AI Giants Tighten Corporate Grip as Costs Surge

Anthropic’s new enterprise reports will tell managers what teams produced, what it cost and where workflows broke down. The reports turn employee surveys and internal analytics into another way to measure work. The company says its tools will recommend ways to improve results, and they won't use data for model training by default. For workers, the immediate question is who gets to define productivity—and who gets watched while someone calculates the answer.

“It's not enough to just spend a bunch. You now have to optimize it, make it work,” Nicholas Arcolano, Jellyfish's chief scientist, told TechCrunch. Spending on agentic coding tools has grown 10x since the beginning of the year, reaching about 2% of total engineering costs. Arcolano said “most CFOs didn't plan for” that growth.

The workplace becomes a dashboard

Anthropic is also creating pooled group budgets to help teams manage token spending. The tools aim to help customers assess AI’s return on investment, a concern among large businesses. Management gets reports, cost controls and recommendations. The people whose work gets measured become the subject of those reports.

Jellyfish reported on agentic coding tools used by 276,000 customers. Claude Code and other tools built on Anthropic models ranked as the most widely used in its survey, accounting for nearly a third of proposed code changes submitted for review. Anthropic said its coding tool, Claude Code, recorded an operating profit in the second quarter. Net revenue retention was 900%: spending by the same customers in Q3 2025 grew nine times in Q3 2026.

The adoption boom isn't evenly shared. Venture-backed companies drove token spending, while most Fortune 500 companies weren't seeing major gains from AI deployments. Nikolas Albarran, Jellyfish's AI product lead, said companies that have standardized workflows around Anthropic tools face more difficulty switching providers than individual programmers or small organizations. “Once you're in, it's going to be harder and harder to switch because you've spent the effort,” he said. The tools differ in small ways, he added, and “that cost to switch is not zero.”

The article describes no grassroots or worker-led response to this corporate rollout. Instead, it focuses on companies trying to measure, budget and optimize AI use, with Anthropic’s reports presented as a lighter version of consultants embedded with customers. That's a market solution to a market problem: more tools to help management make the tools pay.

Cloud contracts, receipts and record revenue

Microsoft reported $10 billion in revenue from commercial arrangements with OpenAI in the quarter ended Sept. 30, according to a regulatory filing on Wednesday. That represented nearly 11% of Microsoft’s quarterly revenue of $92.3 billion. The filing didn't make clear how much came from revenue-sharing payments and how much from Azure cloud services.

Microsoft reported $7.2 billion in receivables from OpenAI at the end of September, compared with $6 billion at the end of June. Those receivables covered amounts owed for earlier transactions, not additional revenue. OpenAI-related revenue totaled $24.1 billion for Microsoft’s full 2026 fiscal year; at the latest quarterly pace, it would amount to $40 billion annually.

Microsoft’s quarterly revenue hit a record $92.3 billion, up 19% from the year-earlier quarter, while net income reached $36.6 billion, up 32% on a GAAP basis. Microsoft Cloud revenue was $63.9 billion, up 30%. Capital expenditures topped $45.2 billion, an all-time quarterly record. Microsoft said its “commercial remaining performance obligation increased 73% to $677 billion,” representing future revenue committed under existing contracts of AI labs.

The companies began loosening their once-exclusive partnership in October 2025. OpenAI committed to buy an additional $250 billion in Azure services under an agreement that year, while an April 2026 agreement required revenue-sharing through 2030, subject to a cap. Microsoft’s ownership stake in OpenAI declined from about 25% in June to 24% in September; Microsoft didn't specify the reason for the dilution.

Wall Street weighs the winners

A Bloomberg Evening Briefing item said technology-company results were making investors nervous after years of concern that trillions of dollars in AI investments might not pay off. Alphabet, Google’s parent company, reported quarterly sales and profit below expectations. Its shares fell 7%; revenue grew nearly 25%, Google Cloud sales more than doubled, and free cash flow was negative $9.1 billion.

Microsoft shares rose 4% in extended trading after it exceeded earnings and revenue expectations. Azure grew 48%, accelerating from 42% in the prior quarter. CNBC’s Jim Cramer called Microsoft “the winner from tonight's two big Magnificent 7 reports” and described powerful technology companies as “nation states.” His CNBC Investing Club portfolio owns shares of Alphabet, Amazon, Apple, Meta, Microsoft and Nvidia. The companies’ scale is treated as spectacle and investment opportunity; workers, teams and customers whose activity feeds that scale remain the material being measured.