Capital & Industrial Strategy

Google closed its acquisition of cloud security firm Wiz for $32 billion, the company's largest deal ever, according to TechCrunch. Shardul Shah of Index Ventures, a Wiz investor, explained the strategic rationale centres on securing enterprise cloud infrastructure as AI workloads migrate to production environments. Wiz's platform provides visibility and threat detection across multi-cloud deployments, a capability Google lacked natively. The acquisition price — more than double Wiz's last private valuation of $15 billion — reflects the premium Google is willing to pay to avoid losing cloud security mindshare to Microsoft and AWS, both of which have been aggressively bundling security tooling into their AI platform offerings.

The deal has cleared regulatory review in the U.S. and EU, with no divestitures required. Google will integrate Wiz into its Google Cloud division, which has been losing market share in AI infrastructure spend to hyperscalers that offer tighter security and compliance guarantees. The acquisition gives Google a credible enterprise security narrative as CIOs evaluate where to deploy AI agents that require access to sensitive data across cloud and on-premises environments.

Google and Accel India reviewed more than 4,000 applications for their Atoms accelerator cohort and found that approximately 70% were 'AI wrappers' — startups with thin layers atop foundation models and no defensible competitive moats, TechCrunch reports. The accelerator selected only five startups, none of which are wrappers. This reflects a broader venture market correction where early-stage AI funding is concentrating on infrastructure, proprietary data, and vertical-specific models rather than application-layer plays that depend entirely on OpenAI or Anthropic APIs. Accel's India managing partner noted that startups without unique datasets, fine-tuning capabilities, or hardware integration are struggling to raise follow-on rounds, even if they achieved early revenue traction.

The filtering rate — 99.9% rejection — is unusually high even for competitive accelerators, and signals that venture discipline is returning after 2024-2025's speculative AI funding boom. Investors are now demanding clear paths to gross margin above 70% and customer lock-in mechanisms beyond prompt engineering. This is particularly acute in India, where AI startups face competition from both U.S. hyperscalers and domestic giants like Reliance Jio, which is subsidising AI API access to capture market share.

Alibaba Group plans to release an agentic AI service for enterprises, banking on national enthusiasm around AI assistants that perform tasks rather than just generate text, Bloomberg reports. The service will allow companies to deploy AI agents that integrate with internal systems, automating workflows in logistics, customer service, and supply chain management. Alibaba is positioning this as a China-specific alternative to OpenAI's agent framework, emphasising compliance with Chinese data residency and censorship requirements. The service is expected to launch in Q2 2026, with pricing tied to API calls and agent deployment volume.

The move reflects intensifying competition among China's tech giants to capture enterprise AI spend before U.S. firms can re-enter the market. Alibaba's Cloud Intelligence division has been losing ground to Tencent and ByteDance, both of which have launched similar agent frameworks in recent months. The Chinese government is also backing this shift through procurement mandates that favour domestic AI platforms, creating a closed ecosystem where Western model providers have limited access. Alibaba's timing suggests it views the next 12-18 months as a critical window to lock in enterprise customers before regulatory barriers potentially ease.

China's second-largest chipmaker is preparing to begin 7nm production, a critical milestone in Beijing's semiconductor self-sufficiency campaign, Reuters reports. The company, whose identity was not disclosed but is widely understood to be SMIC, has been working to replicate TSMC's 7nm process using older DUV lithography equipment rather than EUV machines blocked by U.S. export controls. Initial production volumes will be limited, but the achievement demonstrates that China can produce chips capable of supporting mid-tier AI inference workloads and some training tasks, reducing dependence on smuggled or stockpiled advanced chips. Beijing has directed state-owned enterprises to prioritise purchases of domestically produced semiconductors, even at higher costs and lower yields, as part of a broader industrial strategy to secure AI compute capacity.

The 7nm milestone comes as U.S. export restrictions have tightened further, with the Biden and Trump administrations both expanding controls on chip manufacturing equipment and high-bandwidth memory. China's approach has been to accept performance and cost penalties in exchange for supply chain independence, with government subsidies covering the difference. The strategy is showing results: domestic chip production now meets approximately 60% of China's AI hardware needs, up from 40% two years ago. However, the performance gap remains significant — China's 7nm chips lag TSMC's 3nm and 2nm processes by multiple generations, limiting the country's ability to train frontier models domestically.

BioMap Beijing Intelligent Technology Co. has filed confidentially for a Hong Kong IPO that could raise several hundred million dollars this year, Bloomberg reports. BioMap uses AI to accelerate drug discovery, focusing on protein structure prediction and molecular design. Baidu led the company's Series B round in 2024, viewing it as a strategic bet on AI-enabled biotech. The IPO timing suggests investor appetite for Chinese AI companies remains strong despite broader market volatility, particularly for firms with defensible scientific IP and clear regulatory pathways. BioMap's filing follows successful Hong Kong listings by other AI-biotech firms, which have attracted crossover investors seeking exposure to AI applications with tangible commercial outcomes rather than speculative foundation model plays.

The IPO proceeds are expected to fund clinical trials and expand BioMap's AI model training infrastructure. The company competes with both Western firms like DeepMind's Isomorphic Labs and Chinese rivals including XtalPi, all racing to prove that AI can meaningfully shorten drug development timelines and reduce costs. BioMap's valuation will test whether public market investors are willing to pay premiums for AI-biotech firms before they generate significant revenue from approved drugs, or whether they demand more conservative multiples reflecting the long development cycles and regulatory risks inherent in pharmaceuticals.