The a16z Fintech newsletter

CFO roundtable: AI growth, pricing, and forecasting (June 2025 Fintech newsletter)

Ivan Makarov and James da Costa

Posted July 2, 2025

This content first appeared in the June 2025 Fintech newsletter. If you’d like more commentary and analysis from the a16z Fintech team, subscribe here.

How CFOs are navigating growth, pricing, and forecasting in an AI world

Ivan Makarov, James da Costa

AI is fundamentally transforming enterprises, and few functions are feeling this impact more acutely than finance. CFOs are augmenting labor within their organizations with AI copilots, even as they face growing demands: managing rapid growth, navigating new cost structures and reporting, and making complex decisions around new pricing models.

Building on our partners’ work, including What “Working” Means in the Era of AI Apps and How 100 Enterprise CIOs Are Building and Buying Gen AI in 2025, we sat down and unpacked these changes with finance leaders from AI-native companies, including:

  • Dave Conte, CFO at Databricks: a unified, open analytics platform for building, deploying, sharing, and maintaining enterprise-grade data, analytics, and AI solutions at scale
  • Maciej Mylik, Finance at ElevenLabs: an AI audio research and deployment company making content universally accessible in any language and in any voice
  • Hanson Hermsmeier, VP of Corporate Finance at Together AI: an AI cloud platform that empowers developers and researchers to train, fine-tune, and deploy generative AI models
  • Matthieu Hafemeister, Cofounder at Concourse: a company building AI agents for corporate finance teams
  • Noah Barr, CFO at Ambient.ai: a computer vision intelligence company delivering AI solutions for simplified, automated security

AI isn’t just reshaping products and services — it’s redefining how businesses measure, forecast, and optimize financial performance.

1. Rethinking pricing: The shift from subscription to consumption and outcomes

As we wrote in December, AI is driving a shift toward outcome-based pricing. CFOs are now putting this into practice.

Pricing is increasingly tied to end results:

quote icon

The dramatic difference at Databricks is that our pricing and revenue recognition are based entirely on output, unlike input-based consumption models. If customers don’t derive value, they don’t consume, and revenue doesn’t appear on our P&L.

Dave Conte, CFO – Databricks

AI offers an opportunity to align incentives with your customers:

quote icon

Our pricing is built as a function of increasing profit in absolute terms, but decreasing margin in percent terms — we decrease unit prices as customer commitments increase. We automatically discount via our pricing calculator to encourage larger customer commitments. Locking customers into higher spend helps us de-risk revenue.

Maciej Mylik, Finance – ElevenLabs

For startups, this shift requires rapid experimentation:

quote icon

I changed my pricing more than seven times in the 40 days post-launch. It was really helpful to understand the market and customer’s appetite to pay. Right now, pricing is just a slide in my deck — something I’ll keep iterating on and improving.

Matthieu Hafemeister, Cofounder – Concourse

2. ARR needs reinvention

Traditional ARR metrics struggle to capture the reality of usage-based pricing models. CFOs are adopting hybrid metrics (e.g., ARR plus annualized usage) to reflect true customer consumption.

ElevenLabs tracks committed ARR plus annualized usage to avoid undercounting revenue from enterprise customers:

quote icon

We started annualizing the usage-based revenue and adding it to a new metric — ARR plus annualized usage — because enterprise customers exceed their quotas far more frequently. Without this, we’d be underselling or undercounting what we actually earn.

Maciej Mylik, Finance – ElevenLabs

CFOs and finance teams need to think about revenue recognition differently:

quote icon

With consumption-based models, how do you think about ARR? You might have a committed relationship, but actual usage varies month-to-month, so traditional ARR definitions become challenging.

Noah Barr – CFO – Ambient.ai

Databricks takes a sophisticated approach to consumption forecasting:

quote icon

Unlike SaaS models, where revenue tends to be linear, consumption models are inherently nonlinear — customers surge, then optimize. We manage this volatility by focusing heavily on diversification to avoid customer concentration. … We track contract values closely, and use AI to help understand and forecast our true consumption-based ARR.

Dave Conte, CFO – Databricks

3. Gross margin pressures and cost management

Nearly every AI startup builds on foundation models — such as those from OpenAI, Anthropic, or Mistral — which introduce significant variable costs that scale with AI model usage. Every API call, every token processed, adds to the company’s cost structure. This is a fundamental change in the underlying unit economics of pricing the AI service. The marginal cost of an additional user or usage is no longer zero — and, in fact, varies by user. And while inference costs are dropping dramatically overall, tasks that rely on the newest models with advanced reasoning capabilities still incur relatively high costs.

Changing cost bases mean optimizing, adjusting prices, or facing declining margins:

quote icon

We closely monitor infrastructure spend, and if costs grow faster than usage, we quickly send engineers to optimize it — there’s a continuous cycle of managing cost efficiency.

Maciej Mylik, Finance – ElevenLabs
quote icon

We triangulate pricing decisions through customer value proposition, competitive benchmarking, and cost and return analysis. Given how quickly AI infrastructure and software is evolving, we continually re-evaluate. … Creative pricing and packaging happens quite often based on customer need, term, and deal size — so long as we’re carefully considering the unit economics.

Hanson Hermsmeier, VP of Corporate Finance – Together AI

Startups training their own models must also navigate fixed GPU costs:

quote icon

You have to watch costs closely — these GPU costs are significant. We track regrettable idle GPU time as utilization loss, which directly impacts margin and efficiency. Every hour we have GPUs not being used by customers impacts our margins.

Hanson Hermsmeier, VP of Corporate Finance – Together AI

Or new types of costs, in the case of fine tuning:

quote icon

We have a human-in-the-loop (HILT) team that’s part of COGS. As algorithms improve, effective adjudications per human go up and unit costs come down, but we still have to bias toward false positives to manage risk.

Noah Barr, CFO – Ambient.ai

4. Assessing ROI in an AI world

As AI commoditizes some features, investing in the future is essential.

Invest in the future or risk being disrupted:

quote icon

Not every R&D project can be directly tied to immediate top-line revenue, but through predictive analytics, we measure how certain capabilities, like Unity Catalog, drive higher customer adoption and growth.

Dave Conte, CFO – Databricks

Hanson of Together AI emphasizes the strategic value of R&D:

quote icon

Research projects might not directly translate to immediate revenue, but they create significant long-term differentiation, product development, and stickiness, becoming essential in competitive markets. For example, the investment we made in research around kernels enables us to now provide a unique differentiation to our customers to reduce infrastructure costs and increase performance.

Hanson Hermsmeier, VP of Corporate Finance – Together AI

Long-term thinking is essential for maintaining competitive advantages:

quote icon

Pure text-to-speech will inevitably become commoditized. To maintain long-term defensibility, we need sophisticated product layers — workflows, data-rich features, and APIs — so customers become deeply embedded and find switching difficult.

Maciej Mylik, Finance – ElevenLabs

5. Leveraging AI for advanced financial forecasting

The only constant is change, but AI can help with that too.

It’s hard to plan when everything is changing:

quote icon

It can be a challenge to plan even 12 months out in AI right now. There’s constant innovation, and new use cases appear quickly. You have to remain agile and factor in change as part of your risk management strategy. … The sure constant in AI is change. Models evolve rapidly, and inference use cases we haven’t even considered today will be critical in a year.

Hanson Hermsmeier, VP of Corporate Finance – Together AI

Databricks actively uses its own product internally:

quote icon

We use Databricks itself — AI, machine learning and advanced analytics — to forecast consumption patterns at the customer, workload, and product levels. This is critical not just for financial forecasting, but also for accurately setting quotas for our large sales team. … You can’t achieve the precision required in consumption forecasting using Microsoft Excel — you have to use advanced analytics, machine learning, and AI to build those predictions.

Dave Conte, CFO – Databricks
quote icon

We’ve got a product, Genie, which is basically a natural language query. So you can type in natural language into your data lake. It extracts answers. Genie will understand your data. And then it learns and understands your data more and more and more, the more you use it.

Dave Conte, CFO – Databricks

Despite these advances, forecasting remains challenging:

quote icon

I don’t think anyone has fully cracked forecasting revenue for AI. The market is booming and changing so fast that reliable forecasting feels like more of a sanity check than precise predictions.

Maciej Mylik, Finance – ElevenLabs

Ivan Makarov is a portfolio finance partner on the early stage venture operations team at Andreessen Horowitz.

James da Costa is a partner at Andreessen Horowitz, where he focuses on investing in B2B software and financial services.

More from the a16z Fintech Team

In the Future of Finance x AI panel at New York Tech Week, Moment CEO Dylan Parker, Valon CEO Andrew Wang, and a16z general partner David Haber unpacked the impact of AI on fintech and infrastructure. The conversation covers how LLMs are reshaping enterprise software, bridging the gap between startups and incumbents, and why New York is becoming the epicenter for the next wave of fintech innovation. Listen and watch now »

At a16z’s first-ever Scout Sessions event, a16z general partner Angela Strange led a fireside chat around how the team approaches early-stage investing, including frameworks for evaluating startups and lessons learned from angel investing. She explains why AI is the ultimate unlock for building in “unsexy” domains like compliance. Listen and watch now »

In this episode of a16z Live — recorded live at New York Tech Week — a16z general partner David Haber sat down with Tennr cofounder Trey Holterman and Camber Health cofounder Christophe Rimann, both of whom are tackling some of the most entrenched, frustrating problems in healthcare. The trio discusses why operational rigor trumps demo theater, what it takes to build trust in high-stakes workflows, and the urgency behind integrating tech into healthcare now. Listen now »

What does growth really look like for the average AI company today, and what’s best-in-class? a16z partners Marc Andrusko and Olivia Moore compiled updated revenue benchmarks for gen AI startups in year one, based on dozens of AI companies. Read more »

Recent M&A Deals and Market Intel

Slide completed the year’s largest insurance IPO, raising $408 million in an upsized offering. Shares opened for trading on June 18.

Tipalti announced its acquisition of Statement, an AI-native treasury automation solution, on June 17. The acquisition deepens Tipalti’s treasury offerings with real-time cash intelligence capabilities to enhance cash flow visibility, insights, and forecasting.

Chime opened for trading on June 12, pricing nearly 32 million shares at $27 each — slightly above the proposed range of $24 to $26. The deal was driven by strong demand, with an order book reportedly 20 times oversubscribed, according to Bloomberg.

dLocal announced its acquisition of AZA Finance on June 3. AZA, a fintech company specializing in cross-border payments and foreign exchange solutions in Africa, had previously entered into a strategic partnership with dLocal in February.

Blend announced the sale of its title insurance business, Title365, to Covius on June 9. Blend had previously acquired Title365 from Mr. Cooper in 2021 for $422 million.

Want more a16z Fintech?

Sign up for commentary and analysis on recent news, and compelling trends in the fintech space.

Thanks for signing up for the a16z Fintech newsletter.

Check your inbox for a welcome note.

MANAGE MY SUBSCRIPTIONS By clicking the Subscribe button, you agree to the Privacy Policy.