Tech brief
Custom Silicon and Monetized Agents Reshape Enterprise Cloud
AWS tops $25B in proprietary chip revenue while Meta prepares paid WhatsApp AI agents and Google embeds live 3D models.
Tech
Google is converting Gemini into an interactive visual workspace — shifting enterprise AI past static text responses into live, variable-driven simulations.
BackgroundLarge language models traditionally generated static text, code snippets, or flat images during user sessions. Adding interactive real-time visual rendering allows engineers and researchers to modify variables directly within generative output.
- The interactive simulation feature is enabled by default across Google Workspace enterprise domains, including Israeli tech hubs and academic centers.
- Domain administrators retain policy controls through standard Google Workspace Generative AI management settings to restrict external data sharing.
- The update targets enterprise financial planning, industrial engineering workflows, and advanced scientific research environments requiring dynamic visual modeling.
Tech
Meta is turning WhatsApp into a paid agent interface — betting businesses and power users will pay monthly fees for AI that executes real-world actions.
BackgroundMeta previously offered free conversational AI tools across its social platforms to drive user engagement and ad targeting. Developing paid transactional agents represents a strategic pivot toward direct consumer and enterprise subscription revenues.
- Hatch is engineered to interact directly with external services including DoorDash, Etsy, Reddit, and Outlook, executing automated tasks inside messaging threads.
- Meta plans to follow the Hatch launch with its next flagship foundation model, codenamed Watermelon, scheduled for release in October.
- Tiered pricing targets power users and commercial accounts while keeping standard conversational AI capabilities free across consumer messaging apps.
Tech
AWS is turning custom silicon into a $25B revenue engine — locking in enterprise cloud clients by controlling both the data center hardware and the chip.
BackgroundCloud hyperscalers have invested billions in custom chips to cut reliance on Nvidia and lower operating costs in AI data centers. AWS built its Trainium and Inferentia chip lines across multiple generations, supported by expanding R&D teams in Tel Aviv.
- Amazon expanded its 2026 capital expenditures to $220 billion to fund data center land, power capacity, and server buildouts worldwide.
- AWS officials confirmed that hardware capacity constraints will persist into 2027 despite massive capital investment, affecting enterprise workload deployments.
- Custom chip growth is shifting cloud profit margins away from third-party vendors toward in-house architectures, reshaping how enterprise infrastructure budgets are allocated across global tech hubs.
Unlock the full brief
Sign in to read every signal, takeaway, and source. Free account — Apple, Google, or email.
