{"slug": "how-tabular-foundation-models-could-unlock-the-data-llms-can-t-handle", "title": "How tabular foundation models could unlock the data LLMs can't handle", "summary": "SAP's acquisition of tabular foundation model startup Prior Labs for over €1 billion ($1.16 billion) has spotlighted the emerging category of TFMs, which are designed to handle structured data like spreadsheets that large language models struggle to parse. TabPFN, released in 2023 and commercialized by Prior Labs in 2024, addresses this gap, with IBM's TabularFM and Google's TabFM also entering the field. Despite interest from major tech firms, TFMs remain niche, but could become a major part of enterprise AI.", "body_md": "[tabular foundation models](/tag/tabular-foundation-models/)\n\nLarge-language models have been the foundation for the AI boom, helping companies write code and analyze reams of corporate data that no one considered valuable until their debut. But a ton of corporate data sits in structured, tabular formats like spreadsheets that LLMs are less suited for parsing.\n\nTabular foundation models (TFM) were already on Google and IBM's radar as a possible solution to that problem, but SAP’s announcement last month that it was [ acquiring TFM startup Prior Labs](https://www.thestack.technology/superpowered-spreadsheets-sap-closes-buyout-of-tabular-ai-startup-prior-labs/) and investing more than €1 billion ($1.16 billion) into it brought a fresh spotlight to the category.\n\nTFMs first emerged around the release of TabPFN [ in 2023](https://iclr.cc/virtual/2023/oral/12541?ref=thestack.technology), and Prior Labs launched in 2024 to oversee and commercialise that model. TabPFN was trained on tabular datasets to address the problems that LLMs face trying to process data in an unfamiliar form and make predictions about future outcomes for enterprise users.\n\n“If you're using unstructured text, the problem has been solved. There are enough LLMs out there. When you talk about tabular data, [that] problem has not been solved,” Giuliano Pezzoli, CEO of Schema Labs, told *The Stack*.\n\nIn the years since TabPFN’s release, more researchers invested in the idea; IBM released its [ TabularFM framework](https://research.ibm.com/publications/tabularfm-an-open-framework-for-tabular-foundational-models?ref=thestack.technology) in late 2024 and Google unveiled its “zero shot” model TabFM in June 2026,\n\n[to process an entire dataset as a “single unified prompt.”](https://research.google/blog/introducing-tabfm-a-zero-shot-foundation-model-for-tabular-data/?ref=thestack.technology)\n\n__touting its ability__But despite the attention from some of tech’s biggest companies, TFMs remain a fairly niche part of the AI conversation. Wider understanding of how they work and what they can do, however, could make them a major part of enterprise AI sooner rather than later.\n\n## Get the full story: Subscribe for free\n\nJoin peers managing over $100 billion in annual IT spend and subscribe to unlock full access to The Stack’s analysis and events.\n\n[Subscribe now](https://www.thestack.technology/membership/)\n\nAlready a member? [Sign in](https://www.thestack.technology/signin/)", "url": "https://wpnews.pro/news/how-tabular-foundation-models-could-unlock-the-data-llms-can-t-handle", "canonical_source": "https://www.thestack.technology/how-tabular-foundation-models-could-unlock-the-data-llms-cant-handle/", "published_at": "2026-08-19 15:41:01+00:00", "updated_at": "2026-08-19 16:11:42.327633+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "ai-startups", "ai-products"], "entities": ["SAP", "Prior Labs", "TabPFN", "IBM", "Google", "Schema Labs", "Giuliano Pezzoli"], "alternates": {"html": "https://wpnews.pro/news/how-tabular-foundation-models-could-unlock-the-data-llms-can-t-handle", "markdown": "https://wpnews.pro/news/how-tabular-foundation-models-could-unlock-the-data-llms-can-t-handle.md", "text": "https://wpnews.pro/news/how-tabular-foundation-models-could-unlock-the-data-llms-can-t-handle.txt", "jsonld": "https://wpnews.pro/news/how-tabular-foundation-models-could-unlock-the-data-llms-can-t-handle.jsonld"}}