May 22, 2026 • 4 min read
ZYMP Tech News — May 22, 2026
Alibaba Unveils Zhenwu M890 AI Chip, Surpassing Nvidia H20 Performance
HARDWARE
Alibaba has launched its new Zhenwu M890 artificial intelligence chip, claiming it delivers three times the performance of its predecessor, the Zhenwu 810E. The previous model is widely believed to match the capabilities of Nvidia’s H20 chip, the export version of Nvidia’s AI processing units designed specifically for the Chinese market. The timing of Alibaba’s release is particularly significant as prospects for Nvidia’s more advanced H200 chip entering China remain uncertain due to ongoing export restrictions.
The Hangzhou-based technology giant has shipped more than 560,000 chips in its Zhenwu series to over 400 customers, including major automakers and financial institutions. Industry analysts note that the new chip provides a domestic alternative that is insulated from fluctuations in US export controls, enabling Chinese AI companies to formulate long-term technology roadmaps. While Nvidia still maintains performance leadership with its Blackwell series and forthcoming Rubin processors, Alibaba’s advancement demonstrates China’s growing capabilities in high-end AI computing hardware.
AI Will Help Make Nobel Prize-Winning Discovery Within Year, Says Anthropic Co-Founder
ARTIFICIAL INTELLIGENCE
Jack Clark, co-founder of Anthropic, predicted that an AI system will work with humans to make a Nobel prize-winning discovery within the next 12 months. Speaking at Oxford University, Clark described a “vertiginous sense of progress” in AI development and made several additional predictions. He forecast that companies run entirely by AIs would generate millions of dollars in revenue within 18 months, and that by the end of 2028, AI systems would be capable of designing their own successors.
Clark acknowledged that there remain plausible scenarios in which AI technology poses existential risks, stating there is “a non-zero chance of killing everyone on the planet.” He emphasised the importance of clearly stating that this risk has not disappeared. While suggesting it would be better if humanity could slow AI development to have more time to address its implications, Clark noted this is unlikely to happen given the competitive pressures among nations and commercial actors. Anthropic, valued at approximately $900 billion, recently launched a model called Mythos that demonstrated concerning capabilities in identifying cybersecurity weaknesses.
SpaceX Files Record-Breaking IPO Prospectus, Eyes $1.7 Trillion Valuation
STARTUPS
SpaceX has submitted its S-1 filing to the Securities and Exchange Commission, setting the stage for what could become the largest initial public offering in history. The company plans to trade under the ticker “SPCX” and reported 2025 revenue of $18.7 billion, representing a 33% year-over-year increase. Starlink’s satellite broadband service generated $1.2 billion in quarterly profit, while the launch and AI-related divisions posted losses. The company’s accumulated deficit stands at $41.3 billion, with a Q1 2026 net loss of $4.27 billion.
Elon Musk will retain 85% voting control through Class B shares. His proposed compensation package includes 1 billion performance-based shares tied to achieving the establishment of a permanent human colony on Mars with one million inhabitants. The filing reveals significant AI integration, including a $60 billion option to acquire coding startup Cursor and $1.25 billion in monthly payments to Anthropic for cloud computing services. This development positions SpaceX at the intersection of space infrastructure, satellite broadband, and artificial intelligence.
Nvidia Forecasts $91 Billion Q2 Revenue, Unveils $80 Billion Stock Buyback
BIG TECH
Nvidia has forecast $91 billion in second-quarter revenue, exceeding Wall Street expectations, and announced an $80 billion stock buyback programme. CEO Jensen Huang emphasised the company’s focus on the next generation of data centre chips, highlighting Nvidia’s data centre business as the clearest public indicator of global AI infrastructure demand. Huang introduced the Vera central processor as a major new growth engine targeting a $200 billion market opportunity, with expected revenue of $20 billion by the end of the fiscal year.
The company projects a $1 trillion opportunity tied to its Blackwell and Rubin AI chips through 2027. Nvidia’s increasingly diversified customer base and expanding product lineup are positioned to sustain momentum amid growing concerns about a potential AI spending bubble. The update reinforces Nvidia’s dominance in AI infrastructure while underscoring its strategic push beyond GPUs into CPUs for data centres, where demand for accelerated computing continues to expand. Competition is intensifying as hyperscalers including Google, Amazon, AMD, and Intel race to develop their own custom silicon.
Google Launches Gemini Spark AI Agent for Everyday Tasks
SOFTWARE
Google has introduced Gemini Spark, an always-on AI agent designed to assist with everyday tasks, during its I/O 2026 developer conference. The agent can run in the background, work across Google Workspace applications, connect to third-party software, and eventually interact with local files on user devices. This launch represents Google’s clearest strategic response to the growing agent ecosystem, positioning the company to compete directly with specialised agentic AI platforms.
With Gmail, Docs, Search, Android, and Google Cloud under one corporate umbrella, Google possesses the integrated distribution infrastructure necessary to make AI agents genuinely useful beyond demonstrations. The announcement reflects Google’s broader strategy to embed artificial intelligence throughout its product ecosystem rather than treating it as a standalone feature. Industry analysts note that AI agents are shifting from novelty status to becoming core platform strategies for major technology companies.
ZY Media Productions
IT • Music • Technology