Anthropic is in talks to acquire Israel's Decart for $6 billion, it's greatest acquisition to date
Anthropic is reportedly in early-stage talks to acquire Nvidia-backed Israeli AI startup Decart for approximately $6 billion. Decart develops technology that makes AI inference faster and more efficient, as well as real-time video and world models used in simulation. If completed, the transaction would be Anthropic’s largest acquisition and could establish its first substantial development operation in Israel. Why it matters: The strategic value is not simply Decart’s applications. Frontier AI laboratories are under growing pressure to reduce inference costs, improve model performance and control more of their infrastructure. The reported talks suggest that efficiency technology—and the Israeli talent behind it—is becoming strategically important enough to justify multibillion-dollar acquisition prices. The discussions remain preliminary and should not yet be treated as a completed exit.
Wonderful reportedly seeks $500 million at a $5 billion valuation
Israeli AI-agent company Wonderful is reportedly preparing to raise $500 million at a valuation of approximately $5 billion, only three months after completing a $165 million round. The company is expanding internationally and is expected to approach $100 million in annual recurring revenue this year, according to market estimates cited by Globes. Insight Partners is again expected to participate, although the process remains at an early stage. Why it matters: A round of this size would show how aggressively growth investors are rewarding AI companies that combine fast revenue expansion with enterprise distribution. It would also underline the growing split in venture: exceptional companies can raise enormous sums at increasing valuations, while much of the broader market still faces longer fundraising cycles and greater scrutiny. The proposed round is not included in this week’s completed funding total.
Corma emerges from stealth with a $60 million seed round
Corma emerged from stealth with $60 million in seed funding led by Sequoia Capital, with participation from Khosla Ventures and Coatue. The company is building a foundation model specifically for defensive cybersecurity, arguing that general-purpose models are becoming increasingly capable of attacking organisations but remain poorly equipped to defend them. Corma’s agents are intended to operate across defensive workflows rather than address only a narrow security task. Why it matters: A $60 million seed is another example of capital concentrating around technically ambitious AI infrastructure companies. Cybersecurity is also shifting from adding AI features to existing tools towards building models designed around the domain itself. Israel’s security expertise gives local founders a credible advantage, but the size of the round also creates high expectations around technical differentiation, customer adoption and execution.
Team8 raises $365 million for AI-native enterprise companies
Team8 announced $365 million in new capital: $265 million for Team8 Capital III and more than $100 million reserved for follow-on investments in its highest-conviction portfolio companies. The new flagship fund will back Seed and Series A companies across cybersecurity, software infrastructure, fintech and digital health, taking Team8 to nearly $2 billion in assets under management across eight funds. Why it matters: Venture fundraising remains concentrated around established managers with differentiated sourcing and company-building capabilities. Team8’s close demonstrates continued institutional appetite for Israeli enterprise technology, particularly where AI creates new infrastructure, identity, data and security requirements. The follow-on pool also gives the firm more capacity to support winners as rounds become larger and the time between private financing and an exit lengthens.
Israel launches a national AI action plan
Israel formally launched a national AI action plan designed to strengthen sovereign capability and move the country from startup strength towards national-scale infrastructure. The programme includes a National AI Institute, applied “acceleration reactors,” a proposed computing base of 100,000 accelerators, a domestically developed quantum computer and an effort to attract leading global AI laboratories. Why it matters: Israel cannot outspend the United States or China, but it can concentrate resources around areas where it has genuine advantages: talent, cybersecurity, semiconductors, defence, health and applied research. Execution will matter more than ambition. The key questions are how quickly compute becomes accessible to researchers and startups, how programmes are governed and whether the plan attracts international laboratories without crowding out private investment.
Corma emerges from stealth with a $60 million seed round
Corma emerged from stealth with $60 million in seed funding led by Sequoia Capital, with participation from Khosla Ventures and Coatue. The company is building a foundation model specifically for defensive cybersecurity, arguing that general-purpose models are becoming increasingly capable of attacking organisations but remain poorly equipped to defend them. Corma’s agents are intended to operate across defensive workflows rather than address only a narrow security task. Why it matters: A $60 million seed is another example of capital concentrating around technically ambitious AI infrastructure companies. Cybersecurity is also shifting from adding AI features to existing tools towards building models designed around the domain itself. Israel’s security expertise gives local founders a credible advantage, but the size of the round also creates high expectations around technical differentiation, customer adoption and execution.
