Strategy
226 analyses
Google Cloud and OpenAI announced GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna in a private preview under existing cloud commitments. Pricing, regions, and general availability were not disclosed.
In Berlin, Jensen Huang introduced the DSX POD, a modular project that, according to Nvidia, shortens the deployment of an AI factory from six months to 14 days, provided energy and utilities are ready.
OpenAI models enter private preview in Google's Model Garden, purchasable through the Marketplace and on the same cloud invoice. This is the third major cloud to sell GPT since the end of exclusivity with Microsoft.
The American bank will centralize work in Hyderabad, previously divided among various Indian vendors, under Infosys' Build, Operate and Transfer model. Integration of over 1,000 professionals is expected in October.
Anthropic's own report reveals a shrinking entry point within the lab. In India, TCS still hires fresh graduates, but the data raises questions about the pyramid supporting integrators.
DeepSeek has launched an MIT-licensed version of V4.1-Flash, charging $1.20 per million output tokens at peak times and $0.75 per thousand calls for agents.
Japanese bank launches bootcamp to turn sales staff into developers; HSBC considers cutting 70% of consultants in the UK. The difference lies in who pays for the transition.
Anthropic's fifth Asia-Pacific office arrives with a MoU for model assessment and incident response, coinciding with the EU's information requests to the company.
Canadian lab launches Cohere Cyber division in Prague, acquiring AI startup AISLE's team for undisclosed value.
Michael Roberts, head of HSBC's corporate and institutional bank, stated in Bangkok that agents will trade markets with humans overseeing them. According to Bloomberg, the bank may cut around 20,000 jobs due to AI.
Brian Moynihan revealed that Bank of America hired 1,000 contractors to address software flaws exposed by AI like Mythos, amidst a workforce reduction of nearly 3,500 employees over the year due to automation.
Nvidia published the code for its 2024 acquisition, enabling GPU partitioning and team quotas. Neoclouds like Nebius and Crusoe already integrate the open version.