AI Industry Development & Policy Dynamics

Vercel CEO Guillermo Rauch announced that Context.ai, a third-party AI platform used by Vercel, was compromised. The attackers used Google Workspace OAuth to obtain employees’ account credentials and then gained access to internal environment variables. The investigation shows the attack was carried out in three stages. Vercel has taken emergency measures and provided security recommendations, including reviewing activity logs and rotating environment variables.

Elon Musk was subpoenaed in Paris on April 20, 2026 for a voluntary interview regarding France’s investigation into the social media platform X, mainly concerning his actions involving its algorithm and AI chatbots Grok. The investigation includes alleged conspiracy to possess child sexual abuse material and denial of crimes against humanity. French prosecutors have searched the X company and expanded the investigation to other countries.

Sarvam AI, an Indian startup, aims to raise $320-$350 million, valuing it at $1.5 billion. Major backers include Nvidia and Amazon. Selected as India's national AI champion, it has received substantial support, including 4,096 Nvidia GPUs. Founders have rich experience in AI solutions.

Global AI model call volumes declined to 206 trillion tokens last week. China's calls dropped to 444.1 trillion, while U.S. volumes rose to 490.8 trillion, surpassing China's for the first time in two months. Four of the top nine models are Chinese, with DeepSeek V3.2 second in calls.

AI software editor Cursor developer Anysphere is in talks for a new round of $2.0 billion in funding, with an estimated valuation of $50 billion. Within three years, Cursor grew from zero to $2.0 billion in annualized revenue, becoming the fastest-growing B2B software success story. Nvidia is participating in this round, highlighting the integration of AI infrastructure and showing rising market interest in the AI application layer. Taiwan’s startups can learn from this successful model.

Despite the U.S. NSA banning Anthropic within the Pentagon, it still uses its strong model Mythos, sparking disagreements and skepticism among government agencies. The NSA’s use and the Pentagon’s ban create a self-contradiction, exposing internal inconsistencies in U.S. AI governance. The CEO of Anthropic has met with White House officials to discuss usage boundaries and safety concerns, and in the future may adjust government procurement processes and transparency standards.

David Meyer of Databricks highlights the limitations of top AI models in routine enterprise tasks, contrasting their success in complex problems. Fundamental differences in data types impact performance, leading to a shift towards smaller, efficient models tailored for specific workflows to improve reliability and cost-effectiveness in AI applications.

Decentralized exchange Orca announced that it has completed key rotation and confirmed that users’ funds are safe. This was done because the cloud platform Vercel was attacked. The attack method used a third-party AI tool’s OAuth integration to enter the Vercel system. A supply-chain vulnerability made it difficult for traditional security measures to be detected. Vercel reminded users to review environment variables to strengthen security protections, and noted that encryption projects’ reliance on cloud infrastructure creates a new security risk.

The Claude Haiku 3 model officially stopped service on April 19, 2026. Developers need to update the model ID in their API requests to Haiku 4.5 and note two breaking changes. Enterprises should strengthen AI model lifecycle management to avoid service interruptions caused by model deprecation. It is recommended that developers update their code immediately and monitor changes in costs.

Anthropic’s Claude Opus 4.7 new tokenizer splits the same text into more tokens, increasing input and output costs by 37–47%. Although official rates remain unchanged, businesses need to review contract terms and implement cost monitoring, as a model upgrade may lead to budget overruns. In addition, insufficient transparency will attract the attention of regulators and become a new regulatory concern.

German Chancellor Friedrich Merz urged the EU to ease industrial AI regulations while ensuring consumer oversight, linking this to Germany's €800 billion "Made for Germany" initiative to boost AI in industries and reduce reliance on foreign tech.

Samsung Electronics will hold an installation ceremony for its Texas foundry, aimed at manufacturing Tesla AI chips using 2-nanometer technology. The facility, initially delayed, gained momentum after a $16.5 billion contract with Tesla for AI chips.

Google is collaborating with Marvell Technology to create two custom chips to enhance AI workload efficiency and lessen dependence on Nvidia's GPUs. One chip will be a memory processing unit, while the other will be a new tensor processing unit for AI inference, with designs expected by 2027.

The NSA is utilizing Anthropic's Mythos AI despite Pentagon supply-chain concerns, reflecting internal government divides over its cybersecurity benefits. The model's capability to exploit software vulnerabilities raises global financial system risks, prompting discussions among finance leaders and urging banks to enhance system testing.

Tools for Humanity launched World ID 4.0, enhancing its biometric verification system with key rotation and tiered options. With 18 million users, it integrates with Tinder and Zoom. The company supports expansion through funding, despite a 97% decline in WLD token value.

Canva announced a deepening of its collaboration with Anthropic, enabling drafts from Claude Design to be directly edited in Canva, addressing the problem that AI-generated content is difficult to edit again. This upgrade, called Canva AI 2.0, supports converting AI-generated content into an editable format and introduces an HTML import feature to accelerate website development. In addition, Canva is transforming into an AI collaboration system, with more than 250 million users each month, showing growing market demand.

The “World” iris-scanning project co-founded by OpenAI CEO Sam Altman announced major updates recently. World ID will integrate across multiple platforms to enable real-person digital identity verification, breaking away from the former model that was limited to digital wallets, and strengthening the security of online meetings and signed documents. In addition, World ID has partnered with TInder to offer special discounts to users, and is using Concert Kit to tackle the problem of automated ticket scalpers in the ticketing market.

At a recent Silicon Valley conference, several AI startup CEOs shared their views on the current issues with using AI agents, saying they face two major challenges: token waste and system confusion. Experts noted that companies need to judge more carefully when to use large language models, to avoid unnecessary waste of resources. In addition, the collaboration of multiple AI agents often leads to message-passing and state-consistency problems, indicating that standardization still needs improvement. Although Huang Renxun mentioned token compensation metrics, feedback shows that this does not equal productivity; the real value lies in effective task design.

Dogecoin grows as a payment coin with a strong community and expanding merchant adoption. Tron enables fast, low-cost stablecoin transfers and supports scalable decentralized applications globally. Bittensor combines blockchain incentives with AI training, rewarding decentralized

According to reports, in Q1 2026 the total global venture capital funding is close to $300 billion, of which AI-related companies account for about $242 billion alone, or 80% of venture capital. This shows that AI has become the primary focus of venture capital. As funding concentrates on AI, other areas such as crypto have been squeezed; industry players need to adjust their strategies, integrate AI more deeply into their businesses, and expect a trend of infrastructure consolidation.