State of Gen AI: 2026 Market Analysis and Forecasts

Executive Summary

LONDON, March 2026 — The generative AI market has reached what Gartner calls an inflection point. Worldwide AI spending is forecast to total $2.59 trillion in 2026, a 47% year-over-year increase, while GenAI model spending alone is projected to grow 80.8%. Yet beneath the headline capital flows lies a widening execution gap. McKinsey reports that 88% of organisations now use AI in at least one function, but nearly two-thirds remain stuck in pilot mode and only about 6% qualify as high performers achieving more than 5% EBIT impact. A widely cited MIT (Project NANDA) study found 95% of pilots delivered no measurable P&L impact, a contested headline figure that defines "failure" narrowly as no rapid revenue or P&L impact. The defining story of 2026 is not adoption — it is the migration from experimentation to scaled, workflow-redesigned deployment, and the shift in compute economics from training to inference.

Key Takeaways

• Gartner forecasts worldwide AI spending of $2.59 trillion in 2026, up 47% year-over-year, with 2026 identified as the enterprise "inflection year."
• Bloomberg Intelligence projects the generative AI market to reach $2.3 trillion by 2032, with inference — not training — becoming the dominant revenue driver.
• McKinsey finds 88% of organisations use AI, but only ~6% are high performers with material EBIT impact; workflow redesign is the single biggest value lever.
• An MIT (Project NANDA) study reports 95% of GenAI pilots delivered no measurable P&L impact, though the methodology has drawn criticism for its narrow six-month ROI window.
• Klarna's OpenAI-built assistant scaled to the work of 853 full-time agents and $60 million in savings by Q3 2025 — a benchmark enterprise case study.
• Model concentration is loosening: the share of tokens requested from Google, OpenAI and Anthropic fell to 33% in June 2026 from 72% a year earlier, according to OpenRouter data cited by Bloomberg.

Market Sizing and Growth Drivers

The 2026 market is defined by two parallel realities: extraordinary capital commitment and uneven returns. According to Gartner, worldwide AI spending will reach $2.59 trillion in 2026. Analyst John-David Lovelock framed the moment bluntly: "Up to this point, AI spending has primarily been driven by technology companies and hyperscalers. Enterprises have yet to really flex their spending potential. That is coming and 2026 will be the inflection year." Gartner separately projects GenAI model spending to grow 80.8% in 2026.

Bloomberg Intelligence raised its long-range forecast by $500 billion, now projecting a $2.3 trillion generative AI market by 2032 — roughly 22% of total technology spending. The revision is driven by accelerating token consumption, the expansion of coding and customer-service agents, and a faster-than-expected shift of compute from training to inference. BI's Mandeep Singh noted: "The generative AI market has reached an inflection point where inference — not training — is becoming the dominant revenue driver." Within that forecast, coding agents alone are projected to approach $100 billion by 2032, up from $1 billion in 2024.

The sustainability question looms over the buildout. According to a report from research firm Exponential View cited by Bloomberg, global AI sales (excluding China) reached $25 billion in Q1 2026, exceeding the industry's estimated $21 billion in depreciation costs for a second consecutive quarter — but margins remain thin, with depreciation consuming more than two-thirds of revenue.

Table 1 — Gen AI Market Sizing and Forecasts (2026)

The Scaling Gap: Adoption Without Impact

The central paradox of 2026 is that near-universal adoption has not translated into proportional financial returns. McKinsey's flagship State of AI 2025 survey, fielded across 1,993 respondents in 105 nations, found that 88% of organisations now use AI in at least one business function, up from 78% in 2024. However, nearly two-thirds remain in experiment or pilot mode, and only about 6% qualify as high performers achieving more than 5% EBIT impact.

Related: Moonshot AI $20B Valuation 2026: China's Largest LLM Raise Reshapes AI Race

McKinsey identifies the decisive lever: out of 25 attributes tested, the redesign of workflows has the single biggest effect on an organisation's ability to see EBIT impact from generative AI. Yet only 21% of firms using gen AI have redesigned any workflows — nearly 80% are simply layering AI atop existing processes. Among high performers, one-third allocate more than 20% of their digital budget to AI, against just 7% of other organisations.

On agentic systems, McKinsey reports 23% of organisations are scaling an agentic AI system somewhere in their enterprise, with a further 39% experimenting — but in any single function, no more than 10% are scaling agents. Deloitte's State of AI in the Enterprise 2026, surveying 3,235 leaders, reinforces the governance lag: only one in five companies has a mature governance model for autonomous AI agents, even as agentic usage is poised to rise sharply.

For deeper context, see our Gen AI analysis: "OpenAI Adjusts Pricing with $100/month Pro Plan Launch in 2026".

The ROI Problem in Hard Numbers

The most provocative datapoint of the cycle came from MIT's Project NANDA study, reported by Fortune, which found that 95% of pilots delivered no measurable P&L impact, with only 5% of integrated systems creating significant value. According to the report itself, the research drew on 52 structured interviews, 153 senior-leader survey responses and a review of more than 300 public AI deployments conducted between January and June 2025; the study found that more than half of generative AI budgets target sales and marketing, while the biggest ROI sat in back-office automation. It also found that buying from specialised vendors succeeds about 67% of the time, while internal builds succeed only one-third as often.

The finding is not uncontested. As the Marketing AI Institute noted, the study defined success narrowly — measurable KPIs and ROI within six months of pilot — ignoring efficiency gains, churn reduction and pipeline velocity. The WRITER 2026 Enterprise AI Adoption Survey of 1,200 executives found that AI super-users deliver 5X productivity gains, yet only 29% of organisations see significant ROI from generative AI and 23% from AI agents, with 59% of companies investing over $1 million annually. The Thomson Reuters 2026 report found generative AI use nearly doubled to 40% of professionals — yet only 18% knew their organisation was tracking ROI at all.

Additional coverage: DoorDash, Spotify & Uber Expand AI Integrations in 2026

Case Study: Klarna's Customer Service Assistant

The most-cited enterprise deployment remains Klarna. At launch in February 2024, Klarna reported its OpenAI-built assistant handled 2.3 million conversations — two-thirds of customer-service chats — doing the equivalent work of 700 full-time agents, cutting repeat inquiries 25%, resolving issues in under two minutes versus 11 previously, across 23 markets and 35-plus languages, with an estimated $40 million profit improvement for 2024. By Q3 2025, CEO Sebastian Siemiatkowski reported the agent did the work of more than 853 full-time agents and had saved the company $60 million. Klarna illustrates both the upside of scaled, workflow-embedded deployment and the back-office ROI thesis MIT identified.

Competitive Landscape

The model layer is fragmenting. Data from OpenRouter, cited in a Bloomberg report, showing the share of tokens requested from Google, OpenAI and Anthropic models falling from 72% to 33% in a year signals intensifying competition from open-weight and specialised models, with implications for pricing and enterprise vendor strategy.

Related: Gen AI Market Size and Forecast Statistics 2026-2030

Table 2 — Enterprise AI ROI Benchmarks (2026)

Practical Business Implications

For enterprise decision-makers, the 2026 evidence points to four actions. First, prioritise workflow redesign over tool layering — McKinsey's data make this the clearest predictor of EBIT impact. Second, reallocate budget toward back-office automation where MIT found the strongest returns, rather than defaulting to sales-and-marketing pilots. Third, favour vendor partnerships over internal builds for first deployments, given the ~2x success differential. Fourth, instrument ROI from day one: with only 18% of professional-services firms tracking returns, governance and measurement maturity are now competitive differentiators. Sectors from health tech operations to education technology are facing the same scaling discipline.

Forward Outlook

The near-term outlook hinges on whether inference revenue can sustain hyperscaler capex approaching $750 billion in 2026. The first signs are positive — sales now exceed depreciation — but margins are thin and the model layer is commoditising fast. Expect 2026–2027 to reward operators who treat AI as a workflow and governance challenge rather than a procurement exercise. As agentic systems move from experiment to production, the governance gap Deloitte identifies will become the dominant risk vector. The capital intensity of the buildout also ripples into adjacent markets, from critical minerals to data-centre power and the connectivity infrastructure underpinning distributed inference. Private capital is following the trend into adjacent data-rich verticals, as seen in genetics asset consolidation.

For deeper context, see our AI analysis: "AWS Expands AI Solutions with New Integrations".

Frequently Asked Questions

How big is the generative AI market in 2026?

Gartner forecasts worldwide AI spending of $2.59 trillion in 2026, up 47% year-over-year, with GenAI model spending growing 80.8%. Bloomberg Intelligence projects the generative AI market to reach $2.3 trillion by 2032.

Why do most AI pilots fail to show ROI?

MIT's Project NANDA study found 95% of pilots delivered no measurable P&L impact, attributing this largely to misallocated budgets and a failure to redesign workflows. McKinsey similarly found only ~21% of firms have redesigned workflows, the single biggest driver of EBIT impact.

What separates AI high performers from the rest?

McKinsey found high performers redesign workflows and allocate budget aggressively — one-third commit more than 20% of their digital budget to AI, versus just 7% of other organisations. Only about 6% of firms qualify as high performers with material EBIT impact.

Is the AI infrastructure buildout financially sustainable?

Bloomberg reports global AI sales (ex-China) reached $25 billion in Q1 2026, exceeding the ~$21 billion in depreciation costs for a second consecutive quarter. However, depreciation consumes over two-thirds of revenue, leaving thin margins against power, labour and financing costs.

What is the most credible enterprise Gen AI case study?

Klarna's OpenAI-built customer service assistant is the most-cited deployment, scaling to the equivalent of 853 full-time agents and $60 million in savings by Q3 2025 — exemplifying the back-office automation ROI that MIT identified as most valuable.

Sources include company disclosures, regulatory filings, analyst reports, and industry briefings.

Related Coverage

• More Gen AI Analysis
• Latest Technology News

Analysis based on company announcements, investor disclosures, regulatory filings, Reuters, Bloomberg, Financial Times, CNBC, SEC documentation, and publicly available market data as of publication.