The AI Pricing Reckoning: What Happens When the Music Stops?
We have been talking about nothing but economics and market forces lately, and it's no accident. It is budget season at US law firms, and these conversations mirror what's happening in boardrooms across the industry.
Innovation teams and IT departments are sharing their questions with us:
What does the future look like for AI tool pricing?
How much budget should we reserve?
Will technology costs rise or fall in 2026?
What's the right blend of tools, and how many seats do we actually need?
The Duopoly Everyone Is Talking About
Two platforms dominate conversations right now: Harvey and Legora. They have captured the mindshare and budgets of the world's most influential law firms. And, a handful of firms are considering Microsoft Copilot.
IT teams tend to prefer Copilot — it's simply easier to deploy when your entire infrastructure already runs on the Microsoft stack. Meanwhile, innovation teams and partners are drawn to Harvey and Legora.
This isn't just about features; it's about the power of marketing, the influence of investor-clients, and the safety of critical mass.
The rack rate pricings for these leading tools betray a profound economic paradox:
Copilot: ~$30/month per user
Harvey and Legora: At least ~$300/month per user
For large firms, this translates to potentially millions of dollars — a massive chunk of IT budgets that historically haven't been allocated to technology at this scale. Some are justifying these costs by positioning AI as a strategic expense or an HR expense. Sometimes the calculation doesn't even get that sophisticated. Firms simply recognize they need these tools to have credible conversations with clients about leveraging AI to manage cost pressure.
The One-Year Gamble
One important fact is flying under the radar: nearly none of these contracts are multi-year commitments. The bulk are one-year arrangements because firms are aware of a systemic risk and therefore keeping their options open, watching for who else enters the market and what might replace current solutions.
The risk of churn for these platforms is substantial, especially as OpenAI and Anthropic make their tools increasingly accessible. These foundation model providers could effectively replace the Harveys and Legoras of the world — which is exactly why these verticalized platforms are racing to prove their indispensability.
Right now, two simultaneous games are playing out:
Law firms are asking: How do we avoid vendor lock-in? What's our exit strategy if prices spike?
Vendors are asking: How do we create sufficient lock-in that firms will accept a minimum of 20-25% annual price increases?
Both sides are betting on different futures.
What Happens After One Year?
Since the verticalized tools are charging 10x (or greater) of the generalist tools, standard economics would suggest prices should decline.
Demand appears relatively saturated — Harvey claims to have 50+ Am Law 100 firms, while Legora likely has around 10-20%, with the remaining 30% as holdouts. This follows the textbook innovation diffusion curve.
Supply is increasing with new entrants. New entrants to the AI playing field are pricing at $100-150/month rather than $300, creating downward pressure.
So prices should fall, right?
Not so fast.
None of these companies are profitable.
The companies appear to be following the Silicon Valley playbook: spend heavily on marketing, make the tools feel indispensable, capture market share, and then raise prices. It's the Uber model — VCs subsidize the service until market dominance allows for premium pricing.
This free ride won't last forever.
The Uncomfortable Questions Nobody's Asking
Two truths should keep law firms awake at night:
First: There is no free lunch. VCs answer to limited partners who expect returns. AI companies need to turn a profit eventually, or someone will be left holding the bag.
Second: Nobody knows exactly when the inevitable will happen. The forces are dynamic and changing week by week as external pressures evolve.
If law firms have to pay the true, non-subsidized price for AI, they'll face a stark choice: fundamentally change their business model or find new revenue sources to bridge the gap. History shows it is far harder to find new revenue than to cut costs — just ask any management consulting firm.
The critical variable is VC patience. Let’s assume it will be several more years before AI products can turn a profit, then where will returns come from for the investors? Are there other portfolio companies that will cash out and satisfy limited partners in the interim?
The uncomfortable answer appears to be: not really. So much money right now is concentrated in AI. If AI doesn't return capital quickly enough, we could see a significant market correction.
What Happens When the Music Stops?
At the macro level, the legaltech market is a tiny fragment of the broader technology market. We will certainly be impacted by what happens if the bigger “AI bubble” bursts, see, for example, the analysis from Julien Garran of MacroStrategy Partnership, who argues that the current AI bubble is estimated to be 17 times the size of the dot-com bubble of the late 1990s and four times the size of the subprime mortgage bubble that triggered the 2008 global financial crisis. (https://futurism.com/future-society/ai-hype-investment-subprime-bubble).
At the micro level, every law firm should be considering the question: What happens when AI pricing changes, and your lawyers have already integrated AI into their core workflows? Will you be able to pay dramatically more money for AI, or will you have an alternative approach? Will CoCounsel, Lexis+ AI and vLex become adequate substitutes?
Most firms haven't seriously gamed out this scenario. Many are focused on adoption, on not falling behind, on next quarter's metrics. But a handful of the smartest firms are quietly building optionality:
training lawyers on multiple platforms,
developing relationships with multiple vendors,
ensuring data portability and IP ownership, and
creating their AI strategies that aren't dependent on any single provider's continued subsidy.
The AI revolution in legal services is real and transformative. But it is currently riding on the generosity of venture capital subsidies (which cannot last forever).
The music is still playing. But, what your plan for when it stops?
