Last week, TIME hosted the first-ever TIME100 AI Leadership Forum in New York.

The TIME100 AI Leadership Forum, the newest extension of TIME’s growing Leadership Forum series, convenes the world’s leading organizations building AI solutions across various fields ranging from investment banking and venture investing to healthtech and media.

The following are my reflections from the forum and an eventful week in AI investment banking.

For the last three years, the conversations about AI have largely focused on efficiency: how to AI-enable organizations, how to move faster, how to do the existing work for less. Soon the conversation will turn to new solutions AI systems now make possible to build, and how to charge for them.

A few days ago, Ted Smith, co-founder and president of Union Square Advisors, joined Bloomberg’s The Close and discussed this shift. He compared this period to 2007, the year Union Square Advisors was founded, when the software industry was moving from the license-and-maintenance model to subscriptions. That transition was turbulent, Ted said, but the companies that doubled down on their core value propositions and worked through the pricing and revenue model changes came out the other side just fine. We are about to do the same thing with AI. The pricing model is going to change, and it will not be solely per-seat subscriptions anymore.

The data already shows the shift. In a survey of 240 software and AI companies, per-seat pricing fell from 21% to 15% of companies in a single year , while hybrid models, a base fee plus usage, rose from 27% to 41% to become the dominant approach. The reason is this: the seat was always a proxy for a person doing a job, and AI is cutting that link. At Alphabet, nearly 75% of new code is now generated by AI and reviewed by engineers, up from about 50% only a few months earlier. When the software increasingly builds itself, the customer does not need to add seats. So for what, exactly, should a vendor charge?

The answer, increasingly, is the result. Some vendors charge by usage, billing for the tokens an AI system consumes. Others have gone further, to outcome-based pricing, where the customer pays for what the software accomplishes rather than for how it’s delivered and used. When Zendesk introduced outcome-based pricing for its AI agents , it began charging roughly $1.50 for each support ticket the AI resolves on its own, with no human involved. Intercom prices the same way. This is not a niche experiment. Gartner projects that, in its best case, agentic AI could drive 30% of enterprise application software revenue by 2035, up from 2% in 2025 , inside a global software market that already exceeds $1.4 trillion a year . The seat is no longer the unit that matters.

The companies that get to participate in what comes next, the new business models and the repricing, will be the ones that did the foundational work first.

The hardest part of transitioning to AI-native systems is not adopting new ways of working. It is clearing out the legacy systems. Aging technology, the accumulated tech debt of decades, is weighing many companies down against newer, AI-forward competitors.

McKinsey estimates that tech debt amounts to 20% to 40% of the value of a company’s entire technology estate before depreciation, that about 30% of CIOs see more than a fifth of their new-product budgets diverted to servicing it, and that the companies carrying the most of it are more likely to see modernization efforts stall or fail outright.

We’ve been working on this challenge for some time, and this work has become one of our real advantages. Since its founding in 2007, Union Square Advisors has completed nearly 200 transactions representing more than $125 billion in value as of June 2026. For most of that time, the bank has operated as roughly a 30-person team. Over the last two years, with the leadership of the CEO, Mike Meyer, the firm shifted quickly to embrace AI, rebuilding the foundation of the systems and workflows we use from the ground up. Today, everything from our CRM platform to various internal databases to key operational systems and productivity tools is AI-native. We also launched the USA AI Solutions and Analytics offering that powers our team’s M&A and capital markets advisory work for both corporate and investor clients. We estimate that 50% of our analytics and deal workflows today are powered by our AI systems. In addition, we are finalizing the rollout of the firm’s AI Center in Q2, which includes an AI agent store – the same idea as an app store, where any team member can build an agent and share it across the firm.

What we built internally is an AI-first firm, the same shift many of the companies at the TIME100 forum are making in their own way. This is also the work we’re seeing that’s top of the list for some of the largest technology investors in the country. Today, investors are largely spending their time in two ways. The first is identifying and allocating capital toward the newest AI-native companies. The second is helping the companies they already own work down their tech debt so they stay competitive, enabling them to remain sellable when the cycle turns.

That cycle is already turning. The clearest place to watch it is the capital markets, which is where we work.

In the last few weeks, three of the most valuable private companies in the world have moved toward the public markets at once. SpaceX publicly filed its IPO prospectus with the SEC and, according to Reuters, is targeting a mid-June listing. The company expects to raise roughly $75 billion at a valuation near $1.75 trillion, which would be the largest IPO in history. Anthropic confidentially filed its IPO paperwork on June 1 , just days after raising $65 billion at a $965 billion valuation, and OpenAI is widely rumored to be preparing its own filing. Goldman Sachs forecasts that US IPO proceeds will quadruple to a record $160 billion this year , and that software now makes up about a quarter of the IPO backlog.

Ted Smith also shared his insights on the AI-driven IPO and M&A markets with Bloomberg The Close . His view is that a historic run of IPOs will be followed by a surge in M&A. The logic is this: It is hard to do large acquisitions in the months right before going public, when a deal can complicate the filing and regulatory process. But once these companies are public, they have both cash and stock to use for acquisitions, and there is a deep bench of strong companies sitting behind the three leaders. Ted shared that this is the start of a golden period for AI dealmaking, and that the pipeline is not “turning off anytime soon.”

From where we sit, neither is the work behind it.

Thank you to the TIME leadership team and congratulations on the successful launch of the TIME100 AI leadership series.