.png)
.jpg)
Frameworks, core principles and top case studies for SaaS pricing, learnt and refined over 28+ years of SaaS-monetization experience.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Welcome to Monetizely’s 2026 Monetization Outlook. We’ve put together a list of major market forces and our 13 key predictions for the year. We hope you enjoy this article.
We first set out to explain our view on market forces and then dive into our predictions.
In 2026, the tech industry is encountering a convergence of powerful economic and technological forces. These major forces include an AI-driven plunge in software costs, a brewing Main Street recession from past monetary tightening, new crypto adoption laws that could rewrite the rules of finance, and a looming federal debt bubble.
AI’s Deflation Bomb in Software Development
Rapid advances in artificial intelligence are unleashing what many describe as an “atomic bomb-esque” deflationary shock on the software industry. Generative AI tools (like coding assistants and GPT models) are dramatically lowering the cost of producing software, automating tasks that once required large teams of developers. In effect, software is becoming commoditized – easier and cheaper to create – eroding the pricing power of traditional software companies[1]. GitHub’s AI pair-programmer Copilot, for example, already contributes to nearly half of new code written, showcasing how much development can be offloaded to algorithms[3]. As one Goldman Sachs analysis put it, businesses “might not have to buy so much software” when they can build more in-house with AI, threatening the growth of today’s packaged software providers[4].
The deflationary impact of AI isn’t just theory – it’s happening now. Microsoft’s CEO Satya Nadella has called software “the most deflationary force in the world” for businesses, noting it helps companies do more with fewer resources[5]. With AI supercharging that effect, the economics of software are being upended. Generative AI can generate code, test, and even maintain software with minimal human input, meaning one skilled programmer armed with AI can accomplish what used to require an entire team. In vivid terms, tasks that once took 20 mid-level programmers might soon need only 3 or 4 people when assisted by advanced AI like ChatGPT[7]. This huge leap in productivity acts as a “great deflation bomb” for white-collar tech work – slashing labor needs and, by extension, the cost of software development[7].
The result is fierce price compression across the software sector. SaaS (Software-as-a-Service) vendors face pressure as AI-driven alternatives undercut their products at a fraction of the cost[8]. Entire business models are being disrupted: why pay high subscription fees for a software tool if an AI system can deliver similar functionality cheaply or even free? Already, enterprise customers are tightening software budgets – partly cyclically (after the pandemic boom and amid higher interest rates) and partly to invest in generative AI itself[2]. This shift diverts spending away from traditional software.
As we’ll see, this deflationary force hits at a time when other storm clouds are gathering over the economy, compounding the challenges for the tech sector.
Main Street Recession and Demand Crunch
While tech grapples with AI-driven deflation, the broader Main Street economy is showing signs of recession. After aggressive interest rate hikes in 2022–2023, the lagged effects of monetary tightening are materializing in many local businesses and consumer sectors. Economic pain is mounting quickly for America’s small businesses, which are often the first to feel the squeeze of higher borrowing costs and cautious spending. In fact, recent data shows an alarming divergence between small firms and large firms – with the former shedding jobs and struggling, even as some big companies continue to grow[11]. This “Main Street bust,” as Axios described it, is raising the chances of a recession that hits everyday businesses and workers, even if headline GDP or corporate earnings have held up so far[11].
Key indicators are flashing warning signs. In late 2025, small firms (those with fewer than 50 employees) cut a net 120,000 jobs in a single month, even as larger employers still added some jobs[13]. That was the worst small-business payroll drop since the pandemic onset, indicating real stress on Main Street. Bankruptcy filings among small businesses have also surged to record levels, with 2025 seeing the highest number of Subchapter V (small business) bankruptcies in the six-year history of that program[14]. High interest rates are a major culprit – they’ve driven up loan costs, rent, and other expenses for local shops, even as customers turn frugal. As one economist noted, “small firms are the leading indicator” for the national economy, and right now they’re signaling weakness[15]. Unlike big corporations, small businesses have fewer levers to pull when the economy softens: they can’t easily access global markets, raise prices without losing customers, or cut large efficiencies[16]. This makes them especially vulnerable to a tight monetary environment.
Crypto Policy Opens the Door for Stablecoins
In a surprising turn, governments are increasingly embracing crypto assets through new regulations, with a focus on stablecoins – cryptocurrencies pegged to fiat currencies like the US dollar. After years of uncertainty, lawmakers in major economies have started to pass laws that integrate stablecoins into the financial system. For instance, the United States enacted the GENIUS Act in July 2025, creating a regulatory framework for USD-backed payment stablecoins[17]. This landmark law clarified that stablecoins are not securities or bank deposits (and thus not automatically covered by FDIC insurance), while setting standards for issuers to ensure these tokens maintain a stable value[18]. Crucially, it allows both banks and licensed nonbank entities to issue stablecoins with proper oversight[19]. Similar moves are happening globally – over 70% of jurisdictions worldwide advanced new stablecoin regulations in 2025 alone[20]. The broad motive is to legitimize stablecoins for mainstream use, seeing them as a bridge between traditional money and the innovation of crypto.
Policymakers are increasingly viewing stablecoins as a strategic asset. By allowing fintech and tech companies to offer digital dollars, governments hope to bolster their currencies’ global role and even tap into new demand for their debt. A senior U.S. official explained that dollar stablecoins “have the potential to ensure American dollar dominance internationally… and create potentially trillions of dollars of demand for U.S. Treasuries, which could lower long-term interest rates.”[21] In other words, if people around the world hold digital dollars (via stablecoins) instead of, say, foreign currencies, they’ll indirectly be funding the U.S. government by buying the Treasury bills and bonds backing those stablecoins. Indeed, Fed Governor Stephen Miran noted in late 2025 that widespread stablecoin adoption is already increasing demand for U.S. Treasuries and other dollar assets, which “lowers borrowing costs for the U.S. government”[22]. He likened the effect to a new form of the global savings glut that kept U.S. interest rates low in the 2000s[23].
Clear rules on stablecoins pave the way for big tech firms and startups alike to integrate digital dollars into their apps – enabling instant, low-cost payments globally. We may soon see stablecoins used for everything from e-commerce to remittances to DeFi (decentralized finance) services, all under a legal umbrella. It might even spur innovation in how software services are monetized – imagine subscription fees or cloud credits payable in stablecoins seamlessly.
Another wild card is how monetary policy adapts to the rise of crypto. We already see hints that the Federal Reserve may adjust its approach in a stablecoin-rich world. Governor Miran suggested that if stablecoins significantly raise the pool of loanable funds (by attracting global capital into dollars), the Fed’s neutral interest rate (r) could be lower, meaning the Fed might need to keep its policy rates lower than otherwise to support the economy[25]. Essentially, stablecoin-driven liquidity could be deflationary, giving the Fed room – or forcing it – to ease up. This comes at a time when the Fed was already expected to pivot due to the slowing economy.
In sum, crypto adoption laws are unlocking a new era of digital finance at the same time the Fed stands at a crossroads. For tech firms, especially those in fintech or Web3, this is a green light to innovate with stablecoins and digital payments.
The Massive Federal Debt Bubble Looming
The final major force casting a long shadow is the massive U.S. federal debt burden – a bubble in government liabilities that many fear is unsustainable. Years of heavy deficit spending (including pandemic stimulus and other outlays) have pushed the U.S. national debt to record levels, now roughly equal to the nation’s entire GDP and climbing[30]. This by itself is worrying, but what’s truly alarming is the skyrocketing cost of that debt in an era of higher interest rates. In Fiscal Year 2025, U.S. federal interest payments crossed the \$1 trillion mark for the first time ever, up nearly threefold from just five years earlier[31].
There’s also a linkage between the debt bubble and the other forces we discussed. The push for stablecoin adoption, as noted, is partly an attempt to create new demand for Treasuries[21]. In essence, the government might be using tech and financial innovation (like crypto) as a pressure valve for the debt problem. If trillions in global savings can be channeled into U.S. bonds via digital dollars, it could help sustain the debt a while longer by keeping interest rates lower than they’d otherwise be. Indeed, the phenomenon of stablecoins bolstering the dollar’s global role could marginally ease the debt burden by lowering borrowing costs[22]. However, this is not a panacea. The scale of U.S. deficits – running around \$2 trillion each year going forward[34] – means debt will keep piling up. At some point, investors could demand higher yields, or the Federal Reserve may feel compelled to intervene (potentially buying bonds via quantitative easing to cap yields, effectively “monetizing” the debt). Such actions carry risks of their own, like reigniting inflation or distorting markets.
Setting The Stage For Our 13 Predictions for 2026
In narrative terms, we are setting the stage for a new era. The exuberance of the 2010s (with cheap money and ever-rising software valuations) has given way to a more fraught landscape. Deflation in software means innovation must focus on quality and unique value, not just scale, since scale is cheaper now. A recessionary climate means efficiency and resilience are paramount – survival for many firms will depend on prudent management and clear value propositions. Crypto’s rise offers new opportunities to those who embrace it, potentially enabling the next generation of fintech and web infrastructure, but it also could redraw the competitive map between tech firms and banks. And the fiscal cloud of high debt will be in the background of every major decision, possibly prompting shifts in taxes, interest rates, or government spending that can ripple through the tech ecosystem.
1. Accelerating Deflation Could Put Many SaaS Companies Out of Business
As AI-generated code and automated solutions proliferate, enterprises “might not have to buy so much software” anymore[36]. This deflationary pressure means many incumbent SaaS vendors could see shrinking revenues and margins. They face a double squeeze: customers spending less on third-party apps as they build AI-powered tools in-house, and big platforms bundling formerly paid features for free (or cheap). As a result, some SaaS companies – especially smaller, single-feature providers – may not survive this wave of price compression. Even Goldman Sachs analysts note that AI threatens to “deflate the software industry,” reducing the influence (and profitability) of today’s software firms[37]. In short, accelerating price deflation and feature commoditization pose an existential challenge: adapt business models or risk being left behind.
Cost declines in AI are also fueling this trend. The cost to run advanced AI models has been plummeting, enabling cheaper alternatives to traditional software. For example, OpenAI’s latest reasoning model saw its operation cost drop 80% in just two months[38].
In practical terms, what used to be a paid human-driven service delivered via SaaS might now be automated by an affordable AI agent. This puts heavy deflationary pressure on pricing across the board. Companies that cannot offer unique value – beyond what an open-source model or quick in-house solution can do – will struggle to justify their subscriptions. The end result is likely a shakeout where only the most resilient or innovative SaaS players survive, while many others consolidate or shutter due to an unsustainable business model in the deflation era.
2. AI Automation Creates New Software Categories from Former Services
Advances in AI are turning what used to be human-only services into automated software products. In other words, tasks traditionally performed by professional service firms or outsourcing providers are now being “productized” by AI-powered platforms[41]. Modern AI – especially large language models with tool use and reasoning – can handle work that was previously too unstructured or complex for software. A vivid example is business process outsourcing (BPO): activities like customer support, data entry, claims processing, etc., which companies used to hand off to service firms, can now be done by AI bots at scale[42]. As Andreessen Horowitz observes, “AI has become exceptionally good at handling work that couldn’t previously be done adequately with software,” enabling startups to unbundle the BPO and offer it as a SaaS-like product[41].
This trend is spawning entirely new categories of software. For instance, AI voice agents and browser bots are tackling call center and back-office tasks, respectively, that used to require armies of human workers[43]. The “service” (the human labor) is being replaced by “software” (AI algorithms) in areas like legal document review, marketing content creation, bookkeeping, and more. We’re seeing AI-driven legal research tools, copywriting assistants, and autonomous customer service agents – essentially SaaS products filling roles once offered only via consulting or outsourcing contracts. Crucially, these AI solutions often deliver results faster and at lower cost, which is why enterprises are eager to adopt them. By in-housing functions with AI and eliminating external service fees, companies can save money while still performing the work – a powerful incentive fueling this shift[42] The implication is that some service firms (e.g. in consulting, outsourcing, agencies) will need to reinvent themselves, as AI software handles more of their traditional tasks. Meanwhile, vendors who embrace AI to create these new product categories (think AI co-pilots or “agents” in every domain) have an opportunity to disrupt legacy service providers and capture new markets.
3. Re-emergence of User-Based Pricing (Seats) – Predictability Over Pure Usage
In the near term, enterprise software pricing is swinging back toward the familiar per-user (seat) model, after a period where usage-based schemes gained traction. The reason comes down to predictability and simplicity. Customers have grown wary of unpredictable metered pricing for AI features (charges per API call, per conversation, etc.), which can spike costs in unexpected ways[46]. High interest in AI notwithstanding, CIOs and CFOs prefer a pricing model they understand – and nothing is more straightforward than paying a set fee per user. Even Salesforce, which had experimented with usage and per-chat pricing for its new AI “Agentforce” product, found that clients pushed back. Salesforce CEO Marc Benioff admitted they started by trying per-conversation charges, “but customers have pushed for more flexibility” and clarity[48]. Now Salesforce is largely standardizing on seat-based licensing (with some caps or credits embedded) for its AI offerings[49]. This reflects a broader industry trend: in 2025, many vendors reverted to selling AI capabilities as user licenses to make enterprise buyers comfortable.
Why the retreat to user-based pricing? Simply put, enterprises need cost stability. With AI features, usage can vary wildly – a handful of “power users” might generate a huge share of the consumption[51]. That makes purely consumption-based bills hard to forecast and budget for. Procurement and finance teams loathe “blank check” scenarios where heavy usage blows past budgets. By paying per seat (often with a generous allowance of AI usage included), customers regain a sense of control: costs are fixed or at least bounded[53]. As Gartner software licensing experts note, this shift back to seats indicates buyers “want some certainty in the pricing model before investing” heavily in AI tech[50]. It’s essentially a trust-building phase. Indeed, a recent analysis observed that seat-based AI licenses with usage caps are becoming the norm so that vendors and customers share risk: users get predictability, while vendors protect themselves by embedding fair-use limits or credit top-ups if needed[54]. In summary, for the short term, selling AI add-ons “per user” is seen as the pragmatic approach, even if the long-term vision is to move to value or usage metrics. Major players embracing this (e.g. Salesforce’s Agentic ELA bundle based on seats + credits[55]) signal to the market that per-user pricing is acceptable – even for cutting-edge AI – until buyers are ready for more exotic models.
4. Monetization Becomes a Dedicated Discipline (Pricing & Billing Complexity)
With the explosion of new pricing models and billing complexities (from usage tiers and AI credits to hybrid subscriptions), companies are realizing that monetization strategy needs dedicated focus and expertise. In 2026, leading firms treat monetization not as a one-time pricing decision but as an ongoing discipline – blending product, finance, and engineering considerations[56]. The days of a simple price list are over; modern SaaS businesses might be running multiple models simultaneously (seat licenses, consumption metrics, one-time add-ons, etc.), and the infrastructure to support that is non-trivial[58]. According to industry analysis, this “explosion of complexity” has turned monetization systems from back-office plumbing into a strategic asset[60]. Companies that excel at managing pricing complexity – with proper tooling, telemetry, and cross-functional processes – are pulling ahead of those using spreadsheets and ad-hoc fixes[59]. In short, pricing and monetization is now a first-class problem requiring its own team (“monetization architects” or pricing ops) and technology stack.
The rise of AI has only heightened this need. Zuora’s 2025 monetization trends note that many firms rushed to add AI features and struggled to bill for them, ending up with “ad-hoc meters” and manual revenue recognition that leaked revenue[62]. The “winners,” however, were those that “treated AI monetization as a discipline” – insisting on accurate usage metering, robust billing rules, and cross-team coordination so that pricing experiments didn’t break the finance systems[57]. We also see the emergence of specialized monetization roles and platforms. It’s not uncommon now for a SaaS scale-up to have a Director of Monetization or a dedicated pricing product manager whose job is to align product value metrics with billing. These professionals use sophisticated billing systems that can handle multi-dimensional pricing (e.g. a base subscription + consumption credits + outcome-based bonuses). The complexity is evident in scenarios like hybrid models: one Metronome report found that under the hood of a seemingly simple seat price, you might have “grandfathered plans, regional variants, AI add-on credits, and enterprise overrides all operating at once”[58]. Managing this requires a concerted strategy. Accordingly, boards and executives are starting to treat monetization capabilities as a competitive differentiator – investing in it the way you’d invest in product development. The endgame is clear: those who can engineer flexible, customer-aligned pricing quickly (and reliably) will outperform those who cannot adapt. Monetization, in effect, has become a science of its own.
5. Initial AI Feature Arbitrage Closes – AI Add-Ons Become Table Stakes
In the early phase of the AI boom, software vendors enjoyed an “AI arbitrage” – the ability to sell AI features as high-margin add-ons due to their novelty. That window is closing fast. What used to be a paid differentiator (e.g. “Pro plan includes AI insights” for an extra fee) is increasingly bundled into core platforms at little or no additional cost. In 2024–2025, generative AI went from a premium feature to table stakes in many applications[64]. Tech giants accelerated this shift. For example, Google initially charged $20–30 per user for its Workspace AI (Duet/Gemini) but then abruptly ended those fees – rolling generative AI features into all paid Workspace plans for free[65]. They raised base subscription prices slightly, but effectively signaled that AI assistance (drafting emails, generating slides, etc.) is now a standard part of the product, not a separate SKU. An IDC analyst called Google’s move “an unusually bold first move… a sign we will experience AI everywhere”, noting that Google is “disrupting a market while some application vendors are still trying to monetize generative AI”[65]. In other words, laggards who attempt to charge distinctly for what users increasingly expect as built-in may find themselves outmaneuvered.
Similarly, Microsoft is embedding AI across its product suite. Microsoft still prices Copilot as an add-on in some cases, but it announced that many new AI capabilities will be included in Microsoft 365 and other flagship products, accompanied by broad-based price increases on those subscriptions (e.g. ~16% hike planned by mid-2026 to reflect AI value)[67]. This strategy means AI features aren’t really optional extras anymore – they’re part of the baseline offering and overall fee. The rationale is clear: generative AI is becoming as fundamental as the UI itself. It’s a must-have, not a nice-to-have; thus vendors use it to enhance their core product appeal or justify moderate price uplifts, rather than trying to nickel-and-dime for each AI capability. Analysts observe that this is effectively commoditizing many AI features – the value shifts to higher-level outcomes or platform stickiness, while the AI functionality alone won’t command a premium for long[69].
That said, big platforms can still leverage distribution to monetize AI at scale (e.g. Microsoft can charge a bit more from millions of Office users). But smaller software providers will struggle to charge separately for similar AI functions once those become expected. If Google Docs offers AI writing help at no extra cost, a niche docs SaaS can’t easily charge $20/month just for an AI writer. This dynamic forces vendors to either bake AI in and compete on product experience (not on AI itself), or find truly unique AI capabilities that merit standalone fees. Industry commentators put it plainly: “AI is becoming table stakes, but not every vendor agrees on how to price it”[70]. Right now we see contrasting tactics – some like Google bundle AI to drive adoption, others like Microsoft add fees – but the overall direction is that yesterday’s AI differentiators are tomorrow’s standard features. The initial arbitrage opportunity is closing as the market normalizes.
6. AI “Agents” Enter a Refinement Phase to Deliver on Their Promise
2023’s hype declared autonomous AI agents would revolutionize work, but 2026 finds organizations in a sober refinement period for these technologies. Early agent implementations often fell short of the sweeping promises, revealing that a lot of groundwork is needed for them to truly deliver value. One major realization is that organizational data readiness is critical – successful agents need access to clean, relevant internal data and systems. Companies now recognize they must “get agent-ready” by organizing their proprietary data and integrating knowledge bases so agents can actually act intelligently[71]. An IBM report on AI agents notes that the real wins will go to firms that “take their private data and organize it so that the agents are researching against your documents”, turning enterprise knowledge into fuel for agentic workflows[71]. Many organizations are currently in the process of building these data pipelines and governance structures. In the meantime, agents are being used in more limited, pilot capacities while data and compliance catch up.
Another key insight: not all agents are the same, yet many were being “brushed with the same strokes” in hype discussions. There’s a spectrum from co-pilots (AI assistants that help a human user, but operate under constant supervision) to true autonomous agents that could independently carry out complex tasks. Right now, most so-called “agents” are closer to co-pilots or have very constrained autonomy. They often rely on large language models with some added planning and tool-use capabilities – still far from a general AI employee. As IBM’s AI experts put it, what’s currently marketed as agents are basically LLMs with “rudimentary planning and tool-calling”, a far cry from the vision of an independent reasoning entity[73].
Reaching that vision will require further breakthroughs in contextual reasoning, reliability, and safety. We’re seeing a pause for refinement: teams are improving agent algorithms, defining guardrails, and differentiating use cases (where a co-pilot is enough vs. where full autonomy is feasible). A Bain analysis mapped out scenarios from “AI enhances SaaS” to “AI fully cannibalizes SaaS,” highlighting that many workflows will remain human-in-the-loop for now, until agents prove they can handle exceptions and decisions robustly[74]. Additionally, enterprise buyers are learning that an “agent” for one purpose can be very different from an agent for another – there’s confusion to clear up. Surveys show almost every enterprise is exploring agent tech, but almost none feel it’s production-ready to completely self-drive processes yet (one study found 96% of companies plan to expand AI agents, but only 1% say their AI is mature enough for independent decisions so far[76]).
In sum, the industry hype around agentic AI has tempered into a practical phase of optimization. Companies are assessing where agents work and where they need more training. Data integration, model tuning, and careful scope definition are the focus, so that the original promise of agents – truly autonomous workflow execution – can eventually be achieved safely. It’s a bit of a reality check period. We anticipate that after this refinement and perhaps the next generation of model improvements, some organizations will manage to deploy deeply agentic systems that function as a core of their software. But in early 2026, most “agents” are still co-pilots or narrowly-scoped bots, and that’s okay – it’s part of the maturation cycle.
7. Agentic-First Software Players Could Disrupt Category Leaders
While incumbents refine their products with incremental AI additions, a new breed of agentic-first software companies could leapfrog them by building AI agents into the core of their solutions. These upstarts design their software architecture around autonomous or semi-autonomous agents from the ground up, rather than as an add-on. If they succeed, they can deliver step-change improvements in productivity or user experience that challenge the dominance of today’s category leaders. We’ve seen this pattern before with technological shifts: those who fully embrace a paradigm can eclipse those who only layer it on. For example, consider an ERP or CRM built entirely around AI-driven processes vs. a legacy CRM that added a chatbot. The former could potentially “disrupt the software status quo” by solving problems in fundamentally more efficient ways. According to Bain & Company, incumbents that don’t proactively infuse AI and agent capabilities risk “disruption, obsolescence, and losing out to entrants.” Tasks that are easy to automate will be easy for others to copy, so if a market leader doesn’t do it, a newcomer will[75]. We’re already seeing entrants in specific niches (like AI-native project management tools, recruiting tools, etc.) positioning their agentic approach as a competitive advantage against older players.
What does it mean to be “agentic at the core”? It means the software isn’t just a set of user interfaces with some AI on the side; instead, it orchestrates work autonomously on behalf of the user as a primary function. These systems leverage AI to handle multi-step workflows across other apps or data sources. For the end-user, an agentic application might feel less like using software and more like delegating tasks to a smart colleague. If done well, this can massively reduce the manual effort required, which is a compelling value proposition. Startups in this vein can attract customers by promising outcomes (e.g. an agentic bookkeeping software might close your books automatically, not just provide reports). Incumbents have a hard time with this shift because it often cannibalizes their existing UX-driven models and requires deep AI expertise. Hence the door is open for disruptors. Gartner forecasts that by 2035, agentic AI could account for 30%+ of software revenue[77], implying a huge new market for those who master it. We expect some of today’s category leaders will adapt and survive – the likes of Salesforce, for instance, are investing heavily in AI agents and have the data moats to compete. But others may be caught flat-footed.
The next few years will likely see an arms race: incumbents trying to “agent-ize” their offerings versus startups that are agent-native from day one. History suggests that when a technology shift is big enough, at least a few newcomers will break through. Those that truly build agentic capability at the heart of their software – effectively turning software into a more autonomous service – stand to capture market share from slower competitors. As Bain’s analysis put it, “winners will be the organizations that scale agent orchestration best… companies must pick a lane: become the neutral agent platform or supply the unique data that powers it”[75]. In plain terms, either you’re building the AI brains that run things (and potentially displacing others), or you become a data provider to those brains. Companies that fail to do either could see their competitive position slip. This prediction is that we will indeed witness some category shake-ups where an AI-agent-centric newcomer vaults into leadership by doing things the old guard’s software was never designed to do.
8. Procurement Adapts to Buying Bundles of Usage Credits/Tokens
Enterprise procurement teams are evolving their practices to accommodate credit- or token-based pricing for software and cloud usage. Rather than just buying a number of seats or a flat annual license, they’re now often negotiating “bundles of credits” that represent a certain amount of usage (compute, API calls, AI model tokens, etc.) which can be drawn down. This is a significant shift in how software is bought and sold, brought on largely by cloud and AI services. Traditionally, procurement disliked variable usage pricing due to unpredictability, but the model of pre-purchasing credits provides a middle ground: it offers volume commitment and cost visibility to the buyer, while giving flexibility in how the usage is consumed. In 2026 we see a normalization of this approach. Many major vendors have introduced credit systems – Microsoft with Azure and AI credits, Salesforce with its “flex” AI credits, OpenAI switching its enterprise plans to pooled credits, and so on[78]. Startups and smaller SaaS are following suit, often because selling in “credits” abstracts the complexity of underlying usage (which can vary by feature) into a single understandable unit for customers.
Crucially, procurement teams are learning to accept and even prefer this model after initial hesitation. According to industry commentary, the main roadblock was that “nothing scares procurement more than runaway costs with no visibility,” and initially usage tokens felt like that[80]. But now that large incumbents are validating credit-based pricing as standard, buyers are coming around[81].
One analysis notes that “Salesforce and OpenAI are doing the hard work of educating the market”, making credits a familiar concept and showing that with proper caps/monitoring, unpredictability can be managed[82]. Indeed, many enterprise deals for AI or cloud include negotiated committed-use credit bundles to ensure predictability[83]. Procurement groups have adapted by asking the right questions: How many credits do we need for the year? What happens if we exceed them? Can we top-up at the same discounted rate? – treating credits almost like a new currency to budget for. They also demand tools from vendors for tracking usage against credits in real-time (to avoid surprises). As a result of these changes, it’s now common in board presentations to see, for example, a line item like “500k AI credits” as part of a software contract.
This trend mirrors how, in cloud contracts, companies long ago moved to buying reserved instances or committed cloud spend. Now that mindset is moving up the stack to SaaS and AI APIs. Procurement is essentially learning to buy “capacity” instead of just licenses. The benefits are mutual: vendors get revenue upfront and commitment, buyers get bulk discounts and some cost certainty. A notable statistic: industry experts estimate that “most AI spend is now in the form of credits… also referred to as tokens, AI units, generative credits, etc.”[79].
We’ve basically created a new procurement vocabulary overnight. Going forward, we expect nearly every enterprise software RFP that involves variable usage will offer a credits-based option. It’s becoming a default structure for bridging the gap between pure pay-as-you-go and fixed price models.
9. Ongoing Confusion over Outcome-Based Pricing
Among all the evolving pricing schemes, outcome-based pricing remains the most nebulous and misunderstood for many. The idea sounds simple: charge customers only when a certain desired outcome or result is achieved (e.g. a sales AI charges per actual conversion it helped generate, rather than per user or per attempt). While this model is touted as the ultimate alignment of vendor and customer interests, in practice it’s rare – and when discussed, it often leads to confusion.
Many people still conflate outcome-based pricing with other models, or struggle to define measurable outcomes in the first place. Even in the AI era, truly outcome-tied deals are “mostly – for now – dead” or at least extremely uncommon, as one industry expert bluntly put it. Research confirms this: as of 2026, the vast majority of enterprise AI deals use usage or hybrid pricing, with purely outcome-based arrangements remaining rare[85]. Enterprise buyers are often uncomfortable with the concept of paying directly for outputs or business results because it introduces new uncertainties (how do we agree on attribution? what if outcomes are below target due to customer-side factors, etc.)[85].
The confusion is visible in the market dialogue. Some vendors market their pricing as “outcome-based” when it’s essentially usage-based by another name (for example, charging per API call and claiming the outcome is each successful response). True outcome pricing would be more like “pay us 10% of cost savings we deliver” – which is complex to implement. Internally, even teams that aim for outcome-centric models often retreat to simpler proxies.
A blog by Chargebee, for instance, highlighted that while outcome pricing is appealing in theory (and many leaders want to get closer to value-based pricing), the messy reality of AI products is that different users derive different outcomes, making a generic outcome metric elusive[87]. They cite Zapier’s AI Agents: one customer might use it to deflect support tickets, another to accelerate content creation – the “outcomes” are apples and oranges[89]. Zapier sensibly chose to charge by activity (per task run) and simply communicate the various outcome possibilities, rather than try to price each possible outcome separately[88]. This story is playing out across the industry.
So, people remain a bit confused about what outcome-based pricing truly means and when it works. There’s debate and learning in progress: e.g. how to structure contracts, how to measure baseline vs. achieved outcome, who assumes the risk. In many cases, vendors find it easier to stick to input-based pricing (seats, tokens) and just prove ROI through case studies, instead of literally charging on ROI.
Outcome-based models do exist (some cybersecurity firms charge per prevented breach, some marketing firms charge per lead or sale), but they tend to be bespoke deals rather than standard pricing across a customer base. Analysts note that “truly outcome-based pricing remains rare” and often requires high customer trust and data sharing to pull off[85].
Until more of that infrastructure and trust is in place, expect the confusion to continue. Customers will keep asking for outcome-based pricing (“why don’t you charge us only if your AI actually improves X metric?”) and vendors will continue finding that easier said than done. Education is needed on both sides to clarify where outcome pricing is feasible and how it’s different from usage pricing. It’s likely a longer-term evolution, not an immediate switch, despite the buzz.
10. CTOs Must Consider the Cost Implications of Technical Decisions
The role of the CTO (Chief Technology Officer) is broadening to include a much sharper focus on cost and unit economics of technology choices. With the rise of cloud services and especially AI APIs – which can be very expensive at scale – technical architecture decisions directly influence the company’s margins and cost of goods sold (COGS).
In 2026, it’s no longer viable for engineering leaders to be ignorant of cost; in fact, boards and CEOs are demanding that “CTOs understand the cost of their decisions.” Concretely, this means when selecting an AI model or designing a feature, the CTO must ask: How much will this scale cost to run? Are we using an overly large model where a smaller one would suffice? Can we batch or cache to reduce API calls? The financial impact of such choices can be huge.
One SaaS CFO recently gave a pointed example: their product team initially wanted to include a powerful (but costly) AI feature for all users, but finance intervened, saying “not unless you want to blow up margins.” They calculated that using the AI without limits could result in “$10k of costs on a $500 plan,” which is obviously unsustainable[92]. This scenario is playing out across many companies – the technology team might build something, but if the variable cloud/AI costs per user are too high, the business can’t scale it profitably. Hence, CTOs are now expected to be on top of these numbers and design with cost-efficiency in mind.
Additionally, many organizations are forming FinOps (Financial Operations) teams or similar to bridge the gap between engineering and finance. The CTO often partners closely with such teams, or even leads them, to ensure cost observability and optimization is part of the development process. The emergence of AI cost management as a discipline underscores this trend. Unlike traditional software (which had near-zero marginal cost per user), AI features have a tangible cost per use – e.g. every time an AI writes an email or analyzes data, some compute expense is incurred.
Without controls, this can lead to what some call “AI cost sprawl,” where cloud bills skyrocket unexpectedly[94]. CTOs now need strategies to prevent that: choose efficient algorithms, set usage limits, instrument the code to track expensive operations, etc. In board meetings, gross margin impact of AI and cloud spend is a hot topic, and the CTO is held accountable for managing it.
We see a cultural shift where engineers are being educated about cost (“write efficient prompts/code, because tokens = money” is a 2026 engineering motto in AI-heavy teams). In summary, technical leaders must blend engineering excellence with cost discipline. A good CTO today can talk about model architectures and GPU-hour costs in the same breath. Those who do will help their companies maintain healthy margins even as they deploy advanced tech, whereas those who ignore costs might deliver great functionality at an untenable cost structure. The mandate is clear: technology strategy and financial strategy have converged, and CTOs are at that intersection.
11. High-Quality Professional Services Become a Key Premium Offering
As software products become more commoditized and automated, premium value is increasingly delivered via services – particularly high-quality professional services wrapped around the product. In other words, customers are willing to pay top dollar for expert support, customization, and strategic guidance that ensure they achieve success with the software, even if the software’s features alone don’t justify a premium price. We see SaaS companies doubling down on their professional services, customer success, and consulting offerings as differentiators. In a world of deflationary software (see Prediction #1), the revenue mix may shift such that the “service” component (human expertise) carries the premium, rather than the software bits which might be cheap or even free. For example, a SaaS analytics platform might charge relatively low subscription fees but offer a “white-glove” analytics consulting package at a much higher rate for enterprises that need help interpreting data and integrating insights into their business. Those willing to pay are essentially paying for outcomes and assurance, delivered by people, not just for software features.
This trend is partly a response to customer demand: businesses have realized that simply buying AI or software doesn’t automatically yield results – you need the right implementation and process changes. Vendors that can provide that hands-on assistance stand to command higher loyalty and additional revenue. It’s also a hedge against the declining margins on pure software.
Professional services (if delivered efficiently) can have healthy margins and are harder to commoditize because they involve tailored solutions and human relationships. There’s evidence of this emphasis in various sectors. For instance, cloud providers now offer extensive professional services to help with cloud migration and architecture, knowing the cloud resources alone are getting cheaper. Software integrators and vendors are partnering to package more consultative services on top of products. Customers, in turn, are evaluating vendors not just on features, but on the support and expertise available. A “premium” vendor is one who will work closely with your team and guarantee success, versus a low-cost provider who just hands over a tool.
In 2026, even born-in-the-cloud SaaS companies that once prided themselves on being hands-off are embracing services. It’s not uncommon to see SaaS firms building out sizable professional services teams or partner ecosystems to offer implementation projects, custom integrations, training programs, etc. at a premium.
High quality is key – merely offering services isn’t enough; it has to be excellent service that drives value. If done right, it creates a virtuous cycle: the better the service, the more value the customer sees, which justifies renewals and upsells. It also differentiates the vendor in a crowded market.
We anticipate that as basic software functionality becomes inexpensive (or copied by AI), the human element – consultants, customer success managers, industry experts – becomes the new battleground for premium offerings. Companies that excel in customer-facing services will outcompete those that don’t, especially for enterprise clients who expect that partnership. In summary, the “premium” dollars in B2B tech are shifting: instead of paying purely for more features, customers will pay for better service. The software is the enabler, but the service is the multiplier that ensures the customer actually realizes the value.
(No direct external source was found for this specific point, but it is inferred from industry trends in software pricing and value delivery.)
12. Margins and COGS Take Center Stage in Board Meetings
After years of “growth at all costs” mentality in tech, the pendulum has swung back toward profitability and efficiency, meaning metrics like gross margins and Cost of Goods Sold (COGS) have become top agenda items in board meetings. Board directors and investors are drilling into questions like: “What is our gross margin on the core product vs. the AI features? Are our cloud infrastructure costs under control? How can we improve our unit economics?” This heightened scrutiny is driven by several factors. One is the macro environment – interest rates and funding conditions have made profitability more important for valuation. Another is the aforementioned rise in variable costs (cloud, AI compute, etc.) which can erode margins if not managed. So, boards are keenly interested in understanding the cost structure of the business at a granular level.
For SaaS companies, gross margins were traditionally very high (80%+), but adding heavy AI functionality can drag those down, as each usage incurs expense. Software that once had near-zero marginal cost now has a noticeable COGS, and leadership must adapt. It’s not just theoretical – there have been cases where offering an AI feature caused gross margin to drop, prompting course corrections. In one field report, a pricing exec noted “AI overages are real money we have to eat,” contrasting it with seat license overages which just mean lost revenue[93]. That sentiment illustrates why margins are a hot topic: every 1% change in gross margin can translate to millions in profit difference at scale, so it’s firmly a board-level concern. We also hear CFOs in boardrooms explaining how they plan to “monetize to avoid eating costs”[92] – for instance, ensuring customers who use a lot of AI pay proportionally, to protect margins. The board’s role is to challenge and support such strategies. They’re asking if pricing models are properly aligned to costs (see Predictions #3, #8) and if not, what the plan is to fix it.
COGS visibility is another component. Boards want to see COGS breakdowns – how much is cloud hosting, how much is third-party API spend, etc. – and then see initiatives for optimization. It’s become common to present cost optimization projects alongside product roadmaps. Some companies have even tied a portion of engineering KPIs or compensation to cost efficiency improvements, underscoring how strategic it is. In summary, expect board meetings in 2026 and beyond to be far more finance-focused when it comes to product decisions. High-level growth metrics alone won’t satisfy; boards will often zero in on questions of sustainability and efficiency: “Are we scaling revenue without scaling cost at the same rate? What does our gross margin look like by product line, and how can technology help improve it?” The companies that can show a credible story of maintaining or improving margins while growing will be rewarded (in stock price, investor confidence, etc.). Those that cannot will feel pressure to change course. This renewed prominence of margins and COGS is essentially a return to business fundamentals – even in high-tech, a business must eventually make more than it spends, and that truth is front and center again.
13. Stablecoins Start Making Inroads into B2B Payments
Stablecoins – digital currencies pegged to stable assets like the US dollar – are moving from the crypto fringes into mainstream B2B payment flows. In particular, companies are beginning to use stablecoins for cross-border B2B transactions as a faster, potentially cheaper alternative to traditional banking. A recent industry survey found that among firms already using stablecoins, 62% use them to pay suppliers across borders and 53% use them to accept payments from overseas business partners[96]. These are significant numbers, indicating that a substantial portion of early adopters are leveraging stablecoins specifically for B2B payments. The appeal is clear: a USD-backed stablecoin (like USDC or USDT) can be sent globally in minutes, 24/7, with low fees, and without the friction of correspondent banks and forex conversions. For businesses dealing with international vendors or subsidiaries, this can streamline treasury operations. It essentially brings the speed of crypto with the familiarity of a dollar unit, thus “entering” the B2B payments space in a practical way.
In 2025, we’ve seen major developments that boost confidence in stablecoins for enterprise use. Some large financial institutions and fintechs are embracing stablecoin settlement. For example, Visa announced pilots for settling transactions in USDC stablecoin, and fintech companies are offering business accounts that support holding and sending stablecoins. There have also been regulatory signals (such as the proposed U.S. Stablecoin legislation) that provide more clarity, encouraging corporates to experiment. McKinsey’s analysis suggests stablecoins are on the cusp of materially shifting the payments industry, forecasting that 2026 could be a tipping point with broad adoption in cross-border transfers[98] The cost savings on FX and fees, plus near-instant settlement, are hard for finance departments to ignore – especially for high-value international B2B payments where a few basis points saved is meaningful.
Moreover, stablecoins can help companies manage treasury across countries by providing a digital dollar that’s easier to move than actual dollars in bank accounts. Early adopters in, say, the export/import business have used stablecoins to speed up payments and avoid having capital tied up in transit. One report by William Blair highlighted that cross-border B2B payments are the most compelling use case for stablecoins in commerce, due to these efficiency gains[100]. Of course, adoption is not without hurdles – firms need to set up digital wallets, ensure compliance (KYC/AML), and be comfortable with the custodial aspects. But the survey data shows universal awareness and rapidly rising interest: 95% of companies expected growing interest in stablecoins in the next year, according to the EY study[101]. In practical terms, we predict more corporations will start holding a small portion of treasury in a stablecoin form for transactional purposes, and more deals (especially with crypto-friendly counterparties or in regions with weaker banking infrastructure) will be settled via stablecoin. It’s a gradual entry, but the trend is clear – stablecoins are moving from niche to normal in B2B money movement, heralding a new era of digital finance integration in corporate settings[102].
Sources:
1. Lucas, M. (2024). The Deflation of Software. Goldman Sachs Global Institute[35][1].
2. Monetizely (2025). The Deflationary Impact of AI on the Software Industry[8].
3. Distil AI / Rich Karlgaard (2023). The Great AI Deflation Bomb[7].
4. Axios (2025). Main Street bust threatens the entire economy[11][36].
5. Reuters (2025). Fed’s Miran: Stablecoin adoption could put downward pressure on interest rates[25][22].
6. Brookings – Nellie Liang (2025). Stablecoins: Issues for regulators as they implement GENIUS Act[17][18].
7. Reuters Breakingviews (2025). US stablecoin power-play could epically backfire[21].
8. Committee for a Responsible Federal Budget – CRFB (2025). Trillion-Dollar Interest Payments Are the New Norm[31][33].
[1] [2] [3] [4] [9] [10] [35] The deflation of software | Goldman Sachs
[5] [6] [8] The Deflationary Impact of AI on Software: Why Artificial Intelligence is Reshaping SaaS, Pricing, and Productivity
https://www.getmonetizely.com/blogs/the-deflationary-impact-of-ai-on-the-software-industry
[7] The Impact of AI Deflation on the Industry - Distil AI
https://distilinfo.com/distilai/2023/04/21/the-impact-of-ai-deflation-on-the-industry/
[11] [12] [13] [14] [15] [16] [36] Main Street bust threatens the entire economy
https://www.axios.com/2025/12/04/main-street-bust-threatens-the-entire-economy
[17] [18] [19] [24] Stablecoins: Issues for regulators as they implement GENIUS Act | Brookings
https://www.brookings.edu/articles/stablecoins-issues-for-regulators-as-they-implement-genius-act/
[20] Global Crypto Policy Review Outlook 2025/26 Report | TRM Labs
https://www.trmlabs.com/reports-and-whitepapers/global-crypto-policy-review-outlook-2025-26
[21] [28] US stablecoin power-play could epically backfire | Reuters
[22] [23] [25] [26] Fed's Miran: Stablecoin adoption could put downward pressure on interest rates | Reuters
[27] Survey: Recession By July 2024 Likely Despite Resilient Economy | Bankrate
[29] Goldman Sachs Sees Fed Delivering First Rate Cut in Q3 2024
[30] [31] [32] [33] [34] Trillion-Dollar Interest Payments Are the New Norm-2025-12-16
https://www.crfb.org/blogs/trillion-dollar-interest-payments-are-new-norm
[35] [36] [37] [39] [40] The deflation of software | Goldman Sachs
[38] [74] [75] Will Agentic AI Disrupt SaaS? | Bain & Company
https://www.bain.com/insights/will-agentic-ai-disrupt-saas-technology-report-2025/
[41] [42] [43] [44] [45] Unbundling the BPO: How AI Will Disrupt Outsourced Work | Andreessen Horowitz
https://a16z.com/unbundling-the-bpo-how-ai-will-disrupt-outsourced-work/
[46] [47] Enterprise AI Pricing Models Explained: Why Predictability Is Key
https://www.uctoday.com/unified-communications/ai-pricing-models-enterprise-predictability/
[48] [49] [50] [53] [54] [55] [77] Salesforce opts for seat-based AI licensing • The Register
https://www.theregister.com/2025/12/12/ai_agents_salesforce_pricing/
[51] [52] [78] [79] [80] [81] [82] Why everyone’s switching to AI credits
https://www.growthunhinged.com/p/ai-credit-pricing
[56] [57] [62] The Top 5 Monetization and Finance Trends of 2025 - Zuora
https://www.zuora.com/guides/the-top-5-monetization-and-finance-trends-of-2025/
[58] [59] [60] [61] [63] The Hidden Complexity of Modern Monetization | Metronome blog
https://metronome.com/blog/hidden-complexity-of-modern-monetization
[64] [65] [66] [69] Google ups Workspace price, makes Gemini AI features available for free – Computerworld
[67] [68] Microsoft 365 to include more AI tools – at a higher price | CIO Dive
https://www.ciodive.com/news/microsoft-365-ai-tools-higher-price/807189/
[70] AI Pricing: What’s the True AI Cost for Businesses in 2025? - Zylo
https://zylo.com/blog/ai-cost/
[71] [72] [73] AI Agents in 2025: Expectations vs. Reality | IBM
https://www.ibm.com/think/insights/ai-agents-2025-expectations-vs-reality
[76] Agentic AI vs Copilots | Shift from Assistance to Autonomy
https://dextralabs.com/blog/agentic-ai-vs-copilots/
[83] [84] [85] [91] [92] [93] AI Pricing in Practice: 2025 Field Report from Leading SaaS Teams | Metronome blog
https://metronome.com/blog/ai-pricing-in-practice-2025-field-report-from-leading-saas-teams
[86] [87] [88] [89] [90] How AI-Era SaaS Proves Value When Outcome-Based Pricing Fails
https://www.chargebee.com/blog/ai-saas-pricing-outcome-value-based-models/
[94] [95] The AI Cost Crisis: What AI Cost Sprawl Is And How To Fix It
https://www.cloudzero.com/blog/ai-cost-crisis/
[96] [97] [101] [102] Stablecoins in focus: navigating the new digital financial landscape
[98] Stablecoins payments infrastructure for modern finance - McKinsey
[99] Stablecoins Are Quietly Reshaping the Future of Payments - ETA SLF
https://etaslf.com/stablecoins-are-quietly-reshaping-the-future-of-payments-is-your-business-ready/
[100] How Stablecoins Are Transforming Global B2B Payments
https://www.williamblair.com/Insights/How-Stablecoins-Are-Transforming-Global-B2B-Payments