70% Want AI Budget Cuts, But Gartner Sees 47% Global Growth

So, What Actually Happened?

Thursday morning the spending desks are reading a Wednesday that handed enterprise teams a contradiction in one news cycle. A new global survey says seven in 10 companies are ready to slash AI budgets after 73% of executives reported AI investments missed expectations. We scanned 190,000 articles this week so you don't have to. Meanwhile Gartner forecasts global AI spending will grow 47% in 2026, and Qualtrics closed a $6.75B Press Ganey Forsta acquisition on the same day Commure crossed $7B on a $70M round.

The Bottom Line: When 70% of executives are willing to cut AI budgets but global AI spend still grows 47% next year, the operating question shifts under your feet: the money is moving from broad experiments to a narrow set of named outcomes, and the buyers writing the next checks already know which vendors and which workflows have receipts. The CIO walking into Friday with one page mapping which AI line items have a named ROI owner, which are still in pilot purgatory, and which got absorbed into an M&A move underneath them runs the rest of the year. The rest will be defending last year's budget.

The Tracks That Matter

1. The ROI Reckoning Just Hit AI Budgets

A global G-P survey landed Wednesday with a number Wall Street had been waiting for: seven in ten executives are ready to slash AI budgets if business goals miss this year. 73% reported at least some AI investments did not meet expectations over the past twelve months. The number of leaders describing their company as ”aggressively” using AI fell from 60% to 42% year over year. That is a procurement shift, not a survey blip.

The contrast that sharpens the read is what most CFOs still have on the 2026 CapEx file: an unbroken AI line item that has not been tagged with named owners or measurable outcomes. 88% of executives told the same survey they worry employees are ”performing productivity” with AI rather than producing value, and 69% said the time spent reviewing AI-generated work went up, not down. The ”AI is paying for itself” framing that anchored most 2025 budget conversations just got a named operating-floor counter-reference from inside the workforce that uses the tools every day.

The strategic implication: the CFO and head of platform just gained a named ”show me the ROI per AI line item” row on the next operating review. The question is no longer whether to invest in AI. It is which specific workflow has a documented owner, a measurable outcome, and a kill date if the outcome does not land. The audit-committee question for Q3 is going to be brutal for any AI program that survived 2025 on enthusiasm instead of receipts.

Here's what works: Ask the CFO and head of strategy together: for our top three AI line items in the 2026 budget, do we have a named ROI owner, a measurable outcome, and a Q3 kill date, or are we still rolling 2025 enthusiasm into 2026 line items the audit committee is about to question? G-P just put the question on the calendar.

2. Commure Just Hit A $7B Healthcare AI Valuation

Commure closed a $70 million round at a $7 billion valuation on Tuesday, the sharpest healthcare-AI valuation signal of the week. The provider-facing platform sells AI assistants and workflow automation directly into hospitals and clinics, the kind of customer that buys on documented EHR savings rather than horizon hype. The valuation lands inside a market most enterprise software CFOs had still mentally filed under ”pilot territory.”

The contrast is what most enterprise health-tech procurement desks were modeling six months ago: a horizon of three years before AI-clinical-workflow vendors would clear the valuation bar of incumbent EHR platforms. After a single round closed at $7B, the procurement narrative shifts. The ”we are evaluating AI co-pilots in a multi-vendor pilot” framing just got a named single-vendor counter-reference, and the same week Qualtrics closed a $6.75B Press Ganey Forsta acquisition explicitly framed as the experience-management layer for healthcare AI. Two operating-floor signals, one shared frame: the healthcare-AI buying decision is consolidating around named platforms, not pilots.

The strategic implication: the head of digital health and the chief financial officer just gained a named ”incumbent platform vs. consolidated AI vendor” question to put on the next vendor review. The contracted AI vendor with the EHR integrations, the compliance posture, and the multi-hospital references is starting to win the renewal cycle most pilots were supposed to graduate into. The chief medical information officer who walks into the next review with a defensive five-vendor pilot list will spend the rest of the quarter explaining why none of them can scale to a system-wide commitment.

Here's what works: Ask the CMIO and CFO together: for our top three healthcare-AI workflows currently in pilot, do we have a named consolidated-vendor candidate with EHR integration and a multi-hospital reference, or are we still budgeting for five parallel pilots that each need a separate compliance review? Commure just priced what a consolidated healthcare-AI bet looks like.

---

Try It Yourself

3. Qualtrics Just Bought $6.75B Of Experience Data

Qualtrics closed the $6.75 billion Press Ganey Forsta acquisition on Tuesday, the largest experience-management consolidation move in five years. The pairing pulls 38,000 healthcare provider organizations and hundreds of thousands of public-sector responses per year onto one AI-powered experience stack. The framing was explicit: not a feature acquisition, but a buy of the data corpus the next agentic AI layer will read from.

The contrast that sharpens the read is what most experience-management buyers still have on the architecture diagram: a survey tool sitting next to a contact-center stack sitting next to a clinical-feedback tool, each running its own AI assistant on a thin slice of the customer record. After Qualtrics absorbed Press Ganey and Forsta in one move, the ”best-of-breed XM stack” framing that anchored most 2025 procurement decks just got a named consolidated counter-reference, and the AI agent reading the full experience corpus belongs to one vendor.

The strategic implication: the chief customer officer and the chief data officer just gained a named ”where does our experience-data corpus live and who reads it” question on the next architecture review. Owning the question matters because the agentic AI layer is moving from ”summarize this dashboard” to ”act on the customer record across channels,” and the vendor with the consolidated corpus has the structural advantage on the action layer.

Here's what works: Ask the chief customer officer and chief data officer together: for our experience-management stack today, do we have a named owner for the consolidated customer-experience corpus that the agentic AI layer reads from, or are we still routing through three vendors with three partial views? Qualtrics just made consolidation the procurement signal.

4. OCC Just Telegraphed AI Governance Rules For Banks

The clearest US bank-regulatory signal of the week sits on a JD Supra alert that the OCC will issue formal AI governance guidance. The Office of the Comptroller of the Currency's annual operating-risk report telegraphed a specific direction: third-party-AI risk frameworks, model-risk management extensions, and explicit board-level reporting expectations. The ”we will publish guidance later” framing the OCC has held for two cycles just got replaced by a named regulatory trajectory.

The contrast is what most US bank chief risk officers still have on the AI-readiness file: an SR 11-7 model-risk-management posture extended informally to cover generative-AI tooling. After the OCC telegraphs a specific guidance package, the informal extension stops being defensible at the next bank exam. The ”we are following SR 11-7 for AI” framing that anchored most 2025 risk-committee decks just got a counter-reference, and the Q1 2027 bank examiner is going to ask which specific AI workflows have a named third-party-risk and board-reporting owner.

The strategic implication: the chief risk officer and head of model risk just gained a named ”third-party-AI risk policy and board-reporting owner” row on the AI governance scorecard. The vendor-selection question for AI in banking is no longer ”does this vendor pass model-risk review.” It is ”which specific OCC guidance line item does this AI workflow sit on, and who is the named board-reporting owner of the third-party-risk posture behind it.”

Here's what works: Ask the chief risk officer and head of model risk together: for our top three AI workflows in customer-facing banking operations, do we have a named third-party-risk policy and a board-reporting owner, or are we still extending SR 11-7 informally and hoping the OCC guidance lands later? The OCC just told you it is coming.

5. Trump And Kennedy Want To Loosen AI Healthcare Rules

The sharpest US federal-regulatory signal of the week sits on a Union-Bulletin report that the administration wants to relax safeguards for AI healthcare tools. Trump and Kennedy framed the move as removing innovation friction for diagnostic and clinical-decision AI, walking back the more cautious posture in place at the start of the year. The ”FDA AI/ML clearance bar is the floor” framing that anchored every health-AI procurement deck in 2025 just got a named counter-reference.

The contrast is what most hospital chief medical officers had pencilled into the 2026 health-AI procurement calendar: a single rising regulatory bar that would only get harder. After the federal posture flips toward loosening, the procurement question reverses. State-level health-AI laws fill the vacuum, payer-side liability rises, and Plus Therapeutics' new partnership with Ephemeral Technologies on CNS-oncology AI lands in a different procurement environment than the one it was modeled in. The single-federal-bar mental model GCs have been running on just split into a federal-relax, state-tighten, payer-litigate triangle.

The strategic implication: the GC and chief medical officer just gained a named ”post-federal-relaxation procurement risk” row on the AI scorecard. The question is no longer only ”does this AI tool have FDA clearance.” It is ”which state regulators, payer contracts, and litigation exposures are now load-bearing, and who is the named owner of each one when the federal floor moves.”

Here's what works: Ask the GC and CMO together: for our top three clinical-AI tools currently in evaluation, do we have a named state-regulator, payer-contract, and litigation-exposure owner, or are we still treating FDA clearance as the single procurement bar? The federal floor just started moving.

6. Sharon AI Books $950M In Sovereign AI Capacity

The clearest sovereign-AI infrastructure signal of the week sits on an Edge IR report on Sharon AI's $950M Australian deal. The Australian neocloud locked in nearly a billion in AI infrastructure capacity, explicitly framed around regional data sovereignty. The deal lands inside a procurement environment where regulated industries are starting to write ”data must not leave national borders” into AI contracts. The ”AI capacity is a global hyperscaler conversation” framing just got a named regional counter-reference.

The contrast is what most APAC chief data officers still have on the cloud architecture diagram: three hyperscaler regions plus a hope that data residency reads will hold up under regulator scrutiny. After Sharon AI books $950M into Australian sovereign capacity, the regional procurement narrative shifts. Banking regulators, healthcare data authorities, and defense procurement officers in the region just got a named alternative to the three-hyperscaler default. The Teradata Factory launch explicitly framed around data sovereignty as a core architectural decision lands the same week, on the same theme from the platform side.

The strategic implication: the chief data officer and head of compliance just gained a named ”sovereign-AI capacity counterparty” row on the procurement scorecard. The question is no longer only ”do we have a hyperscaler region in our jurisdiction.” It is ”do we have a named regional sovereign-AI counterparty on the file with a contracted-capacity window and a regulator-jurisdiction note, and is that counterparty a real procurement option or still a slide in the strategy deck.”

Here's what works: Ask the CDO and head of compliance together: for our top three regulated AI workloads in the APAC region, do we have a named sovereign-AI capacity counterparty with a contracted-capacity window, or are we still running data-residency on the three-hyperscaler-default mental model? Sharon AI just priced the sovereign-AI option.

7. Gartner Pegs 2026 AI Spending Growth At 47%

Gartner forecasts worldwide AI spending will grow 47% in 2026, the kind of analyst-firm number that lands inside the CFO budget conversation regardless of whether the CFO loves analyst firms. The sharper read is the spread between the headline number and the G-P survey on the same wire showing 70% of executives ready to cut AI budgets. Two numbers, one news cycle, opposite directions.

The contrast is what most finance teams still have on the 2026 AI line item: a single growth assumption applied across every AI-touched line. After the analyst firm pegs topline growth at 47% and a separate operating-floor survey shows the same executives ready to slash budgets, the spread is the story. Both can be true. Aggregate spend grows while individual line items get killed, because the money is consolidating around named outcomes and broad experimentation is getting cut. The same week the CMMI Institute completed a pilot for a new AI Maturity Framework named the maturity side of the same shift.

The strategic implication: the CFO and head of strategy just gained a named ”consolidating spend vs. broad experimentation” question for the 2026 AI budget review. The vendor with the named outcome owner and the maturity rating gets the budget growth share. The pilot without a measurable outcome gets the kill date. The audit committee meeting after Q2 is going to look very different for AI line items depending on which side of the consolidation they sit on.

Here's what works: Ask the CFO and head of strategy together: for each AI line item rolling from 2025 into 2026, can we point to a named outcome owner, a maturity rating, and a kill date if the outcome misses, or is this line still running on aggregate AI optimism that the audit committee is about to question? Gartner and G-P just printed the spread for you.

Signal vs. Noise

🟢 Signal: Risk Management. Risk management is climbing in real influence on Wednesday into Thursday morning while the broad ”AI” label loses ground in the same window. Enterprise buyers are quietly moving the AI conversation from product features down to the risk-and-governance layer, the same layer the OCC just telegraphed it will publish formal guidance on. Most coverage is still keyword-screening for funding rounds and missing where buying authority moved.

🔴 Noise: ”AI” as a single topic. The undifferentiated ”AI” label pulled the most mentions across the wires Wednesday but its real operating influence dropped as the conversation split into named sub-tracks (AI budget reckonings, AI governance for banks, sovereign-AI capacity, healthcare-AI regulatory whiplash). Anyone still tracking ”AI news” as a single signal is reading from a 2024 keyword filter while the operating floor split it into a dozen named files.

From the 190K

We scanned 190,000 articles this week. Here's what no one's talking about:

The G-P survey showing 70% of executives ready to slash AI budgets, the Gartner forecast that 2026 AI spending will grow 47%, and the OCC telegraphing formal AI governance for banks all printed inside one 36-hour window.

Each desk reads these as unrelated stories. The HR-and-workplace wires file G-P. The CFO desks pick up Gartner. The bank-regulatory desks write up the OCC report. Read them on the same morning and a different shape emerges: the AI spending conversation just got named consolidation criteria (kill the broad experiments, fund the named outcomes), a named topline trajectory (47% growth on the consolidated bets), and a named regulatory backstop (OCC governance for banks, on top of the EU AI Act sub-deadlines and state-level health-AI laws). The ”we have AI in the 2026 budget” framing that anchored most mid-market strategy decks just got three named counter-references at once, each from a different layer of the operating stack (workforce ROI, analyst topline, regulator backstop).

The strategic move on Friday is mapping which AI line items in your 2026 budget have a named owner, a measurable outcome, and a regulator-aware compliance posture, and which still run on aggregate enthusiasm. The week just priced the gap.

By The Numbers

• Seven in ten executives are ready to slash AI budgets, with 73% reporting failed AI investments over the past 12 months — The G-P global survey just turned the AI ROI question from a vendor sales objection into a named CFO budget filter for 2026.

• Gartner forecasts worldwide AI spending will grow 47% in 2026 — A topline growth number from the analyst firm CFOs cite in budget reviews, even when they roll their eyes about the methodology.

• Qualtrics closed a $6.75 billion Press Ganey Forsta acquisition pulling 38,000 healthcare provider organizations onto one experience platform — The largest experience-management consolidation in five years, sized to give the agentic AI layer a single corpus to read from.

• Commure raised $70 million at a $7 billion valuation for its provider-facing healthcare AI platform — A clinical-workflow valuation number from a sector most procurement desks still had filed under ”pilot territory.”

• The number of executives describing their company as aggressively using AI to innovate fell from 60% to 42% year over year — A workforce-side number that lands the same week as the Gartner 47% topline forecast, surfacing the consolidation thesis in one sentence.

• Sharon AI secured a $950 million AI infrastructure deal in Australia as sovereign AI capacity locks in regional supply — A regional sovereign-AI capacity number from a procurement environment that is starting to write data-residency directly into the AI contract.

• Confluent reports its Private Cloud with Intelligent Replication delivers up to 73% fewer brokers than open-source Kafka at matched performance — A real benchmark number for the data infrastructure layer most enterprise AI workloads run on top of.

• See what's rising across the 190K-article corpus this week →

Deep Dive: The AI Budget Just Split In Two

Every DJ knows the moment when one track is hiding another. The crowd hears the headline mix, but the working track is doing the load-bearing work underneath. This week the enterprise-AI room had two tracks playing on top of each other, and most desks heard only one. Gartner says AI spending grows 47% in 2026. G-P says 70% of executives are ready to cut AI budgets. Both are true at the same time, and the gap between them is the strategy map for the next two quarters.

The Consolidation Track

The Gartner 47% growth forecast is the topline analyst signal. The G-P survey showing 70% of executives ready to slash budgets is the workforce-side counter-signal. Read together, the shape is: aggregate AI spend grows while individual line items get killed, because the money is consolidating around named outcomes. The 88% of executives worried about ”performing productivity” with AI is the disqualification mechanism. The 82% hiring AI talent in new countries is where the consolidating spend is going. The CFO who walks into the next architecture review with a single AI growth assumption applied across every line is the one who explains in Q3 why three pilots that survived 2025 just got killed and the budget did not follow.

The Governance Track

The OCC report telegraphing formal AI governance guidance is the bank-regulatory equivalent of saying: the third-party AI risk frameworks and board-reporting expectations the OCC has held back for two cycles are now on the calendar. The chief risk officer who walks into Q1 2027 with informal SR 11-7 extensions covering generative-AI tooling is the one explaining to the bank examiner why no specific AI workflow has a named third-party-risk owner. The governance track is not the slow lane. It determines which line items survive the consolidation cut.

The Sovereignty Track

The Sharon AI $950M Australian deal is the regional-infrastructure equivalent of saying: sovereign AI capacity now has a real procurement option outside the three-hyperscaler default. The chief data officer who has only modeled hyperscaler-region growth on the data-residency line is missing the second-order question: which regional sovereign counterparty is on the file behind the AI workload, and is that counterparty a real procurement option or still a slide. The same week Teradata launched an on-premises AI system explicitly framed around data sovereignty, reinforcing the same theme from the platform side.

What Actually Works

1. Tag every 2026 AI line item with a named outcome owner and a kill date. Every AI line item carries a measurable outcome and a Q3 kill date before the audit committee asks. G-P just told you 70% of your peers are ready to slash. Be on the right side of the cut.

2. Put a named third-party-AI risk and board-reporting owner on every regulated AI workflow. Every AI workflow inside a regulated function gets a named owner of the third-party-risk policy and a board-reporting checklist. The OCC just told you formal guidance is coming.

3. Put a named sovereign-AI capacity counterparty on every APAC AI workload. Every regulated AI workload in the region gets a named sovereign-AI counterparty alongside the hyperscaler row. Sharon AI just made it a real procurement option.

4. Put a named consolidated-vendor candidate on every multi-pilot AI portfolio. Every multi-vendor pilot portfolio gets a named consolidation candidate before the renewal cycle. Qualtrics and Commure just printed what consolidation looks like.

The next analyst forecast, the next budget-cut headline, the next regulatory signal will hit by next Friday. The room is still moving. The operator who walks into Friday with the outcome-owner row, the risk-owner row, the sovereign-counterparty row, and the consolidation-candidate row already on the dashboard is the one mixing for the rest of the quarter. The one waiting for the headline number is going to play to a thinner floor by Q3.

What's Coming

The First Top-50 Bank To Disclose A Named Third-Party-AI Risk Owner In A 10-K

The OCC report telegraphing AI governance guidance is the trigger. The next move is the first top-50 US bank disclosing, in an annual report or risk filing, a named owner of the third-party-AI risk policy and a board-reporting checklist tied to specific AI workflows. That disclosure is one to two cycles out, and the chief risk officers who already pre-staged the file absorb it as routine.

The First Sovereign-AI Counterparty Disclosure In A Regulated APAC Procurement

The Sharon AI $950M Australian deal is the trigger. The next move is the first regulated APAC enterprise (banking, healthcare, defense) disclosing a named regional sovereign-AI counterparty in a procurement contract or annual filing. That disclosure is probably one to two cycles out, and the CDOs who already pre-staged the sovereign-watch line absorb it as routine.

The First Public Disclosure Of An AI Line Item Killed For Missing Q3 Outcome Targets

The G-P survey on AI budget cuts is the trigger. The next move is the first mid-market enterprise publicly disclosing a specific AI line item killed because the named outcome target missed at Q3 review. That disclosure is one to two cycles out, and the CFOs who already tagged outcomes and kill dates use it as cover for their own consolidation moves.

For Your Team

Strategic purpose: Friday is the day this week's budget-reckoning, governance-telegraph, and sovereignty signals get translated into one Strategy Map before the next budget or risk review. The work is one named owner per load-bearing layer: the outcome owner on each AI line, the third-party-risk owner on each regulated AI workflow, the sovereign-counterparty owner on each regional load, and the consolidation candidate on each multi-pilot portfolio. Everything else is commentary.

Friday's meeting prompt: ”If G-P just put 70% of executives on record ready to slash AI budgets while Gartner forecasts 47% growth in the topline, if the OCC just telegraphed formal AI governance for banks, and if Sharon AI just locked in $950M of sovereign capacity, who in this room owns the named scorecard across our AI outcome owners, our regulated-AI risk owners, our sovereign-AI counterparties, and our consolidation candidates, and is that owner one person, or four people who have never been in the same meeting?”

The Budget Reckoning Framework:

1. One named owner per load-bearing AI layer. CFO and head of strategy co-own the outcome-owner row on every AI line. Chief risk officer and head of model risk co-own the third-party-AI risk and board-reporting layer. CDO and head of compliance co-own the sovereign-AI counterparty layer. Chief customer officer and chief data officer co-own the consolidation-candidate layer.

2. Named outcome owner and kill date on every AI line item. Every 2026 AI line item gets a measurable outcome and a Q3 kill date before the audit committee asks. G-P put 70% of your peers on record.

3. Named third-party-AI risk policy per regulated AI workflow. Every AI workflow inside a regulated function gets a named third-party-risk owner and a board-reporting checklist. The OCC just told you guidance is coming.

4. Named sovereign-AI counterparty per regional AI workload. Every regulated AI workload in the region gets a named sovereign-AI counterparty alongside the hyperscaler row. Sharon AI made it a real option.

5. Named consolidation candidate per multi-vendor AI portfolio. Every multi-pilot portfolio gets a named consolidation candidate before the renewal cycle. Qualtrics and Commure printed what consolidation looks like.

Share-worthy stat: A G-P global survey just put seven in ten executives on record ready to slash AI budgets if business goals miss this year, while Gartner forecasts AI spending will grow 47% in 2026 anyway. Drop both numbers on the next budget review and the consolidation thesis writes itself in 30 seconds.

Go deeper: Track the budget, governance, and sovereignty signals in real time →

The Track of the Day

”AI is increasingly being measured by trust, accountability and business impact. The future belongs to companies that pair AI with the right expertise, governance and operational discipline to turn opportunity into real business outcomes.”
, Pete Tiliakos, Principal Analyst at GxT Advisors

Today's set: ”Walk the Line” by Johnny Cash, cued at the moment the loud opening sets have cleared and the working tracks have to carry the room. The Gartner topline number was the headline cue this week. The G-P survey on budget cuts, the OCC governance telegraph, the Qualtrics consolidation move, the Commure valuation, and the Sharon AI sovereign deal were the tracks that actually moved the floor. The operator who waits for the analyst headline while the outcome-owner, risk-owner, sovereign-counterparty, and consolidation layers keep shipping is going to play to a thinner room by Q3. The one who walks into Friday with all four rows already on the dashboard is headlining the rest of the cycle.

Yves Mulkers, your data DJ, mixing 190,000 articles into the tracks that actually matter.

We scanned 190,000 articles this week so you don't have to. Data Pains → Business Gains.

Published: May 21, 2026 | Curated by Yves Mulkers @ Ins7ghts