Gemini becomes workstation
Google rolled out updates reframing Gemini from an assistant into an 'interactive workstation' with 3D simulations, new open models and developer tools like TorchTPU. The company also drew clearer product boundaries between Looker and Data Studio and is deepening infrastructure ties with Intel around Xeon and IPUs, signalling a push to own both workflow surfaces and the substrate beneath them. (intelligentliving.co) (unwinddata.com) (igorslab.de)
Google has reframed Gemini from a chat assistant into an "interactive workstation" that can generate live 3D simulations, interactive charts and integrate new developer tools. (blog.google) The Gemini app update, published April 9, 2026, lets users ask for “show me” visualizations that produce manipulable 3D models and physics simulations directly inside chat. (blog.google) On April 2, 2026 Google released Gemma 4, a family of open models ranging from Effective 2B to a 31-billion-parameter dense model released under the Apache 2.0 license. (blog.google) Google also published TorchTPU on April 7, 2026, a stack to run PyTorch natively on Tensor Processing Units with an "Eager First" design to simplify migration and scale on Google’s TPU infrastructure. (developers.googleblog.com) At the same time Google clarified analytics branding: the company reintroduced the Data Studio name for the free reporting tool while keeping Looker as its enterprise, governed analytics platform. (cloud.google.com) Industry commentary noted the change was a reversal of the prior "Looker Studio" rename and described the clarification as overdue for practitioners who needed a clear free-versus-enterprise split. (unwinddata.com) Google and Intel on April 9, 2026 announced an expanded multiyear collaboration committing Google Cloud to continue using Intel Xeon processors and to co-develop custom infrastructure processing units, or IPUs. (newsroom.intel.com) Google’s cloud posts show Gemma 4 will be available on TPUs and through Vertex AI, and Intel’s release says Xeon plus IPU co-development will target efficiency and utilization at scale—linking new software surfaces to explicit hardware commitments. (cloud.google.com 1) (cloud.google.com 2) Intel framed the partnership as building “balanced systems” for AI; the company’s statement said continued Xeon use plus custom IPUs will “improve efficiency, utilization and performance at scale.” (newsroom.intel.com) Google says Gemma 4 will be supported on Vertex AI, TPUs, and managed offerings like Model Garden in the coming days, while the Intel agreement is a multiyear effort—expect staged rollouts of both features and infrastructure through 2026. (cloud.google.com 1) (cloud.google.com 2)
