EXPLORINEER
en

US — TechCrunch

Open source AI is booming. Hugging Face explains who profits.

AI1 min read

A TechCrunch podcast episode talks with Hugging Face's policy leader about how open-source AI models make money. It also covers whether American companies can keep up with Chinese open models.

What happened

TechCrunch's Equity podcast interviewed Irene Solaiman, Chief Policy Officer at Hugging Face. They talked about open AI models, which anyone can download and use. The talk focused on what startups should know before building on these models. It covered how to pick the right tools and the trade-offs involved.

Why open models are growing

Open models are AI programs whose code or model files are shared freely. Companies, governments and startups use them. They promise lower costs, more control over their systems, and more flexibility. As these models get better, a key question is how the companies behind them earn money.

Key facts

  • Open models are used by enterprises, governments and startups.
  • The main benefits named are lower costs, more control and more flexibility.
  • Chinese open models are gaining ground, according to the article.

Why it matters

Open models could change who gets to use powerful AI. Small teams can build on them without paying for closed systems. The article also raises a big question: can American companies keep up with fast-improving Chinese open models?

What to watch

  • How companies make money when their models are free to use.
  • Whether American players can compete with Chinese open models.
  • Which tools and infrastructure startups choose for building on open models.

Glossary

Hugging Face
A company that hosts and shares open AI models and tools that developers can use.
Open models
AI models that are shared freely so others can download and use them.
TechCrunch
A technology news website that publishes news, analysis and podcasts.

Ad

Add 150 words. Learn 10 a day.

Dicter word books send only the words you learn to your review notes.

AIZDNet Korea

Excelra Named a Leader in Everest Group's 2026 Pharmaceutical R&D IT Services PEAK Matrix® Assessment

Excelra, a global life sciences data and informatics company, was named a Leader among 34 evaluated service providers in Everest Group's Pharmaceutical R&D IT Services PEAK Matrix® Assessment 2026. Everest Group cited Excelra's combination of curated data, scientific AI, and domain expertise, including its GOSTAR® database of 37 million bioactivity data points. The company said recent acquisitions expanded its onshore scientific consulting capacity and global customer reach.

AIheise online

Saudi Aramco Takes Nearly Half of AI Company Humain

State-controlled oil company Saudi Aramco has acquired a 49 percent stake in Saudi AI company Humain, contributing about $2.6 billion along with existing AI, computing infrastructure and digitalization assets. The Public Investment Fund (PIF) keeps the remaining 51 percent. Humain plans to build data centers with a total capacity of three gigawatts by 2030, though it has not disclosed a specific investment budget.

AIVnExpress

OpenAI CEO Renames X Account to 'Ram Altman'

OpenAI CEO Sam Altman renamed his X account "Ram Altman" after backlash over remarks about RAM. He said he wants the world to have much more memory and that people should be able to install as much as they like in their computers, adding that he had long called for more memory production. Elon Musk mocked Altman's comments, while Tom's Hardware data showed average US DDR5 RAM prices rose about 500% over 12 months.

AITechCrunch

I tried Apple’s Siri Recaps. They’re surprisingly useful — and a little unsettling

Apple's new Audio Intelligence features, including Siri Recaps, are in testing with the first developer betas of iOS 27.3 and watchOS 27.3. The writer found Siri Recaps useful for capturing verbal details like deductibles and repair timeframes, but noted it mixed unrelated household topics into notes and recorded non-essential conversations. The feature's constant listening raised privacy concerns, even though Apple processes audio on-device when possible and does not identify speakers or store raw audio.