More resources

The 50 Fastest-Growing SaaS Companies of 2025: 10 Lessons

What the highest-growth B2B software companies reveal about modern plans, pricing, and monetization — and why static pricing is holding you back.

---

We analyzed the 50 fastest-growing B2B SaaS companies of 2025 (using Brex's annual ranking). The findings challenge conventional wisdom — and reveal a playbook that separates hyper-scalers from the rest of the SaaS world, helping explain the >5X difference in conversion and growth between top versus bottom quartile SaaS businesses.

1. AI SaaS leaders 'subscriptionify'

Despite the buzz around usage-based pricing and credits, 80% of the fastest growers still use subscriptions as their primary model.

Despite the increasing cost of APIs, AI-native businesses like ElevenLabs, Runway, and Gamma have found ways to "subscriptionify" inherently variable costs — using credit bundles, overage charges, and hybrid structures that give customers the predictability of a subscription while allowing revenue to scale with usage. The subscription model isn't going away; it's becoming more sophisticated.

2. Freemium is the norm, not the exception

This was one of the most striking findings: 85% of high-growth B2B subscription SaaS companies offer an always-free plan, compared to just 20–30% of SaaS overall.

Freemium doesn't guarantee growth. But sustained hyper-growth in B2B SaaS rarely occurs without it.

Why? Because a free plan removes the biggest barrier to adoption: the credit card. When there's no payment gate at sign-up, trial friction drops to near zero. Users try the product because there's nothing to lose — and that drives organic virality. Free users share links, invite teammates, create public content, and embed your product in their workflows. Each one becomes a distribution channel.

The best free plans aren't charity — they're conversion machines. They're designed to create habitual use, deliver repeated "aha moments," and generate frequent, natural opportunities to upgrade. PostHog, Linear, Supabase, and Framer all use generous free tiers to drive adoption, then gate on usage and collaboration features that naturally expand with team size and product investment. This is critical for AI-native businesses where low conversion of free plans is structurally unprofitable and unsustainable.

3. Three Four plans is the new standard

The old "Good / Better / Best" three-tier model is giving way to four plans. Among the companies we studied, the median number of plans is four, and 62% have four or more.

The typical structure: a free tier, two mid-range paid tiers differentiated by usage and features, and an enterprise plan with "contact us" pricing. That enterprise tier matters even if you don't have dedicated enterprise features yet — it signals positioning, enables a sales-led motion, and provides pricing flexibility.

Companies like Vercel, Supabase, and PostHog follow this pattern almost exactly: Free → Pro → Team → Enterprise. Each tier adds meaningful usage headroom and collaboration capabilities, creating natural upgrade pressure as teams grow.

Companies pushing past five plans tend to be those where value scales continuously (credits, tokens, compute minutes) and users self-segment naturally by spend — like ElevenLabs with its character-based pricing.

4. The pricing page is a marketing abstraction

To manage complexity for customers, the public pricing page is increasingly a simplification of what's actually happening underneath. The "hero table" — the comparison grid visitors see — intentionally masks far richer entitlement structures.

Think of it in three layers:

• The hero table: A clean marketing abstraction. One or two usage limits, key feature differentiators. Simple enough to understand in 30 seconds.
• Additional detail: Other usage metrics and features that differentiate plans and drive upsell, sometimes shown in expandable sections or comparison pages.
• Not publicly stated: Niche features, fair-use limits, and experimental gates that provide flexibility to test without committing publicly.

Vercel's pricing page, for example, shows a clean comparison of Hobby / Pro / Enterprise. But underneath, there are dozens of granular limits on build minutes, bandwidth, edge function executions, serverless function duration, and more — each acting as an independent upgrade trigger.

This layered approach lets companies iterate on entitlements without re-architecting their public offer — and test pricing changes with specific segments before broad rollout.

5. Pure seat-based pricing is almost extinct

Among the fastest growers, every single instance of seat-based pricing was part of a hybrid structure. No company relied solely on seats.

Why? Seats are easy to understand but poor at capturing the value most B2B SaaS products now deliver. As products become more workflow- and AI-driven, the value scales with usage, not headcount.

Usage-based components are now pervasive: 88% of subscription companies in the dataset employ some form of usage limit, and of those, 91% gate on three or more different metrics. One third combine usage gates with seat-based pricing in hybrid models.

Linear, for instance, gates on seats but also limits issue history depth and advanced features by tier. Fireflies.ai combines seat pricing with transcription minute limits and storage caps. The seat is rarely the only lever.

6. Soft gating is replacing hard feature paywalls

Traditional feature gating was binary: you either had access or you didn't. The fastest growers have moved to a more nuanced approach — soft gating — where lower-tier users get a degraded version of a premium feature rather than being locked out entirely.

Examples from the dataset: Runway gives free-tier users access to video generation but with a weaker AI model and watermarked exports. ElevenLabs lets free users generate voice content but with fewer characters, slower processing, and no access to premium voices. Ideogram provides image generation on free plans but with reduced queue priority and limited style options.

The logic is sound: soft gating lets users experience premium value, builds habit formation, and creates natural, contextual upsell moments — without the hard stop that can frustrate users and damage brand perception.

7. Usage limits are being "created" from features

This is a subtle but powerful pattern. Companies aren't just gating existing usage metrics — they're engineering new ones by decomposing features into measurable, limitable units.

A feature like "AI meeting notes" becomes a usage metric: "meetings transcribed per month" (Fireflies.ai). A collaboration feature becomes "active shared workspaces." PostHog decomposes its analytics suite into separately metered dimensions: events ingested, session replays, feature flag evaluations — each with its own limit and upgrade trigger.

The smartest implementations use per-day or per-week resets rather than aggregate caps. Why? Daily resets bring users back, create repeated upsell moments, and avoid the situation where a user hits their cap in week one and has no reason to return for three weeks. Ideogram's daily generation credits are one of many examples of this pattern — users return every day to use their refreshed allocation.

8. AI tools are bringing back the command line

Taking us back to the 80s/90s, AI-powered tools have made the CLI (command-line interface) an important new surface for monetization.

As developers and power users spend more time in terminals and AI coding assistants, companies like Cursor, Windsurf, and Vercel are extending their credit-based and subscription monetization directly into the terminal. When a developer exhausts their AI completions or build minutes from the CLI, the upgrade moment happens right where they're working — not on a pricing page they have to navigate to, and not in a marketing email they'll ignore. The terminal is becoming a natural monetization surface because usage, value, and the moment of friction all converge in the same context.

9. Throttle, don't cut: degradation is the new hard cap

This may be the most underappreciated pattern in the dataset. When users hit a usage limit, the fastest-growing companies don't simply cut access — they degrade it. Throttling, queue deprioritisation, model downgrades, and reduced functionality replace the hard wall, keeping users engaged while creating continuous, contextual upgrade pressure.

The degradation patterns are surprisingly varied and sophisticated:

Performance throttling: Vercel queues builds behind higher-tier users and reduces concurrency when limits are approached. Supabase throttles API requests and can even pause projects before triggering overage billing. These aren't billing events — they're runtime behaviour changes that users feel immediately.

Quality degradation: Runway reverts free-tier users to a weaker AI model and adds watermarks to exports when credits are exhausted. ElevenLabs removes access to premium voices and slows processing speed. The product still works — just not as well.

History and retention gating: PostHog reduces data retention windows and pauses session recording after caps are hit. Sentry stops recording error events entirely once the monthly limit is reached — meaning bugs go untracked. Fireflies.ai makes older transcripts inaccessible. These create urgency because the value of the product erodes over time, not just at the moment of the limit.

Soft-to-hard progression: Fathom Analytics and Framer show warnings before enforcement, giving users a chance to upgrade before the limit bites. Framer's progression — warning → publishing disabled → domain locked → CMS creation blocked — escalates gradually rather than slamming a door.

Why this matters: hard caps create a single frustration event. Degradation creates a continuous, lived experience of the gap between free and paid — which is far more effective at driving upgrades and far less damaging to the user relationship.

10. The fastest growers iterate constantly

Perhaps the most important pattern: hyper-scaling SaaS companies are evolving their plans, entitlements, and lifecycle flows at an accelerating pace. They treat pricing and plans as a product surface — not a one-time decision.

This is where the gap between fast and slow growers compounds. Companies that can test a new usage limit, change a gate, or launch a targeted upsell experiment rapidly are running dozens more experiments per year — and each experiment compounds into better conversion, expansion, and retention.

In B2C, Fyxer AI grew from $1M to $30M ARR in 2025, supported by a growth team that ran 514 experiments — more than two per working day. Their four-person growth squad alone shipped 360 of those. One of the successful ones — defaulting to a new higher-priced plan with annual billing communicated monthly — lifted month 0 revenue by 67%. That's not a pricing decision. That's a growth team velocity advantage.

Modern plans and pricing are a system to experiment with and layer on new features when released — not a decision to make once and revisit quarterly.

What this means for your business

Design for growth from day one. Don't hard-code your monetization logic. Build (or adopt) a system that lets you adjust entitlements, test gates, and run upsell experiments as fast as you ship product.

Gate more, but gate smarter. Soft gating, usage-based limits, degradation patterns, and per-period resets outperform binary feature walls. Create upgrade triggers that feel helpful, not hostile.

Treat pricing as a product. The companies that iterate fastest on monetization are the ones growing fastest. If changing a usage limit requires a code deploy, you're moving too slowly.

This is why we built RevTurbine — to give B2B SaaS teams the infrastructure to implement these best practices and continuously optimize monetization. The companies in this analysis throw growth PMs and engineering teams at the problem; RevTurbine is designed to make these same patterns accessible without that overhead. If you're looking to improve your free-to-paid conversion, book a free monetization audit and we'll benchmark your funnel against the patterns we've seen across hundreds of SaaS businesses.

---

RevTurbine is the monetization engine for B2B SaaS — helping founders and growth teams convert, expand, and retain customers faster. Learn more or join the beta.

---

Methodology: This analysis is based on Brex's 50 Fastest-Growing Software Companies of 2025. We examined the publicly available pricing pages, plan structures, and entitlement models of the 50 B2B subscription SaaS companies in the data-set. Conversion benchmarks referenced are sourced from the Growth Unhinged Free-to-Paid Conversion Report. All analysis reflects publicly observable data as of Q1 2025.