humanoid robots; and Amazon is flattening its org chart by 16,000 people.

Here's what's been trending:

• Clawdbot, an open-source AI agent, went viral over the weekend, sparking a Mac Mini buying frenzy as founders discovered they could run an autonomous "Chief of Staff" from a $600 computer

• Meta and Microsoft dropped blockbuster earnings: $60B and $81B quarters respectively, while SK Hynix overtook Samsung in annual profit for the first time ever, and both chipmakers warned the memory crunch extends to 2027

• Tesla is killing the Model S and X to convert its Fremont factory into an Optimus robot manufacturing hub, and quietly invested $2B into Elon Musk's xAI

• Amazon plans to cut ~16,000 roles as part of an organizational flattening effort, continuing the industry-wide squeeze on middle management

As we continue to build at Context Ventures, we've launched Startup Intros with a simple mission: to help early-stage founders find the right investors, faster and smarter.

🤖 Clawdbot Goes Viral: The Open-Source AI Agent That Sparked a Mac Mini Buying Frenzy

Over the MLKJ weekend, an open-source project called Clawdbot (or Moltbot) went from obscurity to the hottest thing on X. The concept: run an autonomous AI agent on a $600 Mac Mini that acts as your personal Chief of Staff: managing email, calendar, building code, writing content, and monitoring your business 24/7.

🔥 From Zero to Viral in 48 Hours

Created by Peter Steinberger (@steipete), Clawdbot (rebranded to Moltbot) launched in early January but exploded January 24-26 when founders started sharing their setups on X. Jason Calacanis dedicated an entire episode of the All-In / Twist podcast to it, calling it "what Siri was supposed to be."

The Result? A literal Mac Mini buying frenzy. Founders were posting receipts for $600 Mac Minis and $12,000 Mac Studios, treating them as if they were hiring a full-time employee rather than buying a computer.

🧠 Why It Matters: Personal AI Agents Are Here

Unlike Claude Co-work (which "can basically edit a spreadsheet," as one guest put it), Clawdbot connects to your entire digital life: Gmail, calendar, code repos, messaging apps, databases, and APIs. It has infinite memory, learns from every interaction, and runs autonomously while you sleep.

The use cases demoed on the Twist podcast were wild:

• A tea shop in Israel automated procurement, HR, payroll, logistics, and inventory, saving an estimated $40-50K/year

• A YouTube creator's bot autonomously detected a trending topic, built a new feature for his SaaS product, and submitted a pull request, all overnight

• A venture fund described running deal flow, portfolio monitoring, and anti-portfolio analysis from a WhatsApp window

💡 The Business Opportunity (and the Risk)

The emerging ecosystem looks like the early app store: a skills marketplace is forming, with community-built tools for X search, competitive research, content creation, and more.

But the security concerns are real. These agents have admin access to your entire digital life, and The Register reports that security researchers warn that even properly configured setups still carry data-exposure risks. Jason Calacanis called it "as dangerous as it gets." Alice (formerly ActiveFence), which already protects 3B people on social networks, is positioning itself as the security layer for AI agent ecosystems.

Translation: We just crossed a threshold. Personal AI agents aren't a research demo anymore; they're running small businesses and writing code overnight. The question isn't whether this changes how startups operate, but how fast. If you're a founder and you're not experimenting with this, you're already behind.

💰 Big Tech's Earnings Blitz: $141B in Combined Revenue & Record AI Spend

Meta and Microsoft reported on the same day and painted the same picture: revenue is surging, profits are enormous, and they're both spending ungodly amounts of money on AI infrastructure. Meanwhile, the companies making the chips that fuel this boom are warning: the supply crunch isn't going anywhere.

📈 Meta: $60B Quarter, $115-135B in 2026 Capex

Meta crushed estimates with $59.9B in Q4 revenue (up 24% YoY) and $22.8B in net income. The stock popped 7%+ after hours.

But the real headline: Meta plans to spend $115-135B in 2026 on AI infrastructure, up from $72.2B in 2025 and blowing past analyst estimates of $110.6B. Zuckerberg teased "agentic shopping tools" and plans to ship new AI models throughout 2026. Reality Labs continues to bleed: $6B in Q4 operating losses, $75B+ since late 2020, but nobody seems to care anymore.

🖥️ Microsoft: $81B Quarter, Azure Up 39%

Microsoft posted $81.3B in Q2 revenue (up 17% YoY), with $30.9B in net income (up 23%). Azure grew 39%, barely beating estimates, and LinkedIn crossed $5B in quarterly revenue for the first time ($20B+ annual run rate). Windows 11 also hit 1B users, up 45% YoY as Windows 10 end-of-life drove mass upgrades.

The Catch? Microsoft spent $37.5B in capex this quarter, above estimates, and the stock dropped 4% after hours. Wall Street wants to see the AI spending turn into margin expansion, not just revenue growth.

🧩 The Memory Crunch: SK Hynix Dethrones Samsung, Supply Squeezed to 2027

Here's the story behind the story: all that AI capex has to go somewhere, and it's going into chips, specifically memory.

SK Hynix overtook Samsung in annual operating profit for the first time ever, posting a record ₩47.2T ($33B) vs. Samsung's ₩43.6T ($30.5B). SK Hynix's dominance in high-bandwidth memory (HBM), the specialized chips inside Nvidia's AI processors, has been the difference maker. The company holds a commanding 57% revenue share in HBM, vs. Samsung's 22%.

Samsung, for its part, still posted 200%+ operating profit growth in Q4, driven by memory chip price hikes and rising HBM demand. But the real warning came from both companies' executives: the global memory supply shortage will persist until 2027 due to extreme AI-driven demand and limited cleanroom capacity.

Translation: Big Tech is printing money and spending it as fast as they can on AI. But the supply chain can't keep up, and the companies making the actual chips that power all this AI infrastructure say they're years away from meeting demand.

When Is the Right Time to Retire?

Determining when to retire is one of life’s biggest decisions, and the right time depends on your personal vision for the future. Have you considered what your retirement will look like, how long your money needs to last and what your expenses will be? Answering these questions is the first step toward building a successful retirement plan.

Our guide, When to Retire: A Quick and Easy Planning Guide, walks you through these critical steps. Learn ways to define your goals and align your investment strategy to meet them. If you have $1,000,000 or more saved, download your free guide to start planning for the retirement you’ve worked for.

Find Your Retirement Timeline

⚡ Tesla Kills the Model S and X: Fremont Becomes an Optimus Robot Factory

In the clearest signal yet that Tesla sees itself as a robotics company that happens to sell cars, Elon Musk announced on the Q4 earnings call that Tesla is discontinuing its Model S sedan and Model X SUV, and converting the Fremont, California factory lines into a 1 M-unit-per-year Optimus robot production hub.

🏭 "An Honorable Discharge"

"It's time to basically bring the Model S and X programs to an end with an honorable discharge," Musk said on the call. "If you're interested in buying a Model S and X, now would be the time to order it."

The two vehicles are Tesla's oldest models, the S launched in 2012, the X in 2015, and together accounted for just 3% of Tesla's 1.59M deliveries last year. The Model 3 and Y, starting at $37K and $40K respectively, now carry the entire vehicle business.

Production winds down next quarter. Musk says headcount at Fremont will actually increase, not decrease, as the facility pivots to a "completely new supply chain" for Optimus, with nothing carried over from the existing auto lines.

🤖 The Optimus Bet

Tesla plans to unveil the third-generation Optimus this quarter; its "first design meant for mass production." Musk has said Tesla will sell Optimus to the public starting in 2027, calling it "the biggest product of all time."

The earnings report described Tesla's tumultuous 2025 as a "transition from a hardware-centric business to a physical AI company." Revenue fell 3% YoY, Tesla's first annual decline ever, while automotive revenues dropped 11%. The Cybertruck saw a 48% sales decline. But EPS of $0.50 beat Wall Street's $0.45 estimate, and the stock rose in after-hours trading.

💸 Oh, and Tesla Just Put $2B Into xAI

Buried in the earnings report: Tesla agreed to invest $2B in xAI as part of its Series E round, despite a November 2025 shareholder vote that failed to approve the deal. The investment is in preferred shares, suggesting Musk pushed it through as a board decision rather than revisiting the shareholder question.

Translation: Tesla is betting the entire company on a future where humanoid robots and AI matter more than cars. The Model S and X were already rounding errors in the vehicle business, but killing them to build robots is the kind of move you only make when you've fully committed to the pivot. Whether Optimus delivers or not, the old Tesla, the luxury EV maker, is officially dead.

🪓 Amazon Cuts 16K: The Org Flattening Continues

Amazon confirmed plans to cut approximately 16,000 corporate roles, its biggest reduction since 2023, as part of CEO Andy Jassy's ongoing push to flatten the management structure and increase the ratio of individual contributors to managers.

The cuts span multiple divisions and aren't concentrated in any single business unit, Reuters reports. Unlike the 2023 layoffs (which hit AWS and devices hardest), these are structural: removing entire layers of middle management in pursuit of what Jassy calls "builder culture" over "process culture."

Amazon isn't alone. Google, Meta, and Microsoft have all reduced management layers in the past 18 months. The thesis is simple: if AI can handle coordination, reporting, and project tracking, you need fewer people whose job is to do so.

Translation: Middle management is being squeezed from both sides: AI from below (automating their tasks) and executive mandate from above (flattening org charts). If your role is primarily "information routing" between people who do the work, the clock is ticking.

🛠️ Product Quick Hits: What Shipped This Week

🧠 Kimi K2.5 Drops as "Most Powerful Open-Source Model": Moonshot AI's native multimodal model features a self-directed agent swarm of up to 100 parallel sub-agents across 1,500 tool calls, with state-of-the-art coding and vision. Available now via API and Kimi Code.

🔍 Google Launches Project Genie: Users can create interactive worlds using Genie 3, Nano Banana Pro, and Gemini, available to Google AI Ultra subscribers in the US. Generations are limited to 60 seconds at 720p, but it's a first glimpse of DeepMind's world model research.

🤖 Airtable Unveils Superagent: Its first standalone product in 13 years deploys coordinating agents that plan work, deploy specialists in parallel, and synthesize output into finished deliverables. TechCrunch reports the company sees it as the start of a comeback.

🧪 Arcee Releases Trinity Large 400B: A US-based startup's open-weight model that it says rivals Meta's Llama 4 Maverick on some benchmarks, built from scratch with a fraction of Big Tech's resources.

🇨🇳 ByteDance and Alibaba Racing to Drop Flagship Models: Both plan to release around mid-February, including ByteDance's Doubao 2.0. The Information reports it's shaping up to be the most competitive model release window in Chinese AI history.

🔎 DeepSeek Eyes Search and Agents: Bloomberg reports job postings reveal plans for a multilingual AI search engine and expanded agent capabilities, a direct challenge to Google and OpenAI.

🎵 Deezer: 39% of Daily Uploads Are Now AI-Generated: Makes its AI music detection tool commercially available after flagging over 13.4M AI tracks in 2025. Daily AI track deliveries averaged 60K in January, up from 10% of uploads a year ago.

🖥️ Nvidia Launches Native GeForce NOW for Linux: Starting with Debian-based Ubuntu, after releasing a Steam Deck app in 2025. The dedicated native app works well in early hands-on.

👔 Personnel Quick Hits: Who's Moving This Week

📸 Sebastiaan de With Joins Apple: The Halide co-founder joins Apple's Human Interface Design team. "So excited to work with the very best team in the world on my favorite products," he said.

📊 SAP Stock Falls 14%+: After Q4 results disappointed on revenue, despite forecasting 23%+ cloud growth in 2026.

📱 iPhone 16 Was World's Best-Selling Smartphone in 2025: Apple took 7 of the top 10 spots, Samsung took 3.

🚫 Nothing CEO Carl Pei: Says Phone 4 won't launch in 2026, "not just going to churn out a new flagship every year for the sake of it."

🌟 Editor's Note

At Startup Intros, our mission is to bring the latest founder-investor news straight to your inbox, keeping you ahead in the fast-paced world of Silicon Valley.

💭 Parting Thoughts

This week crystallized something that's been building for months: AI isn't just reshaping software; it's reshaping the physical world.

Meta and Microsoft are spending a combined $250B on AI infrastructure this year. Tesla is literally shutting down car production lines to build humanoid robots. Amazon is cutting 16,000 managers because AI can route information faster. SK Hynix and Samsung are warning that the memory chip shortage won't ease until 2027. Meanwhile, a solo developer in Austria built an open-source tool that lets any founder run a personal AI agent for the cost of a coffee machine.

The old economy, where you needed headcount to scale, management layers to coordinate, and massive capex to compete,

is being rewritten in real time. The new economy runs on silicon, models, and agents. Some companies are spending billions to get there. Others are spending $600 on a Mac Mini.

The future doesn't belong to whoever spends the most on GPUs. It belongs to whoever figures out how to use them first.

Till next time!

P.S. Raised this week, and we missed you? Want to be featured? Have tips or funding questions? Reply or DM us as we’re here to help.

Startup Intros Trending Thursday: Your trusted source for founder-investor insights, delivered with clarity and focus.