Building on open source AI? Hugging Face’s policy chief on the costs

The piece has to stay inside the podcast page’s facts and hit the section lengths, so I’m counting words against that source before the final text.The draft is over the length cap, so I’m tightening each section to the source facts only.Hugging Face Chief Policy Officer and SVP of Strategy Irene Solaiman spoke with TechCrunch senior reporter Rebecca Bellan on Equity, published November 4, 2026. The episode covers the costs, risks, and trade-offs of building on open source AI, with audio produced by Theresa Loconsolo for TechCrunch’s flagship podcast. TechCrunch Startups's report ties the interview to a mainstream shift: enterprises, governments, and startups are relying on open models for lower costs, more control, and more flexibility.

This piece is for founders, enterprise buyers, and government teams choosing whether to build on open models. It reports the November 4 conversation on cutting costs, when the most powerful models make less sense, and who makes money if models are free to use. It covers why the United States may need deep-pocketed players such as Nvidia for an open frontier model as Chinese open models gain ground, plus the security question from recent attacks on open AI platforms.

Building on: what actually changed

Open models are moving into the mainstream, the shift Equity’s November 4, 2026 episode examines. Enterprises, governments, and startups are adopting them for lower costs, more control, and more flexibility. Rebecca Bellan’s interview with Irene Solaiman asks how companies capture value once increasingly capable models are freely available. Chinese open models are gaining ground at the same time, and the episode asks whether American players can keep up. Broader adoption and tougher competition are the two changes the episode takes up.

Startups now face choices that come before the build. The episode covers what teams need to know before building on open models, from the right tools and infrastructure through the trade-offs in open AI. Recent attacks on open AI platforms add a security choice: how do you keep an open ecosystem secure without closing it off? The episode also asks who makes money when the models are free to use, and whether the companies building them can capture the value they create.

Building on: how it works

A team following the episode’s order picks tools and infrastructure before it settles on a model. Cost is the next filter. Startups, enterprises, and governments are switching to open AI models to cut costs, and the discussion includes when it makes less sense to use the most powerful models. Lower costs, more control, and more flexibility are the three results the switch is meant to deliver. The limit on the most powerful models is one trade-off startups are told to weigh before the build.

When the model is free to use, the episode turns to money. Who makes money, and can the companies building the models capture the value they create? Tools and infrastructure stay in that picture, because startups are told to settle those choices before they build. Security is a constraint taken from recent attacks on open AI platforms. Operators have to keep an open ecosystem secure without closing it off. Irene Solaiman takes up both issues as Hugging Face’s Chief Policy Officer and SVP of Strategy.

Building on: why it matters

Cost is why the set of buyers now reaches past startups. Enterprises and governments are switching with them, and all three are using open AI models to cut costs while seeking more control and more flexibility. The episode pairs that move with a limit: in some cases it makes less sense to use the most powerful models. Builders carry the other exposure. With increasingly capable models freely available, the question is who captures the value, including the company that built the free-to-use model.

Chinese open models raise the competitive stake. They are gaining ground, and the episode asks whether American players can keep up. Funding is the constraint the episode examines. The United States may need deep-pocketed players like Nvidia to build an open frontier model that can compete with China’s. The risk named beside that build is Chinese models becoming the global default. Recent attacks on open AI platforms add the problem of staying secure without closing it off as more buyers depend on open models.

Building on: who is affected

Startups, enterprises, and governments are switching to open AI models to cut costs, each seeking lower costs, more control, and more flexibility. Startups meet the change at the build. Guidance for them covers what to know beforehand: tools, infrastructure, and trade-offs. Enterprises and governments share the limit on the most powerful models, which make less sense for some uses. If Chinese models become the global default, all three sit inside that stake, next to the question of whether American players can keep up.

Model builders face a revenue question: who makes money when the models are free to use, and can those companies capture the value they create? Chinese open-model makers are gaining ground, and American players are asked if they can keep up. Nvidia is among the deep-pocketed players the United States may need for an open frontier model that can compete with China’s. Operators of open AI platforms have to secure the ecosystem after recent attacks without closing it off. Hugging Face is in that discussion through Irene Solaiman.

Building on: what to watch

Four questions from the November 4, 2026 episode remain open. Where does cutting costs stop favoring the most powerful models for startups, enterprises, and governments? Who makes money when models are free to use, and can builders capture the value they create? Why may the United States need deep-pocketed players like Nvidia to build an open frontier model that can compete with China’s, and what is at stake if Chinese models become the global default? After recent attacks, how do open AI platforms stay secure without closing the ecosystem off?

Theresa Loconsolo produced the November 4, 2026 episode. She joined TechCrunch in 2022 after working as one of 2 producers at a four-station conglomerate, writing, recording, voicing, editing, and engineering live performances and interviews, including lovelytheband. Based in New Jersey, she holds a bachelor’s degree in Communication from Monmouth University. Rebecca Bellan covers the business, policy, and emerging trends shaping artificial intelligence, with work in Forbes, Bloomberg, The Atlantic, and The Daily Beast. Equity is on Apple Podcasts, Spotify, YouTube, Overcast, X, and Threads at @EquityPod.