Nvidia addresses H200 orders as Apple invests in AI hardware

Key Takeaways

• Nvidia CEO Jensen Huang is in China to address H200 chip sales, with conflicting reports on whether China will approve commercial imports for major tech firms like Alibaba, ByteDance, and Tencent.
• AI experts at Davos, including Google DeepMind's Demis Hassabis and Anthropic's Dario Amodei, offer widely divergent timelines for achieving human-level AGI, ranging from 5-10 years to within 1-2 years for significant milestones.
• Skepticism surrounds new dedicated AI hardware like Jony Ive and Sam Altman's "iO device," despite Apple and OpenAI reportedly investing in such products for 2026, aiming to overcome past failures of devices like Google Glass.
• Anthropic's report on Claude.ai indicates humans still oversee 52% of generative AI use, positioning AI as a tool rather than a job replacement, with primary usage in software, education, and creative tasks.
• Nvidia's survey in financial services shows 65% of professionals actively use AI (up from 45% last year), with 42% assessing or deploying agentic AI, though challenges like performance and skilled staff persist.
• Current AI identity systems are vulnerable to quantum computing attacks, necessitating the adoption of post-quantum cryptography (e.g., CRYSTALS-Dilithium, CRYSTALS-Kyber) and context-aware access for protection.
• Jim VandeHei warns that AI will profoundly change work and life within months, urging individuals to learn and use AI tools like Gemini, ChatGPT, and Claude daily to adapt.
• Aliisa Rosenthal, OpenAI's former sales head, has joined Acrew Capital as a general partner, focusing on guiding startups in enterprise AI applications.
• Nvidia and Micron Technology are expected to remain leaders in the AI stock market due to their critical roles in providing GPUs and high-bandwidth memory for AI infrastructure.

Nvidia CEO Huang Visits China Amid H200 Chip Sales Block

Nvidia CEO Jensen Huang is visiting China before Lunar New Year. He aims to address reported sales blocks on Nvidia's H200 chips. China is reportedly limiting H200 purchases to research use only, stopping commercial sales. This impacts Nvidia's business in a market that once made up 20% of its data center revenue. Huang will meet buyers to discuss how to supply U.S.-approved chips despite growing tech restrictions. This trip is very important for Nvidia's future in China.

China Approves Nvidia H200 Chip Orders for Top Tech Firms

Chinese officials told major tech companies like Alibaba, ByteDance, and Tencent to prepare orders for Nvidia's H200 AI chips. This means Beijing is likely to approve imports of these key AI components soon. This news goes against earlier reports that China was blocking H200 shipments. The H200 is an older chip that the US government allows to be sold to China. This approval would be a significant success for Nvidia, which aims to restart business in the world's largest semiconductor market.

Nvidia CEO Jensen Huang Heads to China Amid Chip Sales Issues

Nvidia CEO Jensen Huang will visit China in the coming days. This trip comes as Nvidia faces difficulties selling its AI chips in the Chinese market. Both US export rules and Chinese regulations create hurdles for the company. Huang plans to attend an Nvidia company party in Beijing on Monday. He will also meet with potential buyers to discuss challenges in supplying US-approved chips to China. Reports suggest China may only approve H200 AI chips for research use.

Anthropic Report Shows Humans Still Guide Most AI Use

A new report from Anthropic shows that humans still oversee most generative AI use. The study looked at millions of interactions with Claude.ai and found that 52% involved users guiding the AI and making final decisions. This means AI is a helpful tool, but it is not replacing many jobs yet, as businesses use it carefully. The report identified over 3,000 work tasks, with the top 10, mostly software-related, making up a large part of Claude's use. Usage is also growing in education and creative areas, but remains focused on specific tasks.

AI Leaders Disagree on Human-Level Intelligence Timeline at Davos

Top AI experts at the World Economic Forum in Davos had different ideas about how soon AI will reach human-level intelligence. Demis Hassabis from Google DeepMind thinks AGI is 5 to 10 years away and needs more breakthroughs. Yann LeCun, a famous AI pioneer, said that current large language models will not achieve human-like intelligence and a different method is needed. However, Dario Amodei, CEO of Anthropic, believes AI will replace all software developers within a year and reach Nobel-level scientific research in two years. OpenAI CEO Sam Altman also thinks we are nearing superintelligence.

Skepticism Grows for Jony Ive and Sam Altman AI Device

Confusion surrounds Jony Ive and Sam Altman's mysterious "iO device," an upcoming AI hardware product. The author remains doubtful about the need for dedicated AI hardware, suggesting smartphones already handle many AI tasks through apps. Previous AI wearables like Google Glass and Humane's AI Pin faced problems with privacy and performance. Despite the strong reputations of Apple and OpenAI, the author questions why consumers would buy a new AI device when they already own a smartphone. The exact form of the iO device is still unknown, with ideas ranging from a pen to a special audio product.

New Security Protects AI Agents from Quantum Attacks

Current AI identity systems are at risk from quantum computing attacks, which can easily break common encryption like RSA and ECC. Shor's algorithm can crack the encryption in Model Context Protocol (MCP) hosts, letting quantum computers fake AI agent identities. To protect against this, experts suggest using post-quantum cryptography (PQC) such as CRYSTALS-Dilithium for signing and CRYSTALS-Kyber for secure data tunnels. For small edge devices, PQuAKE provides anonymous key exchange. Also, "4D Space" context-aware access adds time and behavior checks to identity, creating a backup defense against attacks.

Nvidia Survey Shows AI Growth and Challenges in Finance

Nvidia's "State of AI in Financial Services" report surveyed over 800 professionals, showing strong growth in AI use. Almost all respondents expect AI budgets to stay the same or increase, and 65% now actively use AI, up from 45% last year. Generative AI use increased to 61%, and 42% are using or assessing agentic AI, with 21% already deploying it. The main challenges for agentic AI include performance, lack of skilled staff, data problems, implementation difficulties, and ethical concerns. Separately, Datalign Advisory is testing "Relationship AI" to help financial advisors find new clients by spotting important life events and communication styles.

Apple and OpenAI Bet on New AI Hardware for 2026

Apple and OpenAI are reportedly investing in new AI hardware for 2026, hoping people are ready for constant digital devices. Previous AI wearables like Google Glass and Humane's AI Pin faced problems with privacy and performance. However, some devices, such as Meta's Ray-Ban smart glasses, are becoming popular because they are useful. Apple and OpenAI have key advantages: Apple's strong reputation for trust and its connected ecosystem, and OpenAI's powerful AI models. For these new AI devices to be widely accepted, they must address privacy worries and prove their value, just like smartphones did.

Jim VandeHei Warns Family About AI's Fast Impact

Jim VandeHei wrote a letter to his family, warning that AI will deeply change work and life within months. He urges them to learn and use AI every day with curiosity. He suggests experimenting with large language models like Gemini, ChatGPT, and Claude for writing, research, and building ideas. VandeHei believes these tools are powerful but emphasizes that human creativity and critical thinking remain vital. He worries that leaders are not preparing young people for these big changes. He advises his children to master AI to boost their skills and secure their future careers.

OpenAI's Former Sales Head Joins Venture Capital Firm

Aliisa Rosenthal, who used to lead sales at OpenAI, has joined Acrew Capital as a general partner. She was OpenAI's first commercial hire in June 2022 and helped build the sales team to over 300 people after ChatGPT launched. Rosenthal left OpenAI to work more closely with new ideas and technical decisions, which she missed as the company grew. She still calls herself an "AGI sherpa," helping companies understand and sell AI products by addressing customer fears and building trust. She believes many opportunities exist for startups in enterprise AI applications, even with large tech companies in the market.

Micron and Nvidia Remain Top AI Stock Leaders

Micron Technology and NVIDIA are expected to stay leading companies in the AI stock market. They are vital for building AI infrastructure, which helps them handle market changes. NVIDIA is the main provider of AI-powered graphics processing units, needed for training and running complex AI models. Micron supplies the high-bandwidth memory that powers these GPUs, forming a strong connection between the two companies. The increasing demand for cloud computing, data centers, and AI services further drives the need for their special hardware. Their strong market position and constant innovation make them good long-term investments.

Sources

• Nvidia's Huang heads to China as H200 chip sales hit wall
• China Tells Alibaba, Top Tech Firms to Prep Nvidia H200 Orders
• Nvidia’s Huang to visit China as AI chip sales stall
• Anthropic Says Most Gen AI Use Still Involves Human Oversight
• AI luminaries at Davos clash over how close human level intelligence really is
• As confusion reigns over Jony Ive's iO device, I remain an AI hardware skeptic
• Quantum-Resistant Identity and Access Management for AI Agents
• WealthStack Roundup: NVIDIA AI Survey Results Indicate Growth, Challenges and Steady AI Budgets
• Why Apple and OpenAI are reportedly betting on AI hardware in 2026
• Blunt AI talk
• OpenAI’s former head of sales is entering VC. She still calls herself an ‘AGI sherpa’
• Micron & NVIDIA: The Kings of AI Stocks