Can you really grow in B2B in 2026 if you aren’t tapping into AI budget? The short answer: Probably not. Here’s the real data.
It may in fact be your last, best change to tap into it.
Something fundamentally different happened in 2025. AI budgets stopped being “innovation fund” line items and became core IT spending. And in 2026? They’re becoming the only growth budget that matters. AI is the growth engine not just of software, but the entire U.S. (and beyond) economy.
Let discuss what’s actually happening—with real numbers from the companies winning right now.
The Great Budget Shift: Where Enterprise Money Is Actually Going
Here’s the eye-opening stat from Andreessen Horowitz’s survey of 100 enterprise CIOs: Enterprise leaders expect an average of ~75% growth in LLM budgets over the next year. And as one CIO put it: “What I spent in 2023 I now spend in a week.”
But here’s what really matters for B2B founders: Last year, innovation budgets made up 25% of LLM spending. That’s now dropped to just 7%. AI has graduated from experiment to core operating expense.
What does that mean practically?
According to ISG’s research, while overall IT budgets are rising just 1.8% (basically inflation), AI spending is increasing 5.7%—with nearly a quarter of enterprises planning 10%+ increases. AI is now capturing 30% of the total IT budget increase despite being a fraction of overall spend.
Gartner forecasts global AI spending will approach $1.5 trillion in 2025, with enterprise software and infrastructure alone hitting nearly $500 billion in 2026. Meanwhile, worldwide IT spending will grow 9.8% to exceed $6 trillion for the first time—but software specifically is projected to grow 15.2% year-over-year.
Here’s the brutal math for B2B companies:
- Average IT budget increase: 2.79%
- Average vendor price increases: 9%
- Companies are underwater before they even think about new purchases
So where do they find budget for AI initiatives? They steal it from somewhere else.
The question is: Are you going to be stealing budget? Or getting stolen from?
The 2025 Stock Market Told the Story: The Great B2B Bifurcation
Want to see what happens when you tap into AI budget vs. when you don’t? Just look at the public markets this year.
The Winners (AI-Native or AI Infrastructure):
- Palantir (PLTR): +142% — AI-native, mission-critical government + commercial
- Cloudflare (NET): +80% — Edge computing + security + AI inference infrastructure
- MongoDB (MDB): +70% — Data platform positioned for AI workloads
- Shopify (SHOP): +51% — E-commerce platform riding AI and merchant growth (arguably not really an AI play yet, and just benefitting from e-commerce tailwinds and dominant market share)
- CrowdStrike (CRWD): +51% — Security is non-discretionary
- Snowflake (SNOW): +45-50% — Data infrastructure = AI infrastructure
The Losers (Traditional SaaS, SMB-Focused, or “Added AI”):
- HubSpot (HUBS): -51% — SMB CRM/marketing automation
- Bill.com (BILL): -43% — SMB financial operations
- Monday.com (MNDY): -36% — Work management
- Atlassian (TEAM): -34% — Developer tools
- Salesforce (CRM): -31% — Enterprise CRM (though Q3 may have been the turning point)
- ServiceNow (NOW): -26% — Enterprise workflows (but AI story emerging)
That’s a 4.7x difference in valuation multiple between high-growth AI companies and slow-growth traditional SaaS. The market is telling us something.
The Winners: What They’re Doing Right
Palantir: The AI-Native Template
Palantir’s story is remarkable. Revenue growth bottomed at 12.7% in Q2 2023—the exact quarter they launched AIP (Artificial Intelligence Platform). By H1 2025, revenue was up 44% year-over-year, with $541 million in net income (up 125%).
The company’s Q3 showed what AI-native means financially: 63% revenue growth with 51% operating margins.Their Rule of 68 score (growth rate plus operating margin) is exceptional. Free cash flow crossed $1 billion trailing twelve months.
Palantir’s CTO Shyam Sankar captures the thesis: “The models are converging while pricing for inference is dropping significantly. This only strengthens our conviction that the value is in the application and workflow layer.”
The lesson? Don’t add AI to existing products. Build AI operating systems from scratch.
ServiceNow: Turning Enterprise AI into Revenue
ServiceNow’s Q3 2025 shattered expectations with $3.41 billion in revenue, up 22% year-over-year. Their AI-related Annual Contract Value is approaching $500 million, and they’re targeting $1 billion in AI-specific revenue in 2026.
What’s working? They’re not just adding chatbots. They’re shipping autonomous AI agents that resolve complex IT outages and HR requests without human intervention. The “Xanadu” and “Yokohama” platform releases introduced agents that execute tasks, not just suggest answers.
CEO Bill McDermott calls it the “enterprise nerve center.” The TAM for ServiceNow is expanding to $275 billion by 2026 as they move beyond IT into AI-powered customer service, HR, and cross-enterprise workflows.
MongoDB: The Turnaround Story of 2025
MongoDB was left for dead in August 2025, trading at $214 with sentiment at rock bottom. Then came Q3: Atlas growth accelerated to 30% YoY, FCF jumped 306% to $140 million, and guidance was raised to $2.43-2.44 billion.
Why the turnaround? AI applications deal with messy, unstructured data—exactly what document databases handle well. As Menlo Ventures noted: “The companies winning in 2025 are infrastructure plays (Palantir, MongoDB, Cloudflare, Snowflake) rather than application plays.”
Infrastructure benefits from AI. Applications get disrupted by it.
Salesforce: Fighting Back with Agentforce
Marc Benioff went all-in on agents, and Q3 showed signs it’s working. Agentforce now serves 18,500 enterprise customers (up from 12,500 last quarter), running more than 3 billion automated workflows monthly and pushing past $540 million in ARR from agentic products.
They added 6,000 new enterprise customers in one quarter for Agentforce—while everyone else was talking about an AI bubble. Current remaining performance obligations are up 15% (versus 10% estimates).
The Futurum Group placed Salesforce at the top of its agentic AI platform rankings, slightly ahead of Microsoft. That matters for 2026.
CrowdStrike: Security = Non-Discretionary AI Budget
CrowdStrike is up 51% YTD despite a major outage in July 2024. Why? Security spending is non-discretionary. CIOs told Morgan Stanley they expect cybersecurity spending to grow 50% faster than overall software spending.
Their Falcon platform now includes endpoint, cloud workload, identity, and SIEM. 49% of customers use 6+ modules.Q3 showed subscription revenue up 21% to $1.17 billion, with net new ARR of $265 million (up 73% YoY).
Their AI SIEM business grew 95% in 2025. The ambitious target: $20 billion ARR by fiscal year 2036.
The Losers: What Went Wrong
HubSpot: The SMB AI Problem
HubSpot fell from $880 to $370—a 51% decline—making it the worst-performing major B2B stock of 2025. The irony? The business is actually performing fine: Q2 revenue hit $760.9 million (up 19.4%), and they raised full-year guidance to $3.08 billion.
The problem is perception. The market is pricing in maximum AI disruption risk for SMB-focused horizontal SaaS. When AI can potentially reduce the need for human marketing and sales users, seat-based pricing becomes a liability.
HubSpot launched Breeze AI in September 2024, including customer agents, prospecting agents, and content agents. But as Benioff acknowledged about Salesforce, agents could be “a victim of their own success”—if agents reduce the need for human users, you sell fewer seats.
The challenge for 2026: How do you monetize AI when your customers expect AI features without paying extra?
The Horizontal SaaS Trap
Look at the broader pattern: HubSpot (-51%), Monday.com (-36%), Atlassian (-34%), Asana (struggling). These are all horizontal tools for broad use cases.
Meanwhile, vertical solutions with deep domain expertise are defensible: Veeva for life sciences, Guidewire for insurance, Procore for construction.
The lesson: Horizontal tools are under pressure. Vertical solutions with domain expertise have moats.
The $2 Trillion Question: Where Is AI Money Actually Going in 2026?
According to Market Clarity’s analysis of Gartner data:
- Total AI spending: $2 trillion globally in 2026
- Over half goes to infrastructure (chips, servers, datacenters)
- Enterprise software: $500 billion as companies add AI to every app
- AI chips alone: $268 billion (up 28% from 2025)
The PwC survey of 308 US senior executives found:
- 88% will increase their AI budgets in the next 12 months
- Over 26% will increase by 26% or more
- 79% say AI agents are already being used in their companies
- 66% see real value through higher productivity
This isn’t tech companies only. It’s across all industries.
The New Rules for B2B Growth in 2026
Rule #1: Don’t Build SaaS and Add AI. Build AI and Add SaaS Economics.
The winners in 2025—Palantir, Cursor, Gamma—didn’t “add AI.” They are AI companies that solve specific problems. The architecture is different. The economics are different. The customer expectation is different.
Gamma reached $100M ARR profitably with just 50 people. That’s $2M ARR per employee. They’re competing against PowerPoint, which was invented before the first website existed. AI enables you to compete in markets that seemed settled—but you need to be 10x better, not 10% better.
Rule #2: AI-Native Companies Are 6-12x More Efficient
According to Iconiq Capital, AI-native companies show superior burn efficiency:
- AI-native at $100M+ ARR: 0.8x burn multiple
- AI-enabled: 1.6x burn multiple
- Non-AI median: 2.0x burn multiple
For companies under $100M ARR, AI-native burns 0.4x their net new ARR, while median burns 2.0x. This advantage compounds over time.
Rule #3: Some / Most Consumption-Based or Outcome-Based Pricing Really Helps
Seat-based pricing made sense when software helped humans do their jobs better. Consumption-based or outcome-based pricing makes sense when AI does the job instead.
HubSpot is moving toward credits-based pricing for Breeze Customer Agent. This transition will define who survives the next phase.
Leaders like Cursor, Replit, etc. have per seat pricing at their core. But the majority of their revenue in many cases, or a large portion of it, comes from usage beyond the monthly fee.
That’s not practical for pre-AI SaaS applications. But if you add so much value with your AI Agent, it just might make sense there.
Rule #4: Solve Problems That Can’t Be Ignored And Were Not Well Solved Before AI
Every winner solves problems that enterprises literally cannot deprioritize:
- Security (CrowdStrike, Zscaler, Cloudflare): You can’t turn off cybersecurity
- Data infrastructure (Snowflake, MongoDB, Oracle): You can’t do AI without data
- Operational AI (Palantir): You can’t run a modern enterprise without AI-powered decisions
- Commerce operations (Shopify): You can’t run an online store without your platform
So, Can You Really Grow in 2026 Without Tapping AI Budget?
Let me give you the honest answer based on all this data:
If you’re infrastructure: You’ll likely grow regardless, but AI workloads are accelerating that growth significantly.
If you’re mission-critical (security, data, compliance): AI is adding to your growth, not replacing it. Budget is flowing your direction.
If you’re horizontal SaaS with seat-based pricing: You’re fighting headwinds. Every customer expects AI features but doesn’t want to pay more. Meanwhile, AI-native startups are attacking your TAM with 6x your efficiency.
If you’re vertical SaaS with domain expertise: AI is an opportunity to deepen your moat. Domain-specific AI is defensible in ways horizontal AI is not.
If you’re SMB-focused: The market is pricing in maximum disruption risk. You need to prove AI enhances your value rather than cannibalizes it.
The bottom line: 76% of AI use cases are now purchased rather than built internally. Enterprise buyers want ready-made AI solutions. AI buyers convert at 47% vs. SaaS’ 25%, indicating AI delivers enough immediate value to short-circuit standard procurement.
The companies capturing this budget in 2026 will be the ones who built AI-native from day one—or successfully transformed before the window closed.
The window is closing.
