AI's costs are landing on investors, workers and buyers at once, and companies that measure and control their own AI will carry them best.
The AI Transformation Brief
AI's bill is coming due, and it is landing on investors, workers and buyers at the same time. OpenAI cut its funding target by $10 billion as AI stocks fell. Moneycontrol Goldman Sachs found that workers who lose AI-exposed jobs stay unemployed longer. Fortune Mistral released its largest AI for companies to run on their own computers. Mistral The companies that measure what AI returns and control where it runs will set the terms.
OpenAI is now seeking about $20 billion at a $1.25 trillion valuation, down from $30 billion at $1.4 trillion last month, and some investors reduced their commitments, the Financial Times reported. Moneycontrol A group of UAE investors that includes state-backed MGX is expected to be the largest participant. Fidelity On October 30 the Nasdaq fell 2.9%, its largest one-day drop since June, after disappointing results from Amazon and Intel. The Business Times
Investors are now asking what buyers have asked all month: what does the spending return? Oliver Pursche of Wealthspire Advisors said companies that miss "the very lofty expectations" are "getting punished." The Business Times Suppliers under pressure to show profits will raise prices, trim free credits and push customers toward longer contracts. Before signing a multiyear AI contract, insist on price caps, a right to leave and a free way to move your records to another provider.
Goldman Sachs economist Pierfrancesco Mei studied more than 15,000 displaced workers who lost jobs between 2016 and 2025. Those displaced from AI-exposed jobs since 2025 took about a month longer to find work and were roughly 10 percentage points more likely to be unemployed a year later. Workers 35 and older took about five weeks longer to return to full-time work. Mei called the damage from earlier waves of technology change "long-lasting." Fortune
The cost of cutting people does not end with severance. In a February Careerminds survey of 600 HR leaders, 35.6% of companies that made AI-led cuts had already rehired more than half of those roles, and 30.9% spent more rehiring than they saved. Careerminds The experienced workers who struggle most to find new jobs often hold the knowledge AI cannot reach. Before cutting a role, test whether AI can do the whole job, and plan retraining for the people affected.
The company that cuts a role answers for whether AI can really do the work. A named manager should check the results for months, and leaders should own the cost of rehiring if the AI falls short.
On October 29, French company Mistral released a downloadable version of its largest AI, built for coding, cyber defense and business tasks. Companies can inspect it, adjust it and run it on their own systems without depending on Mistral to host it. Using it, changing it and building products on it are free; businesses that sell access to the AI itself need a separate license. Companies that prefer a managed service can use it through Microsoft Foundry. Mistral
Running AI in-house gives a company something no hosted service can: no supplier can restrict it or switch it off. Companies still need the right hardware and expertise to run it securely. Euronews With US suppliers under investor pressure, a credible way to run AI yourself is bargaining power, even if you never use it. Test one sensitive task on a self-run AI this quarter, and compare its cost with your hosted contract.
Palmer's October 30 post describes Gemini letting people create and edit documents, spreadsheets and presentations inside the chat, so the chat window becomes the workspace. He notes that an editable document still needs an owner who decides when it is ready, and that the feature has no separate administrator control. Shelly Palmer That owner question runs through every story today.
Treat AI as a cost that must prove itself and a dependency that must be managed.
Ask for a one-page view of AI spending, results, and which supplier could switch your work off.
Investors are now marking down AI suppliers that spend without clear returns.
Expect suppliers to chase profit through higher prices and longer contracts, while downloadable AI gives buyers a credible alternative.
Before cutting any role for AI, run the AI on the whole job for 90 days with a named reviewer.
Keep the experienced staff who hold knowledge the AI cannot reach.
Add price caps, exit rights and a free way to move your records into every AI contract renewed this year.
Count rehiring and retraining costs in every AI savings case.
The shift this edition is that AI's costs now land on investors, workers and buyers at once. The assumption it breaks is that AI spending pays for itself; investors are marking it down, and companies rehiring the people they cut are paying twice. The decision it forces is control: measure the return, keep the people who hold the knowledge, and keep a way to run AI without your supplier.
If your main AI supplier doubled its price next year, what would you do?
Make AI accountable. Measure the results. Redesign the business.