Spend Is Up, Portfolios Are Flat
The defining enterprise software pricing trend of 2026 is that spend is rising sharply while what it buys is not. Gartner forecasts enterprise software spending up about 14.7% in 2026, passing $1.4 trillion globally. At the individual-enterprise level the same pressure shows up as an average SaaS bill of around $55.7 million a year, up roughly 8% year on year — yet application portfolios have stayed essentially flat at about 305 applications. Enterprises are paying meaningfully more for roughly the same software, which means the increase is almost entirely price, not expanded value. That gap is the cost of passive renewal, and closing it is the single biggest pricing opportunity most organisations have. It is also the reason this trend analysis sits at the centre of our enterprise software pricing pillar.
The flat portfolio number carries a second lesson. If application counts are static while spend climbs, then rationalisation — eliminating redundant and unused tools — is not keeping pace with price growth, and the unused share of the average estate is quietly being repriced upward along with everything else. Every shelved licence carried into a renewal is bought again at the new, higher rate. The buyers controlling cost in 2026 are pairing hard price negotiation with disciplined right-sizing, so that the inflation they cannot avoid applies to a smaller, justified footprint rather than a bloated one. Benchmarking the portfolio, as set out in our guide to IT contract benchmarking, is what makes both halves of that discipline possible.
SaaS Inflation Is Running Near 12%
The headline number behind the spend growth is SaaS inflation of roughly 12% a year — close to five times the consumer inflation rate of the major economies, and rising even as general inflation has eased. This is not a one-off correction; it is a structural feature of a market where switching is hard and renewal uplifts compound. An uncapped annual increase of 8–10%, applied each cycle, roughly doubles a unit price within a decade regardless of whether you add a single seat. Vendors present these uplifts as standard and inevitable, embedded in escalator clauses ranging from 5% to 25%, when in reality they are among the most negotiable terms in any contract. Understanding the longer arc — covered in our analysis of software inflation rates and forecasts — is what lets a buyer treat the annual increase as a number to be capped rather than a fact to be accepted.
SaaS inflation near 12% is roughly five times consumer inflation. An uncapped 8–10% uplift applied every renewal doubles your unit price inside a decade — for the same software.
The compounding is what makes the trend so dangerous and so invisible. A 9% uplift looks reasonable in the year it lands; across a five-year relationship it adds more than half again to the unit price, and across a decade it more than doubles it. Because each increase is judged on its own, in isolation, the cumulative effect rarely reaches the level of scrutiny it deserves. The remedy is to model the full-term and ten-year cost of any uplift at signing, so the compounding is visible before it is agreed, and to insist on a cap tied to a published index rather than left to the vendor's discretion. A number you would never accept as a lump sum is exactly what an uncapped escalator delivers, one quiet renewal at a time.
The AI Premium Is Repricing Everything
Artificial intelligence is the largest single force behind the 2026 numbers, and it adds cost in two distinct ways. The first is the premium add-on: AI capabilities layered onto existing subscriptions at a steep markup. Microsoft raised Microsoft 365 prices by $3 a month when it folded Copilot in during early 2025, and enterprise AI tiers routinely carry a significant premium over the base licence. Spending on AI-native applications has more than doubled, reaching an average of around $1.2 million per enterprise in 2026. The risk is paying a large, recurring premium for capabilities whose return is still unproven at scale — which is why AI add-ons should be negotiated as separate, pilotable commitments rather than waved through as part of a renewal.
| 2026 Pricing Signal | Figure | Buyer Implication |
|---|---|---|
| Enterprise software spend growth | ~14.7% (to $1.4T) | Budget pressure is structural |
| SaaS price inflation | ~12% / year | Cap renewal uplifts in the contract |
| Avg enterprise SaaS spend | ~$55.7M (+8% YoY) | Same ~305 apps; increase is price |
| Avg AI-native app spend | ~$1.2M (more than doubled) | Negotiate AI as a separate commitment |
The Shift to Consumption Pricing
The second way AI is changing pricing is structural: it is accelerating the shift away from the predictable per-seat subscription toward consumption metrics — tokens consumed, workflows executed, agents run. Hybrid models that pair a base subscription with variable usage tiers are becoming the enterprise default, and they move the financial risk of forecasting onto the buyer. A per-seat bill is predictable; a per-token bill scales with adoption in ways that are genuinely hard to model and easy to overrun, turning a successful rollout into a budget surprise. The discipline that protects you is the same as for any consumption contract: negotiate caps, committed-use discounts, and real-time monitoring before you sign, not after the first overage. The mechanics of doing this for AI specifically are set out in our guide to AI token pricing.
Consumption pricing is not inherently bad for buyers — for variable or experimental workloads it can be far cheaper than paying for seats that sit idle — but it demands governance the per-seat world never required. You need visibility into usage as it accrues, alerts before a cap is breached, and contractual protection against retroactive repricing of consumed units. The vendors moving fastest to consumption are betting that buyers will not build that governance in time, and that overruns will become a reliable revenue stream. Treating a consumption contract with the same rigour you would apply to cloud spend — commitments, caps, monitoring and a right to renegotiate as usage patterns settle — is what turns the model from a risk into an advantage.
What This Means for Your 2026 Negotiations
These trends point to one conclusion: passive renewal is now the most expensive thing a software buyer can do. With prices rising near 12% and AI premiums stacking on top, the cost of accepting the vendor's default is higher than it has ever been — and the value of preparation correspondingly greater. Three responses follow. First, benchmark relentlessly, because with effective prices inflating you cannot judge a quote without current market data; the method is in our guide to IT contract benchmarking. Second, cap every renewal uplift in the contract rather than fighting it annually. Third, treat AI and consumption commitments as negotiable, pilotable line items, never inevitable add-ons. The vendors raising prices fastest — including Microsoft and its peers — are also the most negotiable for buyers who arrive prepared.
The organisations that come through 2026 with their software budgets intact will be the ones that treated these trends as a planning input rather than a surprise. That means mapping every major renewal against the vendor's fiscal calendar, building benchmarks before the vendor opens the conversation, and resourcing the highest-spend negotiations as the material commercial events they are. The inflation is real and largely unavoidable in aggregate; what is entirely within the buyer's control is whether it applies to a benchmarked, right-sized, capped portfolio or an unexamined one. Our Price Benchmarking Report tracks these movements vendor by vendor. To benchmark your portfolio against current 2026 pricing, request a confidential briefing.