It’s hot IPO summer, and the MANGOS are ripe
- company Google
- company Meta
- company Microsoft
- company Nvidia
- company SpaceX
- lab Anthropic
- lab OpenAI
- person Sam Bankman-Fried
A new cluster of technology companies, grouped under the acronym MANGOS, is preparing to enter public markets in a concentrated window, testing investor appetite and resetting expectations for what a public technology company looks like in 2026, according to a TechCrunch analysis [1]. The acronym MANGOS stands for Meta, Anthropic, Nvidia, Google, OpenAI, and SpaceX, though some analysts substitute Microsoft for Meta [1]. Half of these companies are reportedly heading to public markets simultaneously, creating what the TechCrunch Equity podcast hosts describe as a stress test for valuations and the broader IPO landscape [1]. Meta Platforms, the parent company of Facebook, Instagram, and WhatsApp, generated 97.8 percent of its total revenue from advertising as of 2023 and ranked 31st on the Forbes Global 2000 list that same year [10]. The company has faced sustained criticism since its founding in 2004 over issues including internet privacy, the Cambridge Analytica data scandal, and its role in the United States PRISM surveillance program [2]. In 2021, the company rebranded from Facebook, Inc. to Meta Platforms, Inc. to signal a strategic shift toward developing the metaverse [10]. The IPO activity arrives amid broader shifts in the artificial intelligence sector. A recent analysis of the Hugging Face Model Hub, which tracked 851,000 models and 2.2 billion downloads between June 2020 and August 2025, documented a fundamental rebalancing of economic power in open-weight AI development [4]. The study found that U.S. industry dominance by Google, Meta, and OpenAI has declined sharply in favor of unaffiliated developers, community organizations, and, as of 2025, Chinese industry players such as DeepSeek and Qwen [4]. Separately, a benchmark evaluating 72 language models on petroleum geoscience questions found that top performers achieved over 97 percent accuracy, with Gemini 3 Pro Preview reaching 99.8 percent [3]. The performance gap between open-weight and closed models was narrower than expected, with several lower-cost open-weight models exceeding 90 percent accuracy [3]. The Equity podcast episode also covered a reported $920 million-per-month compute deal between Google and SpaceX, which the hosts framed as a signal of leadership in the AI infrastructure race [1]. Additional topics included Apple's Worldwide Developers Conference announcements, a $250 million settlement, and Waymo's expansion into a market Apple had previously abandoned in its self-driving efforts [1]. The episode concluded with discussion of Sam Bankman-Fried's pardon request and a new Mark Zuckerberg biopic, with the hosts using ChatGPT to cast themselves in hypothetical roles [1].
funding
Background sources we checked (9)
- en.wikipedia.org ↗ Facebook (and parent company Meta Platforms) has been the subject of criticism and legal action since it was founded in 2004. Criticisms include the outsize influence Facebook has on the lives and health of its users and employees, as well as Facebook's influence on the way media…
- arxiv.org ↗ This paper presents FormationEval, an open multiple-choice question benchmark for evaluating language models on petroleum geoscience and subsurface disciplines. The dataset contains 505 questions across seven domains including petrophysics, petroleum geology and reservoir enginee…
- arxiv.org ↗ Since 2019, the Hugging Face Model Hub has been the primary global platform for sharing open weight AI models. By releasing a dataset of the complete history of weekly model downloads (June 2020-August 2025) alongside model metadata, we provide the most rigorous examination to-da…
- arxiv.org ↗ Large language models (LLMs) have raised hopes for automated end-to-end fact-checking, but prior studies report mixed results. As mainstream chatbots increasingly ship with reasoning capabilities and web search tools -- and millions of users already rely on them for verification …
- arxiv.org ↗ Despite widespread deployment of Large Language Models, systematic evaluation of instruction-following capabilities remains challenging. While comprehensive benchmarks exist, focused assessments that quickly diagnose specific instruction adherence patterns are valuable. As newer …
- arxiv.org ↗ The use of large language models (LLMs) in hiring promises to streamline candidate screening, but it also raises serious concerns regarding accuracy and algorithmic bias where sufficient safeguards are not in place. In this work, we benchmark several state-of-the-art foundational…
- en.wikipedia.org ↗ Ray-Ban Meta is a series of smartglasses created by Meta Platforms and EssilorLuxottica (parent company of Ray-Ban). The glasses include two cameras, open-ear speakers, a microphone, and touchpad built into the frame. They are the second generation of a line of smartglasses relea…
- en.wikipedia.org ↗ Meta elements are tags used in HTML and XHTML documents to provide structured metadata about a Web page. They are part of a web page's head section, the term meta indicating that they are a form of self-reference. Multiple Meta elements with different attributes can be used on th…
- en.wikipedia.org ↗ Meta Platforms, Inc. (doing business as Meta) is an American multinational technology company headquartered in Menlo Park, California. Meta owns and operates several prominent social media platforms and communication services, including Facebook, Instagram, WhatsApp, Messenger, a…
Sources covering this (2)
- techcrunch.com — It’s hot IPO summer, and the MANGOS are ripe ↗
- techcrunch.com — SpaceX, Anthropic, and OpenAI’s hot IPO summer · US