Gustavo’s Corner: AI News for CFOs - #20
This edition of Gustavo’s Corner examines how AI is advancing from standalone tools into systems capable of solving complex research problems, completing online tasks, managing business communications, consolidating workplace applications, and supporting self-hosted deployment. Together, these developments highlight the growing capabilities of AI while raising important questions around cost, integration, data access, infrastructure, and geopolitical risk.
OpenAI’s Astra model has reportedly solved 10 open problems across mathematics, quantum complexity, and theoretical computer science, with the results documented in a 249-page paper. The total estimated token cost was just $2,000. Before Astra’s work, none of the problems had seen meaningful progress on their main results for at least a decade. Sam Altman was also reportedly in Washington, DC, last week demonstrating Astra to federal officials.
Google’s Gemini Spark can, with the user’s permission, complete online tasks through logged-in accounts and saved passwords. This includes researching flights, searching for apartments, and handling other browser-based activities through its auto-browse capability. In a separate update, Google has made Gemini easier to access on macOS, allowing users to summon the assistant at any time by pressing the FN key in the Gemini app.
WhatsApp is testing a separate “Offers & Updates” folder designed to reduce business-message clutter in the main chat list. Messages from large organizations, including banks and airlines, would automatically move into the folder after a set period. Business messaging is becoming increasingly valuable: Meta reported that Family of Apps “other revenue” reached $1 billion for the first time last quarter, an increase of 73%. The challenge is keeping those messages commercially valuable without overwhelming users.
Microsoft Copilot is set to become a unified super app for consumers and businesses. Satya Nadella confirmed that Copilot Chat, coding tools, Cowork, and Autopilot agents will be consolidated into a single experience this quarter, although no launch date, pricing details, or platform list were provided. Fortune first reported the project in May. The move follows OpenAI’s introduction of ChatGPT Work, which incorporates its Codex coding tool. Microsoft also reported $90 billion in quarterly revenue and more than 30 million paid Copilot seats. After launching four separate Copilot products, Microsoft now appears to be responding to demand for one simpler, integrated platform.
Moonshot AI’s Kimi K3 is now available as the largest open-weight AI model released to date. Moonshot published the model’s full weights and technical report on Hugging Face, allowing developers to download and self-host the 2.8-trillion-parameter mixture-of-experts system. Although the license permits free use, large hosting providers must negotiate commercial agreements, and one developer reportedly needed 80 RTX 5090 GPUs to run the model. The weights may be open, but the infrastructure entry cost remains substantial.
Why this news is important to CFOs and their teams:
This news shows that AI decisions are becoming broader financial and strategic choices rather than isolated technology investments. Leaders will need to evaluate productivity gains alongside licensing models, infrastructure costs, cybersecurity, vendor concentration, regulatory exposure, and the complexity of integrating AI into existing workflows. The organizations that establish clear governance and investment priorities now will be better positioned to capture AI’s value without losing control of cost and risk
—
/Gustavo
