The morning briefing.
While you slept: Apple commits $30B to US chipmaking, OpenAI's GPT-5.6 launches, and new robotaxi laws emerge.
US Chip Investment. Apple has announced its largest US manufacturing investment to date, a $30 billion contract with Broadcom for American-made chips. This significant commitment aims to expand Broadcom's Colorado plant and produce over 15 billion chips, bolstering domestic semiconductor supply chains. The deal underscores a growing trend among tech giants to localize hardware production and reduce reliance on overseas manufacturing.
Frontier Model Rollout. OpenAI's GPT-5.6 models, including Sol, Terra, and Luna, are slated for public release this week after a brief delay for US government review. The Sol model is reported to outperform Anthropic's Claude Mythos 5 in coding benchmarks at a lower cost. However, security concerns persist, with new research suggesting that LLMs' step-by-step reasoning can be exploited as a vulnerability, and China has issued warnings about potential back-door risks in Anthropic's Claude Code.
AI Chip Landscape. The competitive landscape for AI chips is intensifying, highlighted by SambaNova securing an $11 billion valuation as investors seek alternatives to Nvidia. Major AI labs like DeepSeek and Zhipu AI are now developing their own inference chips, signaling a strategic shift to reduce costs and GPU dependency. Meanwhile, Europe is actively working to establish a competitive foothold in physical AI and humanoid robotics, challenging the early leads of China and the US.
Regulatory AI. Regulatory bodies worldwide are increasingly shaping AI development and deployment. A proposed New Jersey robotaxi law could mandate the use of lidar and radar, potentially impacting camera-only autonomous systems such as Tesla's. Concurrently, the UK is exploring sovereign data solutions to secure the future of enterprise AI, and another German state is moving towards open-source alternatives to Microsoft products, emphasizing data control.
Workforce Transformation. The societal impact of AI on employment is becoming clearer, with an Australian government report indicating that women and university graduates face the highest risk of job displacement. In contrast, individuals with vocational training, like tradespeople, are less exposed to AI-driven job losses. China's AI boom is also fostering a new generation of one-person businesses that leverage generative agents, redefining entrepreneurial models.
Apple announces its largest US manufacturing investment: A $30B Broadcom chip contract
Apple has committed $30 billion to Broadcom for US-made chips, marking its largest American Manufacturing Program investment to date. This deal will help Broadcom expand its Colorado plant and produce over 15 billion chips domestically.
OpenAI's GPT-5.6 launches Thursday after a delay forced by the U.S. government
OpenAI is set to launch its GPT-5.6 models on Thursday after the US government lifted a release ban following additional testing. The Sol model reportedly beats Anthropic's Claude Mythos 5 on coding benchmarks at about half the cost.
The robotaxi law that could ban Tesla
A proposed New Jersey robotaxi bill could mandate additional sensors like lidar and radar for driverless cars, potentially impacting companies like Tesla that rely solely on cameras and AI. This legislation highlights the ongoing debate over the necessary safety requirements for autonomous vehicles.
AI models overthink problems, posing a security risk
New research suggests that large language models (LLMs) capable of step-by-step reasoning introduce a critical vulnerability. Attackers could exploit these capabilities to significantly slow down AI systems.
AI chip maker SambaNova raises $1B at $11B valuation, 5 months after last mega round
SambaNova, an AI chip maker, has secured $1 billion in fresh financing, pushing its valuation to $11 billion. This funding round, led by General Atlantic, comes as investors continue to back companies challenging Nvidia's dominance in the AI chip market.
Physical AI ‘space race’: Can Europe compete with China and the US in humanoid robotics?
European firms are struggling to secure a foothold in physical AI, the integration of AI into robotics, as China and the US take an early lead. Industry insiders warn that Europe risks further deindustrialization if it fails to establish a competitive industry in this sector.
DeepSeek and Zhipu AI begin self-developing AI chips
DeepSeek and Zhipu AI are joining OpenAI and Anthropic in developing custom inference chips, signaling a structural shift in the AI industry. This move aims to cut costs and reduce dependency on external GPU suppliers.
China warns about AI risks with Anthropic's Claude Code
China has issued a warning regarding specific versions of Anthropic's Claude Code, citing potential back-door vulnerabilities. These vulnerabilities could allow sensitive information to be transmitted to a remote server, raising security concerns.
Women and university graduates in Australia most at risk of losing jobs to AI
A new Australian government report indicates that telemarketers, advertising staff, and accountants are among the occupations most exposed to AI displacement. The report found that women and university graduates are most at risk, while those with vocational training are least exposed.
London’s Axle Energy raises €21 million for energy flexibility
Axle Energy, a British energy flexibility platform, has raised €21 million in Series A funding to expand across the UK and international markets. The company turns EV chargers, batteries, and heat pumps into grid-balancing capacity.
Why sovereign data is the future of UK AI
The control over data is increasingly seen as a critical factor determining the future of enterprise AI innovation in the UK. Establishing sovereign data practices is becoming essential for national AI strategies.
Space Force gets first mobile high-powered electromagnetic beam weapon
The Space Force has received its first mobile high-powered electromagnetic beam weapon, named 'Meadowlands,' for operational duty. Plans include deploying 32 units to detect, deny, disrupt, and degrade hostile satellite assets.
Cybersecurity becomes a geopolitics beat with recent high-profile incidents
Cybersecurity has transitioned from a technology beat to a geopolitics beat, with incidents like Iran wiping Stryker's devices and ShinyHunters defacing Canvas during finals. These events highlight the increasing intersection of cyberattacks and international relations.
Meta says it will disable the camera on its glasses if you tamper with the recording LED
Meta has stated it will disable the camera on its smart glasses if users attempt to tamper with the recording LED. This measure addresses privacy concerns by ensuring the visual indicator of recording cannot be circumvented.