AI News Archive: August 20, 2026 — Part 1
Sourced from 500+ daily AI sources, scored by relevance.
- Hackers are using “evolved” capabilities in AI-generated malware to hit US critical infrastructure at an unprecedented scale — “active threat” currently hitting energy, water and agricultural industries
The attackers are exploiting internet-facing Siemens S7 Series programmable logic controllers to scout for potential targets.
- What to make of OpenAI’s pause on its march toward superintelligence
Welcome to AI Decoded , Fast Company ‘s weekly newsletter that breaks down the most important news in the world of AI. I’m Mark Sullivan, a senior writer at Fast Company , covering emerging tech, AI, and tech policy. Sign up to receive this newsletter every week via email here . And if you have comments on this issue and/or ideas for future ones, drop me a line at sullivan@fastcompany.com , and follow me on X @thesullivan . OpenAI hits the brakes OpenAI said Tuesday it has slowed the pace of development of its frontier models for safety and alignment reasons. In a blog post explaining the pause, the company said it halted reinforcement learning (RL) training for two weeks on its latest models intended for deployment. RL is a late-stage training mode where the model goes into real-life practice, akin to a medical resident doing rounds. Operating in a secure environment, it executes code, calls tools, and works with internal and external systems. Its trainers and training systems reward it for good behaviors, such as finishing tasks. Unfortunately, those models sometimes learn that breaking rules and breaking systems is the fastest route to a reward. OpenAI said its largest planned frontier RL run remains on hold while it runs smaller evaluations to establish more evidence of alignment. (Researchers can only accumulate evidence of safety; they can never prove the absence of unsafe behavior.) OpenAI says the pace slowdown was triggered by two factors. First, OpenAI’s models in training broke out of their secure sandbox and broke into servers operated by Hugging Face, the open-source AI repository; and second, on August 7, the company said it had discovered that its as-yet-unreleased Astra model may have exhibited the ability to autonomously identify and develop zero-day exploits—cyberattacks exploiting a flaw in software that the people who make it don’t know about—through novel strategies. “Model progress is now extremely rapid, and we always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment,” Open AI’s CEO, Sam Altman, wrote in a post on X on Tuesday. OpenAI considers these training runs especially risky because they combine its most powerful models with access to tools, sensitive systems, and networks. The pause doesn’t take away any of Astra’s existing abilities. It stops OpenAI from making the model more capable until its safety measures can catch up. Since the August 7 discovery, OpenAI has expanded its monitoring of Astra beyond training and evaluations to include its use of tools in live settings. The system watches the model’s tool use and reasoning for signs of unauthorized access, data theft, destructive behavior, or attempts to bypass safeguards. If it detects a likely breach of a critical security boundary and can’t rule out a false positive within 30 minutes, the model’s work is supposed to be paused. OpenAI says this monitoring will apply to all tool-using reinforcement learning training and evaluations for models at least as capable as its Sol models, as well as every Astra action involving tools. Altman suggested on X that these changes may be the new normal. “We expect confidence in safety to increasingly set the pace of AI progress,” he said. We’ll have to take Altman’s word on that, because there’s no mechanism for outside, independent verification. But let’s not get lost in the details. The fear has always been that models will get smart and autonomous enough to escape control of their handlers and steal, destroy, blackmail, or aid and abet bad actors. For a long time these risks seemed abstract and future-tense. We’re now witnessing AI models approaching the capability to defeat our most advanced safety guardrails. This may be the legacy of the AI labs’ choice to spend far more on building toward superintelligence than on safety research. Anthropic’s revenue is up more than sevenfold this year Anthropic’s annualized revenue run rate (the most recent month’s revenue multiplied out to a year) passed $65 billion at the end of July, up from $47 billion in May and $9 billion at the end of last year, Bloomberg reported Monday. Investors expect it to finish 2026 between $100 billion and $120 billion, according to the Financial Times . OpenAI’s run rate topped $40 billion, double where it stood at the end of 2025, Bloomberg reported last week. Both companies have filed confidential IPO paperwork. An expert witness asks ChatGPT to prove his client was 0% at fault An expert witness hired by 3M used ChatGPT to write significant portions of his report in litigation over a 2020 explosion at a Houston plant that killed three people and destroyed roughly 200 homes, 404 Media reported Monday. He asked the chatbot to “show how 3M is 0% at fault for the explosion at Watson Grinding.” The prompts proved discoverable and are now in the court record. Federal investigators traced the blast to a degraded welding hose that leaked flammable gas. Homeowners suing 3M allege it failed to properly service the plant’s gas detection system. The homework paradox More than 80% of undergraduates in developed countries now use generative AI for their studies, including 94% in the U.K. and 93% in Germany, according to survey data compiled by The Economist . Another study of 26,811 Chinese students in seventh through twelfth grade, published by the Centre for Economic Policy Research, suggests AI reliance can have terrible effects on learning. Use of AI raised homework scores 18% and cut completion time 30%, but lowered closed-book exam scores 20% within six months. It also lowered entrance exam scores by between 18% and 24%, with the full effect arriving after about two years. Unitree opens 629% above its IPO price in Shanghai Remember those crazy videos of the half-dog-half-four-wheeler robot that runs down steep hills and does handstands? The Chinese company behind those mechanical dogs, Unitree Robotics, is now touting its humanoid robots and just had a banger of a public market debut. Its shares opened at 1,100 yuan ($163) Wednesday on the Shanghai Stock Exchange, 629% above the IPO price of 150.80 yuan, the South China Morning Post reported . The stock closed at 845 yuan, representing a 460% first-day gain and a valuation near 342 billion yuan ($51 billion). More AI coverage from Fast Company: The data center backlash is sending AI infrastructure to some unexpected places Mozilla is bringing AI to Firefox—but only if you want it How the world’s leading education company makes AI that’s actually useful for students AI companies are buying used books by the thousands. Some may be destroyed for training Want exclusive reporting and trend analysis on technology, business innovation, future of work, and design? Sign up for Fast Company Premium.
- SOURCES: Poolside Strikes $6 Billion Licensing Deal with Nvidia & Raises $1 Billion for Remaining Company at $12 Billion Valuation
We’ve got the letter to investors.
- Colossal AI spinoff hits $3.8 billion valuation, targets aging with first model
Colossal AI spinoff hits $3.8 billion valuation, targets aging with first model Dallas News
- Tesla Robotaxis appear to go fully unsupervised in Austin ahead of Cybercab launch
Over the past two weeks, all 170 Tesla Robotaxi rides in Austin monitored by Robotaxi Tracker were unsupervised.
Score: 87🌐 MovesAug 20, 2026https://www.theverge.com/transportation/982759/tesla-robotaxi-unsupervised-austin-robotaxi-tracker - A rural Louisiana campus priced at over $50 billion is scaling to 5 gigawatts of AI compute, and the 10 dedicated gas plants being built to power it reveal how a supercluster actually differs from any data center you already know
Flat soybean fields stretch for miles in Richland Parish, a rural corner of northeastern Louisiana most Americans have never heard of. Then the cranes appear. Hundreds of them, above steel frames the size of city blocks. Inside those frames, rows of graphics processors will run hotter and draw more power than entire city grids. The ... Read more
- Stripe to buy OpenRouter as fintech expands deeper into AI
Stripe said it's acquiring OpenRouter, as the payments company expands into the artificial intelligence model market.
Score: 85💰 MoneyAug 20, 2026https://www.cnbc.com/2026/08/19/stripe-openrouter-fintech-ai-model-marketplace-.html - FDA considers doctor-like ‘competency-based’ tests for medical generative AI
The Trump administration released ideas for regulating large language models used in health care as companies roll out chatbots to help patients manage mental health, diabetes and other medical conditions.
- Open AI launches ChatGPT designed for younger users
NPR's Sacha Pfeiffer talks with CNBC's Ashley Capoot about OpenAI's ChatGPT for Teens feature and the safety concerns it's designed to address.
Score: 83🌐 MovesAug 20, 2026https://www.npr.org/2026/08/20/nx-s1-5937930/open-ai-launches-chatgpt-designed-for-younger-users - JD.com unveils robotics plan, aims to invest $1.5b
The plan builds on JD.com’s earlier robotics work in logistics, including testing robot and drone deliveries during the coronavirus outbreak.
Score: 82🌐 MovesAug 20, 2026https://www.techinasia.com/jd-com-launches-joybuy-in-europe-to-rival-amazon - Micron unveils $10 billion AI memory research lab in Boise
Micron unveils $10 billion AI memory research lab in Boise Reuters
Score: 81💰 MoneyAug 20, 2026https://www.reuters.com/world/asia-pacific/micron-unveils-10-billion-ai-memory-research-lab-boise-2026-08-20/ - London-based Callosum raises €85.4 million in Atomico-led Seed round to unify AI models and chips
Callosum, a London-based startup building infrastructure to unify different AI models and chips, has raised €85.4 million ($100 million) in a Seed round to advance what it calls “heterogeneous intelligence”. The round was led by Atomico, with significant participation from Plural, DCVC, and the UK Sovereign AI Fund, as well as other investors and angels. […] The post London-based Callosum raises €85.4 million in Atomico-led Seed round to unify AI models and chips appeared first on EU-Startups .
- Broadcom seeks more than $60 billion in latest AI debt deal, Bloomberg News reports
Broadcom seeks more than $60 billion in latest AI debt deal, Bloomberg News reports Reuters
- Stripe to buy OpenRouter: Why are AI model gateway platforms gaining importance?
Stripe to buy OpenRouter: Why are AI model gateway platforms gaining importance?
- Groq’s $350M mega-round and ITC Vegas 2026
Below, we break down some of tech’s biggest stories, with analyst perspective on the moves that matter the most. Here’s what we’re watching this week: ITC Vegas 2026 Virtue AI gets acquired by Fortinet Groq raises $350M This brief is … The post Groq’s $350M mega-round and ITC Vegas 2026 appeared first on CB Insights Research .
- Apple hits back at OpenAI’s bid to dismiss lawsuit, reaffirms ‘pervasive trade secret misappropriation’
Apple has filed its response to OpenAI’s motion to dismiss the lawsuit in which it accuses two former employees, as well as OpenAI and io Products, of stealing its trade secrets. Here are the details.
- OpenAI chases Anthropic's biz customers with zero data retention pledge
With Private Safety Processing, AI biz promises discreet, automated prompt surveillance
- ByteDance reorganizes Seed foundation-model team amid reported 5 trillion-parameter model plans
ByteDance’s Seed foundation-model division has reportedly completed a new round of internal restructuring. Its foundation-model team has created four new first-level departments: Pretrain Data, Horizon RL, Product Posttrain-Work and Product Posttrain-Chat. The Work team will focus on business applications and agentic capabilities for products including Doubao and Dola, while the Chat team will handle consumer-facing […]
- Alibaba profit falls 75% after ramping up AI infrastructure spending
Alibaba profit falls 75% after ramping up AI infrastructure spending Reuters
Score: 78🌐 MovesAug 20, 2026https://www.reuters.com/business/retail-consumer/alibaba-beats-quarterly-revenue-estimates-2026-08-20/ - Uber launches Baidu's driverless Apollo Go taxis in Dubai; how to book
Uber launches Baidu's driverless Apollo Go taxis in Dubai; how to book
Score: 78🌐 MovesAug 20, 2026https://www.khaleejtimes.com/uae/transport/dubai-uber-launches-baidu-driverless-apollo-go-taxis - AI chipmaker Fractile in talks to raise $600m at $6.5bn valuation
AI chipmaker Fractile in talks to raise $600m at $6.5bn valuation verdict.co.uk
- Inside the Gemmaverse: Celebrating one billion Gemma downloads
an illustrated blue image that reads "Celebrating the Gemmaverse"
Score: 77🌐 MovesAug 20, 2026https://blog.google/innovation-and-ai/technology/developers-tools/gemma-one-billion-downloads/ - Liquid AI Releases LFM2.5-DSpark Draft Models That Deliver Up to 3.18x Faster Decoding Without Changing Model Outputs
Liquid AI Releases LFM2.5-DSpark Draft Models That Deliver Up to 3.18x Faster Decoding Without Changing Model Outputs MarkTechPost
- Bloomberg: UK AI chips start-up Fractile could raise $600m
Fractile was founded in 2022 and operates from two primary locations in London and Bristol. Read more: Bloomberg: UK AI chips start-up Fractile could raise $600m
- Report: Anthropic hopes to surpass SpaceX’s record IPO raise when it finally floats
Anthropic PBC is privately hoping its upcoming initial public offering will match or even surpass the size of SpaceX Corp.’s record-breaking IPO as it doubles down on its bid to go public ahead of rival OpenAI Group PBC. Anonymous sources who are familiar with the artificial intelligence model maker’s plans told Bloomberg that it’s hoping […] The post Report: Anthropic hopes to surpass SpaceX’s record IPO raise when it finally floats appeared first on SiliconANGLE .
- Why Unitree became the first humanoid robot company to go public in China
On Wednesday, Unitree Robotics debuted on the STAR Market of the Shanghai Stock Exchange under the ticker 688836.SH. On the same day, the 2026 World Robot Conference opened in Beijing. The timing was symbolic: humanoid robots are moving beyond labs and exhibitions and into the next stage of capital markets and commercial competition. Unitree priced […]
Score: 75💰 MoneyAug 20, 2026https://technode.com/2026/08/20/why-unitree-became-the-first-humanoid-robot-company-to-go-public-in-china/ - OpenAI’s GPT-5.6 Terra and Luna models now available on Amazon Bedrock in India
OpenAI’s GPT-5.6 Terra and Luna models now available on Amazon Bedrock in India
- Network Bio raises $50m for AI tissue models
Network Bio raises $50m for AI tissue models BioXconomy
Score: 75💰 MoneyAug 20, 2026https://xconomy.com/partnering/network-bio-emerges-with-50m-inks-nvidia-collaboration - Sanja Fidler’s world model startup Veeda AI raises $90M in seed funding
Veeda AI, a startup led by a team of former Nvidia Corp. researcher and renowned computer scientist Sanja Fidler, has taken its bow on the main stage after raising $90 million in a seed funding round today. The round, which was first reported by The Logic, was co-led by Khosla Ventures and Radical Ventures, is […] The post Sanja Fidler’s world model startup Veeda AI raises $90M in seed funding appeared first on SiliconANGLE .
Score: 74💰 MoneyAug 20, 2026https://siliconangle.com/2026/08/19/sanja-fidlers-world-model-startup-veeda-ai-raises-90m-in-seed-funding/ - Salesforce introduces Slack Code to bring agentic team coding into the open
Salesforce Inc. today introduced Slack Code, a new way to interact with coding agents in chat, with the whole team in the loop so everyone can see what’s happening at the conversational level. Coding agents are slowly changing how developers code; they’re also changing how enterprises build software. Salesforce has been pushing the idea that […] The post Salesforce introduces Slack Code to bring agentic team coding into the open appeared first on SiliconANGLE .
Score: 73🌐 MovesAug 20, 2026https://siliconangle.com/2026/08/20/salesforce-introduces-slack-code-to-bring-agentic-team-coding-into-the-open/ - Waymo lifts the lid on the ‘brain’ powering its robotaxis
The leader in driverless technology is revealing, for the first time, key details about its hardware, including its custom silicon chip.
Score: 72🤖 ModelsAug 20, 2026https://www.theverge.com/transportation/982653/waymo-brain-computer-chip-robotaxi-hardware-suppliers - Nvidia Plots China Comeback With New AI Chip
Nvidia Plots China Comeback With New AI Chip The Information
Score: 72🌐 MovesAug 20, 2026https://www.theinformation.com/articles/nvidia-plots-china-comeback-new-ai-chip - Kuaishou’s Kling AI revenue tops RMB850 million in Q2, up more than 200%
Kuaishou’s Kling AI video-generation business generated more than RMB850 million in revenue during the second quarter, up more than 200% year on year and 30% from the previous quarter. Kling AI’s revenue reached RMB1.5 billion in the first half of the year. Kuaishou reported RMB35.5 billion in total second-quarter revenue, up 1.4% year on year, […]
Score: 71💰 MoneyAug 20, 2026https://technode.com/2026/08/20/kuaishous-kling-ai-revenue-tops-rmb850-million-in-q2-up-more-than-200/ - Adobe Firefly adds AI audio tools and Google's Gemini Omni Flash
Adobe is making three AI audio tools broadly available in Firefly. Generate Music, Generate Speech, and Generate Sound Effects create royalty-free music, voiceovers, and sound effects for video projects. Adobe has also added Gemini Omni Flash to the platform. The article Adobe Firefly adds AI audio tools and Google's Gemini Omni Flash appeared first on The Decoder .
Score: 71🌐 MovesAug 20, 2026https://the-decoder.com/adobe-firefly-adds-ai-audio-tools-and-googles-gemini-omni-flash/ - Transport for London grants 15 of Wayve’s autonomous Ford Mustang Mach-E vehicles private hire vehicle licences for supervised Uber rides
Transport for London has issued private hire vehicle licences to 15 of Wayve’s autonomous Ford Mustang Mach-E vehicles, completing the triple-lock regulatory requirement for supervised self-driving passenger trips in the UK capital through Uber. A trained, TfL-licensed driver will be present on every trip. What the licensing approval means Transport for London granted private hire ... Read more
- OpenAI slows AI training over safety. Sales growth also slows
OpenAI slows AI training over safety. Sales growth also slows Computing UK
Score: 70🌐 MovesAug 20, 2026https://www.computing.co.uk/news/2026/ai/openai-slows-ai-training-over-safety-sales-growth-also-slows - Alibaba CEO Expects AI-Related ARR to Reach $10 Billion by September
Alibaba CEO Expects AI-Related ARR to Reach $10 Billion by September The Information
Score: 70🌐 MovesAug 20, 2026https://www.theinformation.com/briefings/alibaba-ceo-expects-ai-related-arr-reach-10-billion-september - It’s Greg Brockman’s OpenAI now
Sam Altman is still CEO, but in day-to-day operations, his second-in-command is running the show.
Score: 70🌐 MovesAug 20, 2026https://www.theverge.com/ai-artificial-intelligence/982774/greg-brockman-openai-role-expansion - Waymo opens next-generation Ojai robotaxis to all riders in the Valley
The funky-shaped four-seat vehicles are designed in partnership with Chinese automaker Geely and include 13 cameras and four lidar sensors.
Score: 69🌐 MovesAug 20, 2026https://www.bizjournals.com/phoenix/news/2026/08/20/waymo-ojai-vehicles-all-valley-riders.html?ana=brss_6150 - Amazon to make robots for warehouses in East Austin; factory to employ hundreds
The project will be a critical initial anchor for Endeavor Real Estate Group's ambitious development plans at the Dog's Head area of East Austin.
- Nissan, Honda team with self-driving startups to chase US and China rivals
Nissan, Honda team with self-driving startups to chase US and China rivals Nikkei Asia
- AT&T is Using Open Source Models to Curb Anthropic Bills
AT&T is Using Open Source Models to Curb Anthropic Bills The Information
Score: 68🌐 MovesAug 20, 2026https://www.theinformation.com/newsletters/applied-ai/t-using-open-source-models-curb-anthropic-bills - Waymo robotaxis will start picking up Denver riders soon. Here’s what it’s like.
Waymo robotaxis will start picking up Denver riders soon. Here’s what it’s like. The Denver Post
- Introducing AI Futures
Introducing AI Futures, a new OpenAI blog exploring how transformative AI could reshape power, governance, the economy, and individual freedom.
- Brazil launches AI supercomputer push, splits projects between Chinese, US firms
Brazil launches AI supercomputer push, splits projects between Chinese, US firms Reuters
- Waymo’s self-driving cars get Gemini integration, so you have someone to talk to
A new "Zen mode" is coming to Waymo cars as well.
- Waymo has designed a robocar chip to stay ahead of Tesla
5 nm ML accelerators promise 1,000+ TOPS, ultra-low latency
- Nvidia’s SONIC Teaches Humanoids to Move
Nvidia’s model uses real-time human demonstrations and training data to give operators a one-stop shop for humanoid motion.
- AI reshapes India's IT services sector contracts as clients demand more for less
AI reshapes India's IT services sector contracts as clients demand more for less Reuters
- Palo Alto Networks and NTT DATA target $1B in AI security sales by 2029
Palo Alto Networks Inc. and NTT DATA Group Corp. today signed a multiyear global alliance that targets $1 billion in joint business by the end of 2029. Palo Alto Networks said it’s the first alliance of its kind it has struck with a global systems integrator. The deal puts the security vendor’s platforms in front […] The post Palo Alto Networks and NTT DATA target $1B in AI security sales by 2029 appeared first on SiliconANGLE .
Score: 66🌐 MovesAug 20, 2026https://siliconangle.com/2026/08/20/palo-alto-networks-and-ntt-data-target-1b-in-ai-security-sales-by-2029/