글

10월, 2026의 게시물 표시

Meta Muse Wonder Price: 40% of GPT-6, 39% More Per Task (2026)

이미지
▲ Meta Muse Wonder joins AI's big three Muse Wonder is Meta's first frontier AI model, released on October 12, 2026. For ChatGPT, Claude, or Gemini users, it adds a fourth serious option. The pitch is price: 40% of GPT-6 Astra's token rate . Is that enough to switch? What Is Muse Wonder? Meta's Frontier Model, 14 Months in the Making It is the first top-tier model from Meta's Superintelligence Lab since the company rebuilt its AI effort in August 2025. It ships in two versions. Muse Wonder is sold through the API (the pay-per-use channel for developers). The stronger Wonder Full is limited to Muse Code subscriptions starting at $50 a month . ▲ Muse Wonder score, token price, task cost Muse Wonder Pricing vs. GPT-6, Claude and Gemini Muse Wonder costs $4 per million input tokens and $20 per million output tokens. GPT-6 Astra and Claude Fable 5.1 charge $10 and $50. On the independent Artificial Analysis Intelligence Index, it scores 52 , one point beh...

DeepSeek Backs Winverse, a $400M Robot AI Lab, After Unitree Stake (2026)

이미지
▲ DeepSeek backs $400M robot AI lab Winverse Robot AI is the software that lets a robot understand instructions and carry out real-world tasks. DeepSeek, the Chinese lab best known for its chatbot, has agreed to lead the first funding round for robot AI startup Winverse AI, Bloomberg reports. For the US-China tech race, it shows where money goes next: off the screen and onto the factory floor. Why Is DeepSeek Investing in Robot AI? This is DeepSeek's second robot bet. In August , it invested 140.8 million yuan (about $20.8 million) in humanoid maker Unitree's Shanghai IPO, according to Reuters. The two agreed to co-develop AI models for humanoid robots. Bloomberg also reported DeepSeek is close to raising at least $12 billion ahead of a planned listing in early 2027 . ▲ Winverse seed round and $400M valuation Winverse Seed Round: A $400M Valuation The seed round is worth tens of millions of dollars and values Winverse at about $400 million . The startup is only m...

South Korea Bank Hacks: 1 Hacker, AI Tools, 9 Banks Targeted (2026)

이미지
▲ AI hacking hits 9 South Korean banks AI-assisted hacking is a cyberattack in which an attacker hands the work of finding flaws and breaking in to artificial intelligence tools. That definition stopped being abstract this month. Security firm CrowdStrike says a campaign against South Korean banks traces back to a single suspect, and that the attacker probably could not have pulled it off without AI. How Did One Hacker Target 9 South Korean Banks? Since late September, at least 9 South Korean banks have disclosed attacks or been reported as targets, according to Reuters. CrowdStrike found two tools behind the campaign. One is ARTEX, an open-source AI agent for penetration testing (authorized hacking that finds weaknesses), posted on GitHub this year by a Chinese developer. The other is Claude Code, Anthropic's AI coding tool. ▲ 9 banks, 25,000 customers, suspect aged 26 What CrowdStrike Found: 25,000 Customers and a 26-Year-Old Shinhan Bank said data on about 25,000 cu...

OpenAI Revenue Gap Explained: $50B, Not $70B, After FT Report (2026)

이미지
▲ OpenAI's $20B revenue gap explained Annualized revenue, or run rate, is a company's latest monthly revenue multiplied by twelve. On October 8, 2026, one such number moved the entire AI trade. The Financial Times reported that OpenAI's run rate was about $50 billion at the end of September, $20 billion below the $70 billion figure that had circulated two weeks earlier. Where OpenAI's $70 Billion Revenue Figure Came From On September 29 , Axios reported OpenAI's run rate was nearing $70 billion, up more than 70% since the third quarter began. The FT, citing financial documents, found OpenAI had told investors roughly $50 billion. A source told CNN the $70 billion "did not come from OpenAI": investors had added 70% growth to a $40 billion August estimate. ▲ What OpenAI told its investors: $50B run rate Why OpenAI and Anthropic Count Revenue Differently The goal was a direct comparison with Anthropic, which counts sales through cloud partner...

Samsung Q3 2026 Earnings Hit $80B, Korea's First ₩100T Quarter

이미지
▲ Samsung Q3 2026 earnings hit $80B Samsung earnings guidance is the company's early estimate of quarterly sales and operating profit, released before full results. On October 8, Samsung said Q3 2026 operating profit reached about ₩107.4 trillion (roughly $80 billion ). If you want to know where AI money is flowing next, this number is the clearest signal yet. Why Samsung's Profit Exploded The engine is AI memory. Data center build-outs are outrunning memory supply, which pushed DRAM (the working memory that holds data while chips compute) prices higher. Analysts say sales of HBM4 (High Bandwidth Memory - chips stacked next to AI processors to move data at ultra-fast speeds) for Nvidia's next platform are now reflected in the results. ▲ ₩107.4T profit, ₩195T revenue Samsung Q3 Earnings by the Numbers Revenue came in at about ₩195 trillion , up 126.6% year over year. Operating profit rose 782.5% from ₩12.2 trillion , a fourth straight record quarter. No Korean...

Boston Dynamics Names Alexa Veteran Prasad CEO, Hyundai Eyes 30,000 Robots (2026)

이미지
▲ Boston Dynamics new CEO Rohit Prasad Physical AI is artificial intelligence that lets robots see, decide and act in the real world. As humanoids move closer to factory floors, the edge goes to whoever makes those machines think. That is why Boston Dynamics just hired the AI leader who helped build Amazon's Alexa as its new CEO. Who Is Rohit Prasad, Boston Dynamics' New CEO? Boston Dynamics announced Prasad on Oct. 6, 2026 , effective Oct. 7 . He is only the third CEO since the company's 1992 founding. Prasad spent 12 years at Amazon as SVP and head scientist for Alexa and AGI. He helped grow Alexa into a product used in hundreds of millions of homes and led Amazon's Nova AI models. Before that, he spent nearly 14 years at Raytheon BBN leading machine-learning research. ▲ 12 years at Amazon, third CEO since 1992 Why an AI Leader After the CEO Gap? The top job had been open since Robert Playter stepped down in February 2026 , with CFO Amanda McMaster se...

Trump Launches Super Intelligence Force: Spy Chief, 120-Day AI Report (2026)

이미지
▲ Trump's new AI command puts speed first The Super Intelligence Force is a White House AI task force created by President Trump on October 4, 2026. It is chaired by Director of National Intelligence Jay Clayton, the nation's top spy chief. The message: Washington puts speed ahead of new rules, even as AI's builders ask to slow down. Why Trump Launched the Super Intelligence Force The launch follows a fight over speed. On September 6 , Anthropic CEO Dario Amodei called for deliberately slowing AI progress, and OpenAI CEO Sam Altman agreed. On September 29 , six companies, including Google, Meta, Nvidia and OpenAI, signed a White House safety accord Trump called "morally binding." The same day, an executive order told agencies to say "Super Intelligence" instead of "AI." ▲ Super Intelligence Force leaders and 120-day mission Who Runs the AI Task Force and What It Must Do Clayton, a former SEC chairman, became DNI in July 2026. His t...

Musk Confirms TSMC Talks on $16.8B Terafab - Intel 14A Plan in Question (2026)

이미지
▲ Intel vs TSMC: who builds Musk's chips Terafab is the giant chip manufacturing project Tesla and SpaceX are building in Texas to make their own AI chips. It began with Intel as a partner. Now TSMC, the world's largest contract chipmaker, is exploring a role in Terafab, and Elon Musk has confirmed the talks. The outcome could decide who builds America's most advanced AI chips. What Is Terafab, Musk's Texas Chip Plant? Terafab sits in Grimes County, Texas. Tesla and SpaceX said in August they would spend $16.8 billion on phase one, and a May filing showed total investment could reach $119 billion . The plant aims to make, package and test logic and memory chips under one roof for more than 1 terawatt of future compute. Intel joined in April, and Musk said Tesla plans to use Intel's 14A process (a next-generation, 1.4nm-class manufacturing node). ▲ Musk confirms TSMC talks on X TSMC and Terafab: Two Deal Structures Former Bloomberg columnist Tim Culpan ...

Pentagon Autonomous Warfare Command Plan: Drone Budget Request Up 240x (2026)

이미지
▲ War has a new star: robots The Autonomous Warfare Command (AUTOWARCOM) is a planned four-star US combatant command devoted to drones, robots and other autonomous weapons. Defense Secretary Pete Hegseth announced it on September 30, 2026 . It marks a shift from costly crewed platforms toward machines America can build cheaply and in bulk. Why the Pentagon Wants an Autonomous Warfare Command Now An IFRI report cited by Reuters found drones caused under 10% of Ukraine war casualties in 2022 and up to 80% in 2025 . AP notes the US is seven months into a war with Iran, whose cheap drones have killed US troops. A Senate defense bill already allows a robotics command. "The pace of war is changing faster than the process to support it," Hegseth said. ▲ 4 stars, $54.6B and 120 days Autonomous Warfare Command by the Numbers Per Reuters, it would be the first new combatant command since 2019 , the military's 12th , with its own budget and acquisition (weapons-buyin...

Micron Q4 2026 Earnings: Revenue Up 4.8x as Memory Becomes AI's Bottleneck

이미지
▲ Micron Q4 earnings: AI's choke is memory A memory bottleneck is when AI data center growth is capped by memory chip supply rather than by compute chips or power. Micron's fiscal Q4 2026 earnings on September 30, 2026 put that idea front and center. The AI buildout's pace is now set inside memory fabs, where Micron's revenue grew 4.8x in a year. What Micron's Q4 2026 Earnings Revealed About AI AI chip shortages usually bring GPUs (graphics processing units, the chips that run AI math) to mind. But Micron President Manish Bhatia said on the earnings call that DRAM, more than logic chips or power, is the principal constraint on data centers. Customers are shipping as many AI chips as their memory supply allows. Big Tech, meanwhile, is spending over $730 billion on AI infrastructure this year. ▲ Micron Q4 revenue $54.2B, up 4.8x Micron Q4 2026 Earnings by the Numbers Revenue for the quarter ended September 3 reached $54.23 billion , up 379% from $11.3 ...