Greetings,

Forecasting AI spend remains a major challenge for businesses. Research shows only 11% of companies can accurately predict their AI costs, and many are still trying to zero in on the right models for their needs. More on that below.

Also in the news:

• Global M&A falls below $1T as dealmaking slows
• Trump plans order to cut diesel costs ahead of midterms
• KPMG: CFOs revamp talent strategies for AI-driven finance
• IBM study: AI-first CFOs drive 23% higher revenue growth

Businesses are finding it increasingly challenging to budget for AI usage as it behaves more like a human worker, taking actions and making decisions that can lead to unpredictable costs. A study by Stanford University, Carnegie Mellon University, the University of California at Berkeley, and Microsoft Research highlights that only 11% of companies can accurately forecast AI spending, with cheaper AI models sometimes incurring higher costs due to inefficiency.

Global dealmaking fell to $986 billion in the third quarter, down from nearly $1.7 trillion in the previous quarter and 13% from a year earlier, according to LSEG. Higher interest rates, uncertainty around AI and geopolitical risks have made companies more cautious, while the number of deals exceeding $10 billion dropped sharply. Still, advisers say companies remain interested in pursuing transactions.

President Donald Trump is set to sign an executive order aimed at reducing diesel costs, a significant concern for farmers and ranchers during harvest season. The order will include a Treasury Department review of diesel taxes, increased availability of tax-exempt red-dyed diesel and urging states to waive certain fuel taxes. The move comes as high fuel prices, driven by the war against Iran and Ukrainian attacks on Russian refineries, have contributed to inflation and political challenges for Republicans ahead of the midterms.

CFOs are revamping talent strategies as AI reshapes finance careers, automating entry-level tasks and creating a divide between AI-native hires and experienced employees. Some companies are addressing this by implementing AI fluency training, rapid job rotations and business-first training for new hires.

AI-first chief financial officers are driving significant value creation, with a recent study showing that firms with AI-first CFOs saw revenue growth rates that were 23% higher than firms without AI-first CFOs. The study identifies five key disciplines for AI-first CFOs: shaping enterprise advantage, scaling AI with trust, embedding intelligence into decisions, enabling dynamic capital allocation and building future options.

Executives from Fidelity, Robinhood, Alpaca and Cahill said tokenized markets will need deeper liquidity, stronger infrastructure and greater interoperability with traditional markets before broader institutional adoption can take hold. Speaking at SIFMA's Digital Assets Conference, panelists highlighted price discovery, hedging, custody and connectivity between tokenized and traditional securities as key requirements, while also pointing to growing demand for 24/7 trading and expanded access to U.S. assets. Fidelity's Cynthia Lo Bessette said the market has not yet reached a "minimum viable ecosystem."

Private credit firms are preparing for a potential surge in borrower distress, with 80% gearing up to take control of troubled holdings, according to market participants. Firms are ramping up restructuring skills, anticipating a spike in borrower stress in Q4 2026 and beyond. Lincoln International reports a significant increase in lender takeovers in Europe this year, with most planning to rebuild assets rather than sell. The US is also seeing a rise in lender-control cases, with $22.3 billion in pre-takeover principal foreclosed in H1 2026.

Don't miss out on your chance to enter SmartBrief's 2026 Innovation Awards in AI, which recognize AI solutions that impact business operations. The awards focus on recognizing organizations that harness AI to tackle complex challenges, improve decision-making and create efficiencies across diverse industries. Entry window closes this Friday, Oct. 9. Enter now.