The Nexairi Dispatch · Friday, June 12, 2026 · Issue #26
The AI gap isn't about tools. It's about tier.
The gap isn't which tools you use. It's whether you redesigned workflows or just layered on top.
Good morning, friends. PwC put a number on the AI value gap this week: 74% of the upside goes to 20% of firms. The spending gap driving it is 681 to 1. State lawmakers are writing AI tax bills right now. And over half of accountants at AICPA's ENGAGE told pollsters they expect a recession by year-end, though that survey deserves a closer look before you clear your calendar. Useful Friday.
💰 FINANCE — 74% of AI's upside goes to 20% of firms
What happened: PwC's 2026 AI Performance Study and Ramp AI Index show 74% of AI's economic value is pooling in 20% of companies. The numbers are blunt: the top 1% of firms spend $7,500 per employee per month on AI. The median company spends $11. That 681-to-1 gap has produced a 7.2x difference in value captured.
Why it matters: The firms pulling ahead didn't buy better tools. They redesigned their workflows around AI instead of layering tools onto existing processes. That distinction is now showing up in the margins: leaders hold a 4-percentage-point profit advantage and capture 7.2x more value. For CPAs and finance advisors, one question now comes before any AI recommendation: has this client redesigned, or just added tools?
What to watch: Watch whether the value gap widens in the second half of 2026 as AI infrastructure costs keep falling. Firms still layering tools rather than redesigning have a closing window. The gap is starting to look permanent.
Outside Nexairi
Accountants at ENGAGE think we're already in a recession — CFO Dive
More than half of accounting professionals at AICPA's ENGAGE 2026 said the U.S. is in or heading into a recession by year-end. Company confidence is still holding. The gap between what practitioners feel and what the numbers show is worth watching.
State lawmakers are writing AI tax bills right now — Accounting Today
Multiple states dropped AI tax legislation this week. Taxes on AI-generated revenue, AI infrastructure and AI-derived profits are on the table. These proposals are early, but if your clients run AI-heavy operations, start flagging the exposure now.
What CPAs can delegate to AI on R&D credits and what they can't — Accounting Today
AI handles the documentation and calculation work in R&D credit claims well enough. Where the eligibility line sits still belongs with the CPA. This piece is specific about that boundary, which matters before you automate any of that workflow.
OpenAI is weighing major price cuts to compete with Anthropic — Bloomberg
Bloomberg says OpenAI is weighing major price cuts across its API and subscription tiers. Anthropic is taking enterprise accounts. For practitioners deciding which platform to standardize on, the cost math is about to change.
AI-powered synthetic identity fraud projected at $3.1 billion this year — CFO Dive
CFO Dive reports projected losses from synthetic identity fraud will hit $3.1 billion this year in U.S. unsecured credit. Criminals are using AI to build fake identities that clear standard verification checks. Updating KYC at onboarding is not enough. The detection needs to run through the full account lifecycle.
IRS is merging its practitioner offices over AICPA objections — Journal of Accountancy
The IRS is combining its return preparer and professional responsibility offices into one unit. The AICPA fought it, saying practitioners will lose direct access to specialized staff. If your firm works with either office, the routing for correspondence and representation work is about to change.
Tool Worth Knowing: Cloudskill (producthunt.com/products/cloudskill)
Cloudskill tracks the prompts, tools and workflows your team depends on. The ones that became de facto firm policy without anyone writing them down. Worth a look if you're building an AI usage policy before a regulator or client asks for one.
Deeper Read
CFOs crave real-time data. Few are there yet. — CFO Dive
Finance teams say they need real-time data. Most are still blocked by fragmented systems and technical debt. This piece explains why, and it's the same reason most finance AI projects stall before they reach production.
The Push for a Public Wealth Fund to Distribute AI Gains — Bloomberg
Lawmakers and tech firms are debating whether the public should get a share of AI profits. The proposals are early. But the core question, who captures AI's economic surplus, runs right through the concentration story above.
How Terry Tao Became an Evangelist for AI in Math — Quanta Magazine
Terry Tao works on problems most mathematicians won't touch. He's also using AI collaborators now. His account of what works and what doesn't is worth more than most commentary on the subject.