The $939B Question: Is AI Eating SaaS or Feeding It?

The Great Spending Showdown: AI vs SaaS in 2025/2026 — What Every B2B Leader Needs to Know

We’re witnessing the most dramatic shift in enterprise tech spending since the cloud migration began 15 years ago. And if you’re a B2B leader, founder, or investor, you need to understand what’s happening — because it’s about to reshape your entire world.

The numbers are staggering: AI spending is set to hit $644 billion in 2025, growing at a mind-bending 76.4% year-over-year. Meanwhile, SaaS — our tried-and-true darling — is projected at $295 billion with a “mere” 18.4% CAGR.

But here’s the uncomfortable truth the original convergence narrative missed: these markets ARE competing. And the evidence is mounting that AI startups aren’t just complementing SaaS — they’re actively hunting traditional SaaS incumbents for lunch.  Net net, few public SaaS leaders are actually seeing growth accelerate from AI.  Very few.

The Predator vs. Prey Reality

The AI Funding Explosion That Should Terrify SaaS Leaders

Let’s start with the brutal math that should keep every SaaS CEO awake at night:

2025 Reality Check: AI vs. SaaS Funding War

• $120+ billion in venture capital went to AI startups through Q2 2025 — with North American AI investments representing 86.2% of global AI funding ($79.74 billion)
• AI funding reached $5.7 billion in January 2025 alone
• Traditional SaaS funding continues steep decline in 2025 — with early-stage funding hitting lowest levels in 5+ quarters while AI dominates late-stage rounds
• AI startups earn 2-3x higher valuations across all stages compared to traditional SaaS companies
• 24 AI startups have already raised $100M+ rounds in just Q1 2025
• OpenAI’s $40 billion round in March 2025 — the largest private funding round in history — alone represents 8.5x the entire SaaS sector’s 2024 funding

Translation: The “smart” money isn’t betting on traditional SaaS players.  Not for the most part, at least.

The Billion-Dollar AI Unicorn Factory

The scale of AI startup funding isn’t just impressive — it’s existential for SaaS:

The AI Billion-Dollar Club:

• OpenAI: $8.4 billion total funding, $157 billion valuation
• Anthropic: $8.4 billion in funding
• xAI: $6.4 billion (two $6B rounds in 2024 alone)
• Sierra (Bret Taylor): $4.5 billion valuation after 175M round
• Scale AI: $29 billion valuation / “sale” with Meta’s $14.3 billion stake
• Waymo: $10.5 billion in funding

Compare this to the largest SaaS IPOs of the past decade — most topped out at $1-2 billion in total funding over their entire lifecycle.

The Klarna Experiment: Early Signal or Outlier?

The most talked-about case study in the AI vs. SaaS debate comes from Klarna, the $20 billion fintech that made headlines by dropping Salesforce and Workday in favor of AI-powered alternatives.

What Klarna reported:

• AI-powered customer service bot automated two-thirds of customer service chats, replacing 700 employees
• Query resolution time dropped from 11 minutes to 2 minutes
• Workforce reduced from 5,000 to 3,800 employees, targeting 2,000
• Average annual revenue per employee jumped to $700,000 from $400,000

CEO Sebastian Siemiatkowski’s take: “Thanks to AI agents + AI engineers getting prolific, you can rebuild most enterprise SaaS functionality, host for super cheap, and get basically 90%+ functionality”

The reality check: Klarna is still largely an outlier. The company later clarified it didn’t actually “replace” Salesforce with pure AI — it consolidated data onto its own tech stack using tools like Neo4j and built new interfaces. Industry observers like Josh Bersin remain skeptical about replicating complex systems like Workday’s payroll and compliance frameworks.

What makes Klarna’s case unique:

• Fintech with significant engineering resources
• CEO actively seeking attention ahead of 2025 IPO
• Company culture willing to experiment aggressively

The broader question: While Klarna grabs headlines, how many other enterprises have the resources, risk tolerance, and technical sophistication to attempt similar replacements? The answer today is: very few. But that could change rapidly as AI tooling improves and more case studies emerge.

The AI-Native Attack Pattern

Traditional SaaS companies face threats from three distinct AI-native attack vectors:

1. Vertical AI Dominance

Vertical AI startups captured over $1 billion in combined funding in 2025 YTD, surpassing infrastructure and horizontal AI categories. These aren’t general-purpose tools — they’re laser-focused solutions targeting specific SaaS incumbents:

Healthcare & Life Sciences: 14 companies in CB Insights’ AI 100

• Disrupting: Epic, Cerner, Veeva
• Advantage: Domain-specific data and regulatory expertise

Financial Services: AlphaSense ($1.4B funding) for market intelligence

• Disrupting: Bloomberg Terminal, FactSet
• Advantage: Real-time AI analysis vs. static dashboards

Legal: Harvey for legal agents, Caseflood for law firm operations

• Disrupting: LexisNexis, Westlaw
• Advantage: Natural language processing vs. keyword search

2. AI Agent Revolution

Nearly 90% of code at high-growth SaaS companies is now AI-generated, up from 10-15% just 12 months ago. This isn’t just about productivity — it’s about economic disruption:

The Agent Economics:

• Traditional SaaS: $3,500 per employee annually across organizations
• AI Agents: Can replace entire job functions for <$100/month
• The math may be devastating for high priced seat-based pricing models

3. The Infrastructure Bypass

Midmarket software companies are caught in a “pressure cooker” with fast-moving AI startups developing applications far more rapidly than traditional software companies on one side, and tech giants investing billions in proprietary AI tools on the other.

The squeeze play:

• Above: Microsoft, Google, Amazon bundle AI at costs small SaaS can’t match
• Below: AI-native startups develop faster with modern architectures
• Result: More than 100 midmarket software companies face survival threats in the next 24 months

The Real Convergence vs. Conquest Numbers

Let’s separate the happy convergence talk from the harsh conquest reality:

SaaS: The Steady Performer Under Siege

• $295 billion market in 2025 with 18.4% growth
• But: Public SaaS company valuations have retreated to 2016 levels while P/E multiples for the broader market have grown
• And: Salesforce shares fell 20% after weak earnings, marking the worst trading day since 2004
• Reality: Revenue growth rates for public SaaS companies are decelerating

AI: The Explosive Disruptor

• $644 billion in 2025 — already 2x the size of SaaS
• 76.4% growth rate vs. SaaS’s 18.4%
• Critical insight: Gen AI reached 2% market share in about a year vs. SaaS’s four years
• Projection: Gen AI primed for 10% of related spending by 2028 — three times faster than SaaS adoption

The Threat Matrix: Where SaaS is Most Vulnerable

High-Risk SaaS Categories

1. Workflow Management Tools

• Threat: AI agents can orchestrate workflows without UI-heavy tools
• Example: Orby AI observes enterprise processes and generates executable automations for complex operations

2. Customer Service Platforms

• Threat: Voice agents indistinguishable from humans handling phone-based jobs at a fraction of human cost
• Market size: $85 billion addressable market for business calls alone

3. Business Intelligence & Analytics

• Threat: Real-time AI analysis vs. static dashboards
• Problem: Most BI tools still require human interpretation

4. HR and Recruiting

• Threat: AI agents answering employee questions that typically consume 50% of HR professionals’ time
• Live customers: Coca-Cola, Sony, Puma using AI for HR automation

Defensive Positions for Incumbents

System of Record Applications: Still defensible due to:

• Data gravity and integration complexity
• Regulatory and compliance requirements
• Change management challenges

But even these are under pressure: AI could make switching between vendors much easier as agents manage the change for you

The Innovation Speed Gap

The most damning evidence against traditional SaaS? Innovation velocity:

AI-Native Development Speed

• Unsurprisingly, the primary reason buyers prefer AI-native vendors is their faster innovation rate
• Companies built around AI from the ground up deliver fundamentally better products with superior outcomes compared to incumbents retrofitting AI
• User evidence: Users who have adopted Cursor (gen AI-native coding) show notably lower satisfaction with GitHub Copilot, according to Andreessen Horowitz’s 2025 enterprise AI survey
• Scale example: One CTO at a high-growth SaaS company reported nearly 90% of their code is now AI-generated through Cursor and Claude Code, up from 10-15% with GitHub Copilot just 12 months ago

The Technical Debt Problem

Traditional SaaS companies face a brutal choice between two expensive paths:

1. Retrofit AI: The Compromise Path

• Legacy databases weren’t designed for vector embeddings and real-time ML inference
• Monolithic architectures struggle with the microservices needed for AI agent orchestration
• Authentication and security models built for human users, not autonomous agents
• Result: “AI features” that feel bolted-on rather than native

Real-world pain points:

• Salesforce’s Einstein took years to develop and still requires extensive configuration
• Microsoft’s Copilot integration across Office required fundamental rewrites of core applications
• Traditional SaaS companies report 2-3x longer development cycles for AI features vs. core functionality

2. Rebuild from Scratch: The Nuclear Option

• Admit 10-20 years of platform development may be architecturally obsolete
• Risk disrupting existing customer base during transition
• Compete with AI-native startups while rebuilding
• Capital requirements can exceed $100M+ for enterprise-scale platforms

3. The Team Reality Check: Your Biggest Blind Spot

Here’s the uncomfortable truth most SaaS leaders won’t admit: your current engineering team may not be equipped for AI transformation.

The skills gap is real:

• Traditional SaaS engineers excel at CRUD operations, API design, and database optimization
• AI development requires machine learning engineering, vector database management, and real-time inference optimization
• Your best backend engineer may struggle with transformer architectures and embedding pipelines

The speed problem:

• Teams that built your current platform over 5-10 years aren’t necessarily the right teams to rebuild it in 12-18 months
• Legacy codebases create institutional momentum toward incremental improvements rather than revolutionary changes
• Senior engineers may resist acknowledging their expertise is becoming obsolete

The hiring challenge:

• Top AI talent commands 40-60% salary premiums over traditional SaaS engineers
• They prefer working at AI-native startups where they can build from scratch
• Your existing culture and processes may repel the type of fast-moving talent you need

The brutal question: If you hired a team of AI-native engineers today, would they recommend rebuilding your platform from scratch? Most honest assessments say yes.

#4.  The AI-Native Advantage: Meanwhile, AI-native startups build with modern architectures optimized for intelligent behavior:

• Event-driven architectures that handle real-time AI decision-making
• Vector databases designed for semantic search and recommendations from day one
• API-first designs that treat AI agents as first-class citizens
• Distributed inference engines that can scale AI workloads elastically

The speed differential is devastating: While incumbents spend 18-24 months retrofitting AI into legacy systems, AI-native startups can build and deploy intelligent features in 3-6 months.

Example: Harvey (legal AI) built a complete legal research and document drafting platform in under 18 months. Traditional legal software companies like Thomson Reuters have spent years trying to add comparable AI capabilities to LexisNexis.

The Investment Reality Check

Where the Smart Money is Moving

AI Startup Investment Patterns:

• Over 70,000 AI-centric companies globally, with 25% based in the U.S.
• 34 of the 100 fastest-growing companies are AI-driven
• At least 23 private AI startups have raised over $1 billion

Traditional SaaS Investment Decline:

• Early-stage SaaS investment down significantly
• 60% decline in SaaS funding compared to AI startups in 2024
• VCs increasingly asking: “Why SaaS when you can fund AI?”

The Survival Playbook for SaaS Leaders

Given this existential threat, here’s what SaaS companies must do immediately:

Phase 1: Acknowledge the Threat (Most Haven’t — At Least Not Fully)

Stop believing the convergence narrative will bail you out. Your competitors aren’t trying to integrate with you — they’re trying to replace you.

Phase 2: Assess Your Vulnerability

High vulnerability indicators:

• Workflow-heavy products
• UI-dependent processes
• Manual data analysis
• Repetitive user tasks
• High customer service needs

Phase 3: The Emergency Response Strategy

For SaaS Leaders:

1. Accelerate AI integration — not features, but core functionality replacement
2. Prepare for pricing disruption — AI economics will force price compression.  And create 100s of new competitors, and improve the quality of your low-end competitors that historically were feature poor.
3. Prioritize data moats — your only sustainable defense
4. Plan for workforce reduction / reboot — if you don’t, AI competitors will undercut you. And / or blow past you.

For Investors:

1. Ask “Could an AI agent do this?”  
2. Favor AI-integrated SaaS over pure-play SaaS
3. Bet on teams that understand both paradigms

Phase 4: The Nuclear Option

For SaaS companies in high-risk categories: Consider pivoting to AI-native architectures entirely, even if it means cannibalizing current revenue.

Historical precedent: Companies that survived the cloud transition were those that disrupted themselves first.

Palantir: The One Public B2B Leader Really Getting It Right in AI

While most traditional SaaS companies struggle with the AI transition, one public company stands out as a masterclass in AI-native transformation: Palantir Technologies.

The numbers tell the story:

• Q1 2025: 39% revenue growth year-over-year, with U.S. commercial revenue surging 71%
• Stock performance: Up over 400% in the past 12 months, 63% YTD in 2025
• Commercial momentum: 136 U.S. commercial deals closed in Q1 2025 vs. 70 in Q1 2024
• 2025 guidance: $3.89-3.90 billion revenue (36% growth), crushing consensus estimates

What Palantir did differently:

1. Ready When AI Hit and Went All-In

When ChatGPT launched in late 2022, most SaaS companies scrambled to add AI features. Palantir was ready. They launched their Artificial Intelligence Platform (AIP) in mid-2023 and bet the entire company on AI transformation.

While competitors debated how much to invest in AI, Palantir made it their primary growth engine, restructuring their entire go-to-market strategy around AIP bootcamps and AI-driven customer success.

2. Dramatically Shrinking Time to Value for AI

Palantir’s AIP Bootcamps solve the biggest problem in enterprise AI: proving value quickly. In 5 days, they deploy working AI solutions on real customer data, converting prospects to paying customers at unprecedented speed.

Bootcamp results speak volumes:

• One major utility company signed a seven-figure deal just days after completing bootcamp
• Another customer signed during day one of bootcamp, then increased to seven figures weeks later
• On track to conduct bootcamps for 140+ organizations, with nearly half happening in recent months

3. Government-to-Commercial Transfer Success

Palantir leveraged their battle-tested government experience to dominate commercial markets. CEO Alex Karp’s insight: “Our early insights surrounding the commoditization of large language models have evolved from theory to fact.”

Their government contracts provided the R&D foundation that now powers commercial AI solutions — a moat competitors can’t replicate.

4. The AI Mesh Architecture

While other SaaS companies struggle with AI integration, Palantir built an “AI Mesh” that seamlessly connects:

• Foundry: Data operations platform
• AIP: Artificial Intelligence Platform
• Apollo: Autonomous software deployment
• Gotham: Intelligence and defense tools

This integrated approach means AI isn’t a feature — it’s the operating system.

The competitive advantage is real:

• Commercial revenue is closing in on government revenue for the first time
• U.S. commercial growth: 68% projected for 2025
• Customer expansion: From pilots to multi-million dollar enterprise deals in quarters, not years

What Other B2B Companies Can Learn

Palantir’s playbook proves three critical points:

1. Architecture matters more than AI features: Their 20-year investment in data infrastructure pays dividends when adding AI capabilities
2. Show, don’t tell: Bootcamps that deliver immediate value beat marketing slides every time
3. Bet the company on AI: Half-measures don’t work in platform transitions

The brutal reality: Palantir’s success isn’t just about having better AI — it’s about having an AI-first architecture that competitors can’t easily replicate. While other SaaS companies add “AI features,” Palantir operates as an AI platform with traditional software capabilities.

CEO Alex Karp’s prediction rings true: “The world will be divided between AI haves and have-nots. At Palantir, we plan to power the winners.”

The market agrees: Palantir has become the poster child for successful AI transformation, proving that with the right architecture and strategy, traditional software companies can not only survive the AI transition — they can dominate it.

The Bottom Line: This Isn’t Convergence — It’s War.  Be Honest Here.

The great spending showdown between AI and SaaS isn’t a friendly convergence story. It’s a high-stakes battle for the future of enterprise software, and traditional SaaS is losing.  Or at least, it’s not really winning.  

Most SaaS leaders have rolled out “AI Offerings” — but have not seen growth accelerate at all.  Salesforce, Atlassian, Box, even HubSpot and Monday have seen no boost from AI to revenue growth.   Not yet at least.

That incremental revenue is going to the new guard:

• AI startups are raising 20x more capital
• They’re developing solutions faster
• They’re delivering better economics
• Enterprise customers are already switching

The convergence narrative is comforting but false. Yes, some SaaS companies will successfully integrate AI. But many more will be displaced by AI-native startups that don’t need to retrofit intelligence — they were built intelligent from day one.

The companies that will dominate the next decade won’t be those playing defense with “AI features.” They’ll be the AI-native startups building intelligent software that makes traditional SaaS look as obsolete as desktop software looks today.

The future isn’t AI + SaaS.

The future is AI replacing Traditional SaaS.  Not entirely, but it’s already starting, and it’s accelerating.

Some categories have felt it much sooner — coding, support.  More will feel it soon.   But very few SaaS and B2B public companies outside of Palantir and perhaps ServiceNow are seeing any revenue boost from AI.

The future has already begun.

---

Sources: Gartner, ISG, McKinsey, Statista, Fortune Business Insights, Precedence Research, BCG, CB Insights, Crunchbase, Andreessen Horowitz, HFS Research, EY, AlixPartners