S&P 100 CEOs on AI: who actually moved this month
Findings · Jul 2026 · 6 min
We keep a standing watch on what the biggest US public-company CEOs actually say and do about AI inside Aim — in their own words, not the press releases. This past month the signal wasn't rhetoric, it was money and machinery: Microsoft committing roughly $190 billion of capex while trimming staff, Jensen Huang doubling his demand forecast for Nvidia's next chips, and Sanjay Mehrotra watching AI reprice Micron's entire memory business. Here's who moved, and what it means.
Microsoft spends $190B on AI while cutting 4,800 jobs
Microsoft is running the year's clearest split-screen. It plans about $190 billion in capital spending this year, mostly for AI data centers, and in the same stretch cut 4,800 jobs, a little over 2% of its staff — with about 1,600 of those cuts coming from Xbox.
The company was careful to decouple the two stories — Chief People Officer Amy Coleman told staff the eliminated roles are not being directly replaced by AI — but investors are watching the capex line after the stock shed about $1.2 trillion in market value from its October peak.
Jensen Huang doubles Nvidia's demand forecast — and opens a Japan front
Nvidia's CEO spent the month making the buildout feel inevitable. Beth Kindig's bull case leans on a number Huang himself put forward: over $1 trillion in cumulative demand for Nvidia's upcoming Blackwell and Vera Rubin systems through 2027 — a figure that doubled earlier projections of around $500 billion. Nvidia's fiscal 2026 revenue already reached $215.9 billion, up 65% year over year.
He also opened a new front. In Tokyo, Huang called physical AI the next frontier and a once-in-a-generation opportunity for Japan, lining up Fanuc, SoftBank, Sony and others behind Nvidia's Cosmos models and detailing a Vera Rubin AI factory due in 2028 with 13,750 Vera CPUs and 27,500 Rubin GPUs.
Computing demand is growing exponentially.
Sanjay Mehrotra says AI turned memory into a strategic asset
Micron's CEO framed the boom as a re-rating of his product: Sanjay Mehrotra said AI has recast memory as a defining strategic asset in the AI era. The numbers back it — Micron reported fiscal Q2 2026 revenue of $23.86 billion and guided Q3 to $33.5 billion at roughly 81% gross margin, against $9.30 billion in the year-ago quarter.
Supply is the tell. HBM is sold out through 2026, with Micron able to fulfill only 50% to two-thirds of key customer demand, and the company now guides capital spending above $25 billion for fiscal 2026, with another step up in 2027. Management expects supply constraints to persist well beyond calendar 2026.
Salesforce buys Fin and folds it into Agentforce
Salesforce put a concrete deployment number on agentic AI. It announced it will welcome Fin — and founders Eoghan McCabe and Des Traynor — to Salesforce later this fiscal year, citing an agent that autonomously resolves 76% of customer support volume end to end across more than 30,000 companies worldwide.
The logic is momentum: Agentforce has reached $1.2 billion in annual recurring revenue, up 205% year over year. Bolting on a service agent that already works gives Salesforce a faster on-ramp for smaller customers.
Apple drops its cash discipline to shop for AI chips
Apple signaled a genuine strategy shift. CFO Kevan Parekh said the company is moving away from its long-running 'net cash neutral' policy to pursue chip-company acquisitions that accelerate its AI server chips — a break for a company that historically prefers deals in the hundreds of millions. It already agreed in January to buy Q.ai for close to $2 billion, its second-largest deal ever.
The urgency is technical: Apple's in-house M2 Ultra chips have not proved sufficient for advanced AI workloads, pushing it to rely on Nvidia chips hosted in Google Cloud while its own Baltra server chip slips past its planned ship date. All of this lands as John Ternus is set to succeed Tim Cook as CEO in September.
Adobe's data shows AI shoppers are now the best customers
For all the capex talk, Adobe offered the month's clearest evidence the spend is converting into behavior. Its analytics team found shoppers referred by AI assistants convert 58% more often than regular visitors, and during Prime Day AI-referred traffic grew 98.3% year over year.
The trend line is the story. AI traffic to US retail sites grew 393% year over year in Q1 2026, and AI-referred visits flipped from converting 38% worse than other channels in March 2025 to 42% better in March 2026 — what Adobe's Loni Stark frames as generative AI becoming a primary tool for product discovery.
Moves worth logging
Google is months behind on Gemini 3.5 Pro, with rivals Anthropic and OpenAI shipping models described as exceeding its capabilities — an awkward stumble for a $4.2 trillion company.
Cisco's Chuck Robbins called the shift to AI a change 'fundamentally larger than the advent of the internet' and warned it can't be run as a side IT project.
NextEra and Dominion's $67 billion merger entered a 180-day Virginia review, with Dominion CEO Robert Blue citing 'unprecedented load growth' from data centers.
Duke Energy laid out a $103 billion 2026-2030 capital plan as it flags data-center power demand growing about 10x the historic pace.
What it means
The month's throughline is that AI has stopped being a talking point and become a capital-allocation decision — Microsoft, Micron, Apple and Nvidia are all reorganizing balance sheets and org charts around it.
The open question for the back half of 2026 is whether the demand signals — Micron's sold-out memory, Adobe's converting shoppers — keep validating the spend, or whether names like Google widen the gap between buildout and payoff. We'll keep tracking what these CEOs say next.