Philippe Botteri, a partner at Accel, launched the firm’s GlobalScape report on the mainstage at WebSummit in Lisbon, Portugal, with the report highlighting the extent to which value on the public markets has concentrated in a small group of elite companies and how a new generation of native AI companies is rapidly accelerating.

Accel is one of the top three most-active investors on The Crunchbase Unicorn Board. The Silicon Valley firm has invested in 17 companies that have joined the board in the AI boom this year, according to Crunchbase data.

Its GlobalScape report, centered on the U.S., Europe and Israel, analyzed the surge in values in public and private AI-driven companies and the capex buildout required to keep growing in the next five years.

Here are some key takeaways from the report, which Botteri presented on stage in Lisbon, where Crunchbase News was also in attendance.

Public market concentration

The “Super Six” group of companies — Nvidia, Microsoft, Apple, Alphabet, Amazon and Meta — represent close to half of the current Nasdaq market cap as of October 2025. Altogether, that totals $20.7 trillion.

“I don’t think we have ever seen such a high concentration in the industry,” Botteri said. These six companies added $4.9 trillion in market capitalization between October 2025 and a year earlier, and showed $600 billion of operating cash flow in 2024.

Public cloud is up 25%

The public cloud index, a select list of U.S.-, Europe- and Israel-based companies built on the prior cloud wave, including UiPath, GitLab, Palantir Technologies and Figma was up 25% year over year as of October.

Those companies are all adding agentic capabilities to their products, Botteri said. However, it’s still early. “The models are probabilistic, so if you run a model 10 times in a row, given the probability nature of it, after 10 actions, you have a divergence.”

Botteri anticipates that we are 12 to 24 months from these tool capabilities delivering improved outcomes with advances in governance and security.

New generation of native AI

On the model and infrastructure front, notable investments for Accel include generative AI company Anthropic, small model developer H Co., and publicly traded AI infrastructure provider Nebius Group, along with numerous investments on the application side.

The U.S. dominates on the foundation model side of AI, but there are also more specialized models — which don’t require tens of billions of dollars to be developed — where Europe can contribute, said Botteri, citing portfolio company H, which has developed a computer-use model to take actions on the computer on behalf of a user.

“On the application [side], it’s very much a level playing field,” he said. The firm’s report claims that European and Israeli AI and cloud investments are two-thirds the size of the U.S. investment market, excluding model companies.

A new generation of native AI applications are growing at an unprecedented rate. Botteri named a few AI native Accel portfolio companies, including coding company Anysphere, maker of Cursor; AI search engine Perplexity; Stockholm-based vibe coding startup Lovable;  Berlin-based business automation platform n8n; Synthesia, which offers AI video creation for the enterprise; and Israel-based security company Cyera.

Energy shortfall

The report also outlined the energy shortfall to deliver AI over the next five years — around 117 gigawatts — the equivalent to powering Italy, Spain and the U.K. combined, said Botteri. For context to understand the buildout required, a nuclear power plant creates 1 to 2 gigawatts of power.

The Super Six group of companies, who are investing in much of this infrastructure buildout over the next five years, are expected to generate around $5.5 trillion in operating cash flow.

That operating cash flow along with the debt markets, will go a long way to addressing the $4 trillion required to build out this capacity, Boterri said.

For this infrastructure buildout, the report estimates the revenue payback should be $3.1 trillion, an increase of 1% to 2% of compound average GDP growth per year.

“If you don’t think that GenAI is going to generate a 1%-2% increase in the global GDP,  then I’m not sure why we’re doing all this,” said Botteri.