Sakura | Emerging Tech

7 technologies that will change everything by 2030 — and most people aren't paying attention yet. A thread on what's coming, who's building it, and why it matters. 🧵👇 1/ AI Agents Not just chatbots — autonomous systems that complete tasks for you. OpenAI, Anthropic, and Google are racing to build them. In 2026, you'll have a personal AI agent handling your schedule, research, and communications. 2/ Quantum Computing IBM, Google, and China are all pushing past 1,000 qubits. Quantum will break current encryption, revolutionize drug discovery, and solve problems classical computers can't. The race is real. The timeline is accelerating. 3/ Brain-Computer Interfaces Neuralink is just the start. Non-invasive BCIs are coming that let you control devices with thought. Medical applications first (paralysis, communication). Consumer applications by 2030. 4/ Fusion Energy 2023 was the year fusion net energy gain was proven. Commonwealth Fusion Systems and Helion are racing to commercialize. If fusion works at scale, it changes everything about energy geopolitics. 5/ Humanoid Robotics Figure, Tesla Optimus, and Boston Dynamics are converging on general-purpose humanoids. By 2028, you'll see them in warehouses and factories. By 2035, they could be in homes. Labor markets will never be the same. 6/ Gene Editing (CRISPR 2.0) CRISPR cured sickle cell in 2023. Next: targeted cancer therapies, hereditary disease prevention, and potentially longevity treatments. The ethical debates will define the next decade. 7/ Autonomous Everything Self-driving cars are just the beginning. Autonomous ships, drones, delivery bots, farming equipment, and mining vehicles are all coming online. The physical world is being automated. If you want to track these technologies — and the ones coming next — follow me. I post threads, analysis, and early signals on emerging tech every week.

Who replaces Jensen? The names reported so far. Follow for the chip story as it breaks. Kress, Nvidia's CFO: some directors favor her as interim CEO (Semafor) Srouji, Apple's hardware chief: among those considered (Semafor) Tim Cook: reportedly floated Bloomberg says an outside firm is running the search. My bet: Kress as interim, and the permanent pick is the real fight.

Chips ripped Monday. Nvidia fell alone. Follow for the fallout. The chip index jumped 6.85%, its best day since July 30 (MarketWatch). AMD +13.3%, Broadcom +8.8%. Only 1 of its 30 stocks fell: Nvidia, -4.27% on the Jensen Huang spy story, on 455M shares, ~3.7x normal volume (Motley Fool). The market is pricing the CEO, not the chips.

Nvidia's Huang scandal is now a chip-export story. Follow for the chip-war angle. Per Bloomberg, the board opened an independent probe and NVDA shares dropped. Per Reuters, Trump called Huang "honorable." Kobeissi tallied Nasdaq 100 futures swinging -$250B then +$300B in 7 minutes. What to watch next: H200 export approvals for China.

NASA just wrote a $10B Moon shopping list. CLPS 2 is a 10-year contract with a $10 billion ceiling. Eight companies — including SpaceX and Blue Origin — can bid to fly NASA cargo to the lunar surface. Separately, Blue Origin got a $94 million base award, plus a $140 million performance option, to land PROMISE: a nuclear-powered South Pole rover that JPL is building from Mars rover spare parts, targeting late 2028. Numbers from NASA's November 7 awards release: nasa.gov/news-release/nasa-a…

Export clearance is the chip story. FT + NYT: US intel flagged a China MSS link around Jensen Huang’s circle. When GPU gates sit with one CEO, allocation risk gets priced in—not the gossip. Sources: Financial Times, New York Times.

Anthropic's IPO just got a public valuation reset WSJ: major buyers pushing the $1.5T plan toward closer to $1T — some even float delay or a smaller deal. Asia opened into that scare (CNBC): SK Hynix −8%+, Samsung −7%, TSMC −9.6%, SoftBank −11.6%. Follow @SakuraTech21 for the AI buildout map, not the panic headline.

SK hynix just wiped 9% in one session. KOSPI −5.28% to 5,383.80. Samsung −7.57%. Asia's AI memory names led the close as Micron paused its $4.5B Taiwan Fab 11E over a labor spat — that DRAM/HBM plant was due 2028. Yonhap for the close; Bloomberg via Reuters for Micron. Follow for the chip-chain numbers.

Texas just elected a data center skeptic. I track the AI power fight daily, follow along. Fox News projects Gina Hinojosa beats Greg Abbott. She's promised a moratorium on new data center approvals until lawmakers pass protections. Why it matters: Texas holds ~25% of US data center capacity due online by 2030, and Jefferies says 13.3GW of announced projects could be at risk. Chips aren't the bottleneck here. Permits and grid hookups are.

SpaceX is now almost Anthropic's biggest landlord. Follow for the AI money trail. Its amended S-1 lists $895.4B in compute and infrastructure commitments. The biggest: Google $121.3B SpaceX $120.5B Amazon $110B Microsoft $44.9B Q3 revenue was about $17B, so that's roughly 13x annualised sales locked in. And unlike the Google, Amazon and Microsoft deals, SpaceX's can be cancelled on 90 days' notice from next March. Source: Anthropic's S-1/A, via DatacenterDynamics and CNBC

OpenAI wants to stop calling them data centers. The new name, per reports: "AI compute campuses." The reason is this Gallup number: 71% of Americans oppose a data center in their local area. A rename won't fix power bills or water fights. Real community deals might.

DeepSeek just open-sourced a 24/7 GitHub coding agent. Follow for daily AI. Per Polymarket and X's news story: it's a research prototype called dsh-superagent. One Router watches a repo, turns issues into tickets, and hands each to its own worker agent in parallel. MIT licence. The real test isn't speed. It's whether you'd let it merge with no human looking.

Big Tech capex just hit $202B in one quarter. Follow for the AI money trail. Q3 2026 cash capex, straight from the 10-Q filings: Amazon $77.3B Alphabet $51.4B Microsoft $39.4B Meta $33.9B That's up 108% from $97.3B a year ago. Amazon alone outspent all four combined in Q3 2024. Filings don't split out AI, but most of the growth is data centers and servers.

He used AI to text her. She used AI to catch him. Now the AIs are dating. 💀 Part 2 soon #AI #Chatfishing #AIdating #Animation #Singlish

AI took my order a bit too literally. Milo Dinosaur means DINOSAUR is it?? 🦖 Part 2 if Dragon Fruit arrives... #AI #AIfail #Singapore #Animation #SGfood

Bought an AI that learns by listening. Now it talks like my Ah Ma and answers my mum for me 😭 Part 2 coming… #AI #VoiceAI #TechHumor #Singapore #Animation #Robots

The AI running our HR trained a copy of me on my own emails. It starts Monday. 🤖 Part 2

My company fired HR and hired an AI. It just denied my sick day in 0.3 seconds. 🤖 Part 1

I booked my vacation 3 months ago. My manager says HR never got it. Then I saw who forwarded my request 😳 Part 1

A genius AI employee still needs a boss. Follow for the fine print. Agents that run your calendar, apps and payments need two things before I'd trust them: a log of every action, and a confirm tap before money moves.

Anthropic's IPO is really a compute bet. The draft reportedly showed an $80.7B net loss in H1 2026 and possible compute spend of up to $750B. Friday's filing should show whether revenue can outrun that.