It’s been a busy week for AI. New models have emerged, new capital has flowed in, government activity has multiplied, and AI’s influence has expanded from medical laboratories to global markets. From Silicon Valley to the Gulf Coast, here’s everything worth knowing from last week’s big AI developments.
1. Anthropic releases Claude Fable 5 with guardrails
Anthropic has released Claude Fable 5. This gives more users access to the company’s cutting-edge public model and tightens controls over its use, including for cybersecurity. The company has also made a more limited Mythos class model available to select cybersecurity and infrastructure partners. Anthropic said more powerful models require additional safeguards because, if left unrestricted, they could aid in vulnerability discovery and offensive cybersecurity efforts.
read more: Anthropic releases Claude Fable 5 and Mythos 5
2. AI designs key components of vaccines
Researchers at the University of Cambridge say they have used AI to design new vaccine components to protect against a wide family of coronaviruses. The study used genetic sequences from coronaviruses considered a potential threat to design “superantigens” that could train the immune system against future variants. Researchers said this approach could help accelerate vaccine development for future pandemics. The claim that this is the first AI-designed vaccine component should only be held if the researcher’s attribution is clear.
read more: World’s first AI-designed vaccine targets future pandemics
3. IPO Confidential Document for OpenAI Files
OpenAI confidentially filed draft IPO documents with the U.S. Securities and Exchange Commission, giving the company the option to go public. The company has not determined the timing and said it may remain private for some time. The filing comes as investor interest in AI companies remains high, with Anthropic moving toward a possible public listing.
read more: OpenAI files sensitive files for IPO, keeps timing open
4. Google signs SpaceX computing agreement
Google has signed a cloud services agreement with SpaceX to secure additional AI computing power for Gemini Enterprise. The deal, revealed in a SpaceX regulatory filing, will see Google pay $920 million per month from October 2026 to June 2029 for access to approximately 110,000 Nvidia GPUs, along with CPUs, memory and related infrastructure. Google said the deal provides bridging capabilities as demand for its enterprise AI agent platform increases. The deal comes as SpaceX prepares for an IPO and follows a separate AI infrastructure deal with Anthropic.
More information: Google signs $920 million per month SpaceX compute deal
5. Dubai’s ‘SME in a Box’ reimagines business setup as an integrated service
Dubai has launched a new initiative to address the complexity of starting and running a business. Through SME in a Box, Dubai Economic and Tourism Authority is creating a more integrated operating environment for small and medium-sized enterprises (SMEs), going beyond traditional business registration support.
The initiative aggregates 18 private sector services into one ecosystem, reducing friction and speeding time to market, rather than requiring founders to use multiple providers individually.
read more: Dubai’s ‘SME in a Box’ reimagines business setup as an integrated service
6. Hub71 startups raise more than $2.7 billion as Abu Dhabi’s tech ecosystem matures
In total, startups in the emirate’s Hub71 innovation ecosystem have raised more than $2.7 billion (Dh9.9 billion) and generated revenues of $1.5 billion (Dh5.4 billion) by the end of 2025, according to a new report. Impact report.
The figures demonstrate the emergence of an innovation ecosystem that is increasingly capable of attracting capital, generating commercial returns and supporting companies through the scale-up phase.
read more: Hub71 startup raises $2.7 billion in funding as Abu Dhabi tech ecosystem matures
7. Most UAE companies overestimate their crisis preparedness, research reveals
Despite their confidence in their resilience, businesses in the Middle East may not be adequately prepared for the next major disruption. According to new the study According to risk management firm Optro, only 19% of organizations in the UAE have a formal disaster recovery plan in place, compared to a global average of 31%. Specifically, less than 40% of organizations reported establishing recovery time objectives (RTO) and recovery point objectives (RPO) for all critical business processes.
read more: Survey finds most UAE companies overestimate their crisis preparedness
8. UAE brings together 50 federal agencies to advance agent AI in government
More than 300 employees from 50 federal agencies participated gathered A specialized workshop will be held in Dubai this week to launch a government-wide effort to transform 50% of government departments, services and operational functions to agent AI models within the next two years.
The program signals the UAE’s intention to put autonomous AI systems at the core of public sector administration. Progress will be monitored intensively over the next three months, with continued support and oversight from government teams.
read more: UAE brings together 50 federal agencies to advance agent AI in government
