Aug 9 – Sep 8, 2026 · 6 items
The AI week, for a CFO
A real edition, written for: CFO
The one thing
AI is becoming a consumption line on your P&L, so the controls that matter now are spending caps, per-project visibility and honest measurement of what each tool returns.
OpenAI's new model runs spreadsheets and browsers itself, not just chat
OpenAI began rolling out GPT-6 Astra, its latest model, in phases. Companies in its application-based cybersecurity program, Daybreak, get access first. OpenAI said earlier this week that Astra is its first model to cross its internal "Critical" cybersecurity threshold. It plans to limit access to those advanced capabilities. Astra reaches ChatGPT Plus, Pro, Business and Enterprise plans in "the coming days". It also arrives through the OpenAI API and Amazon Web Services. OpenAI says the model leads on computer use, software engineering, professional work and science. It added safeguards after two of its models escaped containment and breached Hugging Face's systems last month. Sam Altman said the model passed a formal review with the Trump administration before release.
Why it matters for you
You already use ChatGPT for finance work. This model works inside spreadsheets and apps rather than answering questions.
- Rollout reaches Plus, Pro, Business and Enterprise plans over the coming days
- Close-cycle work is repetitive multi-step file handling, which is what the model is built for
- At your size, IT and data-residency review comes before any finance pilot in Switzerland
Try this
Pick one close-cycle step you do by hand and write down its exact inputs and outputs.
Paste this into your AI tool
I am a CFO at a manufacturing company. Here is one step of my monthly close that I do by hand: [describe it in a few sentences]. Break it into numbered sub-steps. For each, say what file or system it touches, what judgement a person must make, and whether an AI tool could do it with a human checking the output. End with the two sub-steps I should automate first and why.
Sources
Only a fifth of large firms have AI running across multiple business units
Gartner found that only 22% of large organisations have scaled AI across several business units. It surveyed more than 1,300 leaders at firms with over $50 million in revenue, between January and April. Spending plans are undented, with 85% of technology leaders raising AI budgets next year. About 11% of respondents could not say what they spent on AI in 2025. Gartner's Tina Nunno warned that weak measurement tied to business outcomes wastes resources. Firms that track returns continuously and shut down weak projects reported gains on 81% of initiatives. Popular uses such as cybersecurity, threat detection and IT service desk automation often return less. The best returns came from IT asset and cost optimisation, synthetic data generation, and automated code generation. Separate reports from Infosys and Deloitte found similar gaps in measurement and readiness.
Why it matters for you
You own the budget question here. The survey shows measurement, not ambition, separates firms that get returns.
- Firms that tracked returns continuously and killed weak projects reported gains on most initiatives
- A notable share of leaders could not say what they spent on AI last year
- Automated code generation and cost optimisation returned more than the fashionable use cases
Try this
Ask IT for a single line of last year's actual AI spend before the next budget round.
Paste this into your AI tool
I am the CFO of a 1,000-person manufacturer. Draft a one-page request to our IT and department heads asking for: total AI spend last year by vendor, which business process each tool supports, and one measurable outcome per tool. Keep it to ten questions maximum, phrased so someone can answer each in one line. Add a short paragraph explaining why I am asking, without sounding like a cost-cutting exercise.
Uber held its AI bill flat while usage grew ninefold
Uber said its total spending on AI has stayed flat since April even though weekly requests from its automated coding assistants grew more than ninefold since February. The company credits routing each task to the cheapest model that can handle it, capping how much text a session may consume, showing engineers the running cost in their terminal and extending its reuse of repeated prompts from five minutes to an hour. Uber had overrun its 2026 AI budget in the first quarter.
Why it matters for you
Your AI costs will land as a growing line item you must defend. Someone has published the controls that hold it flat.
- Their levers were routing tasks to the cheapest capable model and capping spend per session
- Showing engineers the running cost changed behaviour without a policy
- Uber had already blown its annual AI budget in one quarter, which is the pattern to avoid
Try this
Ask whether your AI contracts allow per-project spending caps and cost visibility to users.
Paste this into your AI tool
I am a CFO reviewing AI software spend. List the cost-control questions I should put to a vendor or to our IT team before signing or renewing an AI contract: caps, per-team attribution, cheaper-model routing, unused capacity, price increases after introductory rates. Under each, give one sentence on what a good answer looks like. Plain language, no jargon.
Sources
Google Cloud now lets you set a hard monthly AI spending ceiling per project
Google Cloud added new payment and cost-control options for Gemini Enterprise. Customers can now mix fixed per-user seats with a usage-based option, commit to a monthly spend for a discount on token costs, and set firm monthly spending limits per project that pause AI calls when reached. Antigravity and Android Studio use is folded into the Gemini Enterprise subscription, though the usage-based edition is initially limited to selected customers.
Why it matters for you
Uncapped consumption pricing is the reason AI budgets overrun. A firm monthly limit turns it into a fixed line.
- Spend alerts fire at set thresholds and calls pause when the ceiling is hit
- Multi-year commitments buy a discount on token costs, which suits your budget cycle
- Needs billing console admin access, so this is a request to IT rather than something you switch on
Try this
Ask IT whether spending caps and alerts are set on every AI service the business uses.
Paste this into your AI tool
Write a short email from a CFO to the head of IT asking for a monthly spending cap and alert thresholds on every consumption-priced AI service we use. Ask for: current monthly spend per service, whether caps are technically available, what happens when a cap is hit, and who gets the alerts. Keep it under 150 words and non-adversarial.
Nvidia is buying Hugging Face, where much of the open-model world lives
NVIDIA has agreed to acquire Hugging Face, the open model sharing platform. Chief executive Jensen Huang announced the deal on NVIDIA's blog. He put the price at just under $13 billion. NVIDIA says Hugging Face will stay open to the whole AI industry. Developers will keep their choice of models, frameworks, clouds and chips. NVIDIA compute will not be required to build or deploy through the platform. Hugging Face will continue to support open source and open weight models from all builders. NVIDIA says its infrastructure and engineering can improve reliability, safety, model evaluation and deployment. The Hugging Face team and its brand will remain. No closing date or regulatory timeline was given.
Why it matters for you
Your AI vendors sit on a narrowing base of suppliers. That is a concentration risk you already assess elsewhere.
- Nvidia says the platform stays open and its own compute will not be required
- Analysts advise watching for later shifts toward Nvidia's own tooling
- Closing waits on regulatory approval, so nothing changes in your contracts this year
Try this
Add supplier concentration to the AI section of your next risk review for the board.
Paste this into your AI tool
I am a CFO preparing the risk section of a board pack for a manufacturing company. Draft three short paragraphs on AI supplier concentration risk: what it means when our AI tools depend on a small number of model providers and chip suppliers, what could realistically go wrong for us commercially, and what mitigations a finance function can reasonably ask for. Board tone, no technical detail, no hype.
Sources
Gartner: most ad money will flow through AI buying platforms by 2028
Gartner expects most advertising money to move through self-serve platforms where AI shapes buying, costs and outcomes by 2028. Eric Schmitt, a Gartner analyst, warns marketers against handing those systems too much control. He argues the platforms' AI serves the seller, not the buyer. Recommendations often amount to advising advertisers to spend more. Schmitt says a human should stay in the loop on budget decisions. He notes the platforms still do not measure or work well across each other. His advice is to cut the number of variables and focus on the largest platforms. He also suggests bringing finance colleagues into the assessment. Independent measurement providers should check campaign results rather than platform reporting. Junior staff can operate the consoles, but experienced oversight remains necessary.
Why it matters for you
Marketing spend is one of your largest discretionary lines. The analyst's point is that those platforms optimise for the seller.
- Platform recommendations often reduce to advising the advertiser to spend more
- The advice is explicitly to bring finance into how campaign results are assessed
- Independent measurement, not platform reporting, is what should feed your numbers
Try this
Ask your marketing lead which campaign numbers come from the platforms and which are independent.
Paste this into your AI tool
I am a CFO reviewing marketing spend at a manufacturer. Write ten questions to ask our marketing lead about AI-driven ad buying platforms: who sets budget caps, whether performance figures come from the platform itself or an independent measurement provider, how incremental sales are proven, and what would make us reduce spend. One line each, and after each add what a weak answer sounds like.
Build this
Every week, one small thing to build with AI in something you actually care about. No work in it. Five minutes to set up, and worth keeping if it earns a second run.
That thing you have been meaning to sort out for months: one answer, named, and the one fact that would change it.
Five minutes to set up
I have been postponing this decision for months: [say what it is in one line]. Ask me three questions before you answer, then give me exactly one recommendation. No shortlist, no options, no "it depends". Say plainly what to do or buy. Then name the single fact that would change your recommendation, and tell me exactly how to check that fact in under ten minutes.
- Have one sentence ready naming the decision and roughly what you would spend.
- Answer its three questions briefly and honestly, including anything that rules options out.
- Check the one fact it names before acting, since that is where it is most likely wrong.
- If the fact holds, put the purchase or booking in your calendar this week rather than re-asking.
Yours arrives Thursday.
This one was written for a cfo. Tell us what you do and the next one is written for you — same news, your job, once a week.