AI Press Review – Thursday, July 17, 2025 OpenAI TSMC WHO Asia China Mira Meta hiring

Today's AI landscape reveals a striking consolidation of power as major tech companies secure unprecedented funding and infrastructure partnerships, while regulatory frameworks take shape across continents. From OpenAI's strategic alliance with Google Cloud to TSMC's record-breaking profits, the industry demonstrates both massive growth potential and growing concerns about market concentration.
1. OpenAI Diversifies Cloud Infrastructure with Google Partnership
OpenAI has officially added Google Cloud to its infrastructure provider roster, marking a significant shift in the company's previously Microsoft-exclusive arrangement 123. The partnership, finalized in May 2025 after months of negotiations, enables OpenAI to access Google's computing resources alongside existing partnerships with Microsoft, Oracle, and CoreWeave 45.
Strategic implications:
Diversification strategy: OpenAI reduces its dependence on Microsoft while securing additional GPU capacity for ChatGPT and API services
Competitive dynamics: Google Cloud gains a major AI client despite direct competition with its own Gemini platform
Infrastructure demand: The partnership underscores the massive computing requirements for training and deploying advanced AI models
The collaboration comes as OpenAI faces capacity constraints, with CEO Sam Altman previously posting urgent requests for "GPU capacity in 100k chunks" 5. The agreement forms part of OpenAI's broader $500 billion Stargate infrastructure program alongside SoftBank and Oracle 2.

2. TSMC Posts Record Profits Amid AI Chip Demand Surge
Taiwan Semiconductor Manufacturing Company reported extraordinary Q2 2025 results, with net profit soaring 61% year-over-year to NT$398.3 billion ($13.5 billion), significantly beating analyst expectations 678. The world's largest contract chipmaker projects continued strong performance with Q3 revenue expected between $31.8-$33 billion 910.
Key performance metrics:
Revenue growth: 38.65% increase to NT$933.80 billion in Q2 2025
Advanced semiconductors: 74% of wafer revenue from cutting-edge 7nm and smaller process nodes
Client portfolio: Serving major AI companies including Nvidia, Apple, and AMD
Despite exceptional results, TSMC faces potential headwinds from U.S. tariff policies, with President Trump's administration considering semiconductor-specific levies 1011. The company has announced $100 billion in U.S. investments to mitigate geopolitical risks 9.
3. China's Zhipu AI Launches Free AI Agent Platform
Chinese AI startup Zhipu AI unveiled AutoGLM Rumination, a free AI agent capable of conducting deep research, web searches, travel planning, and report writing 121314. The platform challenges competitors like DeepSeek while offering services at no cost compared to rivals charging up to $199 monthly 1516.
Technical capabilities:
Proprietary models: Powered by GLM-Z1-Air reasoning model and GLM-4-Air-0414 foundation model
Performance claims: Matches DeepSeek R1 while running 8x faster using 1/30th of computing resources
Accessibility: Available through GLM model website and mobile applications
Founded in 2019 as a Tsinghua University spinoff, Zhipu AI has emerged as a major player in China's AI ecosystem, recently securing three consecutive government funding rounds totaling 300 million yuan ($41.5 million) 1213.

4. Mira Murati's Thinking Machines Achieves $12 Billion Valuation
Former OpenAI CTO Mira Murati's AI startup Thinking Machines Lab closed a massive $2 billion seed funding round, achieving a $12 billion valuation in just five months since launch 171819. The round was led by Andreessen Horowitz with participation from Nvidia, AMD, Cisco, and Jane Street 2021.
Company trajectory:
Talent acquisition: Team includes former OpenAI researchers John Schulman, Barret Zoph, and Luke Metz
Product development: Plans to unveil first product in coming months with "significant open-source component"
Market positioning: Represents one of the largest seed rounds in Silicon Valley history
The funding reflects massive investor appetite for AI startups led by proven executives, with Murati leveraging her experience as OpenAI's former CTO during critical development phases 1822.
5. Pentagon Awards $200 Million AI Contracts to Four Companies
The U.S. Department of Defense announced contracts worth up to $200 million each to xAI, OpenAI, Anthropic, and Google for developing "agentic AI" workflows for national security applications 232425. The contracts aim to accelerate AI adoption across military operations, logistics, and intelligence systems 2627.
Contract details:
xAI launch: Elon Musk's company introduced "Grok for Government" platform targeting federal agencies
Controversy timing: Award comes one week after xAI's Grok chatbot generated antisemitic content
Strategic focus: Pentagon emphasizes AI's role in maintaining competitive advantage over adversaries
The contracts represent a major shift toward commercial AI adoption in defense applications, with officials highlighting AI's "transformative" impact on military capabilities 2325.
6. Meta Announces Massive AI Infrastructure Investment
Meta CEO Mark Zuckerberg revealed plans to invest "hundreds of billions of dollars" in AI data centers to support the company's superintelligence ambitions 282930. The initiative includes two major facilities: Prometheus (1 GW, operational 2026) and Hyperion (scaling to 5 GW by 2030) 3132.
Infrastructure scale:
Prometheus cluster: 500,000 GPUs drawing 1,020 MW of electricity in Ohio
Hyperion facility: 1.5 GW capacity in Louisiana, covering area comparable to Manhattan
Energy requirements: Combined facilities will consume enough power for millions of homes
The investment reflects Meta's strategy to build industry-leading AI capabilities while reducing dependence on external cloud providers 33. The company has also assembled a superintelligence team including former Scale AI CEO Alexandr Wang and other top AI researchers 28.
7. DeepSeek Maintains Cost-Efficiency Leadership
Chinese AI company DeepSeek continues demonstrating remarkable cost efficiency, with its models operating at significantly lower energy consumption than Western competitors 343536. The company's approach challenges fundamental assumptions about AI development economics, potentially democratizing access to advanced AI capabilities 3738.
Efficiency metrics:
Training costs: DeepSeek-V3 model trained for $5.6 million versus Meta's $65 billion AI spending
Energy consumption: 50-75% less power usage compared to Nvidia H100-based systems
Hardware optimization: Utilizes 2,000 chips versus competitors' 16,000+ chip requirements
DeepSeek's innovations have prompted stock market volatility and forced competitors to reconsider their infrastructure strategies 3439.
8. EU Advances AI Regulation Implementation
The European Union released its General-Purpose AI Code of Practice, providing detailed guidance for companies to comply with AI Act obligations taking effect August 2, 2025 404142. Despite industry calls for delays, the EU maintains its implementation timeline 4344.
Regulatory framework:
Transparency requirements: Companies must disclose training data sources and model capabilities
Copyright protection: Enhanced safeguards for copyrighted material in AI training
Voluntary compliance: Initial voluntary period before mandatory enforcement in August 2026
The code represents the world's first comprehensive AI regulation, establishing risk-based compliance requirements for general-purpose AI models 4546.
9. Asia-Pac
ific AI Development Accelerates
Singapore launched new AI initiatives including the Global AI Assurance Sandbox and 800 additional AI training opportunities over three years 474849. The city-state aims to expand its 6,000-strong AI professional workforce while supporting 500 new generative AI projects 47.
Meanwhile, India's AI ecosystem shows remarkable growth with 121 unicorns and over 159,000 recognized startups, driven by the IndiaAI Mission providing GPU access and infrastructure support 505152.
10. WHO Addresses AI in Healthcare at Global Summit
The World Health Organization hosted sessions on AI in healthcare during the AI for Good Global Summit, focusing on traditional medicine applications and global health challenges 535455. The summit emphasized responsible AI deployment in healthcare settings while addressing regulatory frameworks and intellectual property considerations 56.
Healthcare applications:
Traditional medicine: AI-powered analysis of traditional healing practices and compounds
Diagnostic tools: Enhanced screening capabilities for rural and underserved areas
Drug discovery: Accelerated development of new treatments using AI modeling
The initiative reflects growing recognition of AI's potential to address global health disparities while maintaining cultural sensitivity 5556.

Looking Ahead
Today's developments reveal an AI industry entering a new phase of maturation, characterized by massive infrastructure investments, regulatory clarity, and international competition. The convergence of record-breaking funding rounds, strategic partnerships between former competitors, and government-backed initiatives suggests 2025 may be the year AI transitions from experimental technology to critical infrastructure.
Key trends to watch include the outcome of massive infrastructure investments by Meta and others, the implementation of EU AI regulations, and the ongoing U.S.-China technological competition exemplified by companies like DeepSeek challenging Western AI dominance. As the industry consolidates around a few major players while regulatory frameworks take shape, the balance between innovation and responsible deployment will likely define AI's trajectory through the remainder of the decade.
The juxtaposition of OpenAI partnering with Google Cloud despite direct competition, alongside Pentagon contracts worth hundreds of millions, illustrates how AI has become too strategically important for traditional competitive barriers to constrain collaboration. This pragmatic approach to AI development may well characterize the industry's future evolution.
https://www.techzine.eu/news/infrastructure/133125/openai-also-adds-google-as-a-cloud-partner/
https://www.cnbc.com/2025/07/16/openai-googles-cloud-chatgpt.html
https://siliconangle.com/2025/07/16/openai-taps-google-cloud-access-much-needed-gpus-chatgpt/
https://finance.yahoo.com/news/tsmc-q2-profit-surges-60-053818635.html
https://www.ai-buzz.com/zhipu-ai-debuts-free-autoglm-agent-challenges-deepseek
https://finance.yahoo.com/news/mira-murati-thinking-machines-lab-175929212.html
https://techcrunch.com/2025/07/15/mira-muratis-thinking-machines-lab-is-worth-12b-in-seed-round/
https://observer.com/2025/07/mira-murati-ai-startup-raises-2b-prepares-first-product-launch/
https://www.wired.com/story/thinking-machines-lab-mira-murati-funding/
https://defensescoop.com/2025/07/14/pentagon-ai-contracts-musk-xai-google-openai-anthropic-cdao/
https://aimagazine.com/articles/mark-zuckerberg-reveals-100bn-meta-ai-supercluster-push
https://techcrunch.com/2025/07/14/mark-zuckerberg-says-meta-is-building-a-5gw-ai-data-center/
https://www.rcrwireless.com/20250715/ai-infrastructure/meta-ais
https://www.rinnovabili.net/business/markets/deepseeks-energy-consumption-ais-75-power-cut/
https://devops.com/deepseek-disrupts-ai-market-with-off-peak-pricing-model/
https://www.weforum.org/stories/2025/02/open-source-ai-innovation-deepseek/
https://semianalysis.com/2025/07/03/deepseek-debrief-128-days-later/
https://www.nytimes.com/2025/07/10/business/ai-rules-europe.html
https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai
https://www.mddi.gov.sg/newsroom/ai-initiatives-to-transform-life-work-and-business-in-singapore/
https://www.dw.com/en/why-india-has-no-world-class-ai-llm-engine-yet/a-72799462
https://ts2.tech/en/latest-developments-in-ai-june-july-2025/
https://blog.google/technology/ai/google-ai-updates-june-2025/
https://vavoza.com/latest-ai-developments-trending-tools-tech-news-july-17-2025/
https://laotiantimes.com/2025/07/17/automation-anywhere-achieves-the-aws-generative-ai-competency/
https://www.governing.com/workforce/want-ai-to-boost-your-local-economy-its-time-to-strategize
https://www.digitimes.com/news/a20250401PD232/zhipuai-startup-ai-agent-llm-2025.html
https://de.tradingview.com/news/reuters.com,2025:newsml_L8N3TD21U:0/
https://semianalysis.com/2025/07/11/meta-superintelligence-leadership-compute-talent-and-data/
https://www.stephensonharwood.com/insights/neural-network---july-2025
https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/
https://fortune.com/article/fortune-brainstorm-ai-singapore-livestream-2025/
https://www.technologyreview.com/2025/07/04/1119705/inside-indias-scramble-for-ai-independence/
https://opengovasia.com/singapore-enhancing-ai-and-digital-resilience-in-the-financial-sector/
https://www.aimagicx.com/blog/quantum-ai-breakthrough-applications-business-2025/
https://www.linkedin.com/pulse/ai-news-highlights-from-17th-june-2025-grok-ai-wcj5e
https://www.shrm.org/mena/enterprise-solutions/insights/other-tomorrowist-trends-news-july-17-2025
https://www.optica-opn.org/home/industry/2025/july/quix_quantum_poised_to_launch_quantum_computer/
https://fortune.com/2025/07/16/asml-cannot-confirm-growth-in-2026-wiping-out-30-billion/