Skip to main content

Google Cloud London outage, Exchange patch pain and a DeepSeek ban

  1. Cloud

    Google Cloud London zone outage knocks UK banks and apps offline (opens in a new tab)

    Google's status page says a power event caused an outage in europe-west2-b, one of three London zones, from 06:08 UTC, taking down Cloud Run, GKE, App Engine and more and hitting Lloyds, HSBC, Royal Mail and Revolut. Test multi-zone failover.

    Source: The Register

  2. Cloud

    October Windows Server updates destabilise on-prem Exchange servers (opens in a new tab)

    Microsoft says the default limit of 512 simultaneous Endpoint Lookup calls in its October Windows Server updates may not cope under heavy load, breaking OWA sign-in and mail flow on Exchange 2016, 2019 and SE. Don't uninstall; use Microsoft's registry workaround.

    Source: The Register

  3. Regulation

    US bars DeepSeek from federal civilian agency systems (opens in a new tab)

    DHS made public an order requiring federal civilian agencies to remove DeepSeek products, including third-party tools built on its models, and stop buying them. Vendors selling to the US government should check whether any DeepSeek model sits in their stack.

    Source: CNBC

  4. Security

    ConnectWise suspends Cloud Backup after unauthorised access to app identities (opens in a new tab)

    ConnectWise suspended Cloud Backup (formerly SkyKick) and linked services after unauthorised access to app identities. Huntress says a compromised Microsoft 365 API secret was used to get into customer tenants, and in one a ConnectWise app gave itself Global Administrator. MSPs should audit tenants.

    Source: The Register

  5. Business

    AI chip startup Tenstorrent raises $1.5 billion at $6.2 billion valuation (opens in a new tab)

    Tenstorrent closed a $1.5 billion Series E led by Trace Capital and Innovation X, valuing it at $6.2 billion. Investors are still betting big on alternatives to Nvidia, which could mean more choice and cheaper inference hardware for AI teams.

    Source: Bloomberg

  6. Security

    Ahold Delhaize confirms cyberattack took US and European systems offline (opens in a new tab)

    Ahold Delhaize CEO Frans Muller said pulling IT systems offline after a cyberattack affected most of its brands across Europe and the US; US e-commerce at Food Lion and Stop & Shop should return in the coming days. It follows France's supermarket fridge hacks.

    Source: The Register

  7. Regulation

    EU announces first harmonised standard for AI Act quality management (opens in a new tab)

    The European Commission announced that EN 18286:2026, a quality management standard for high-risk AI systems, is now cited in the Official Journal. Following it gives providers a presumption of conformity with Article 17, ahead of high-risk rules from December 2027.

    Source: European Commission

My take

Today's biggest stories are not about clever new models; they are about the boring layers underneath breaking. A single power event in one London zone took banks offline, a routine Windows patch stalled mail, and a backup vendor's leaked secret became a door into customer tenants. If your resilience plan still treats your cloud zone, your patch pipeline and your SaaS vendors as things that just work, this is the week to test that belief. The DeepSeek ban and the EU's first AI Act standard point the same way: who you depend on is now a compliance question, not just an engineering one.