NextFin News - On February 5, 2026, San Francisco-based AI startup Fundamental emerged from stealth with a massive $255 million Series A funding round aimed at revolutionizing how enterprises analyze structured data. The round was led by Oak HC/FT, with significant participation from Valor Equity Partners, Battery Ventures, and Salesforce Ventures. Notable angel investors, including Aravind Srinivas of Perplexity and Olivier Pomel of Datadog, also joined the cap table. The capital injection will be used to scale the deployment of NEXUS, the company’s flagship Large Tabular Model (LTM), which is designed to provide predictive intelligence for the relational data—spreadsheets, ERP systems, and financial ledgers—that forms the backbone of the global economy.
According to TechCrunch, Fundamental was co-founded by alumni of Google’s DeepMind, including CEO Jeremy Fraenkel, who identified a "blind spot" in the current deep learning revolution. While Large Language Models (LLMs) have mastered unstructured data like text and images, they frequently struggle with the numerical distributions and non-linear relationships inherent in tabular data. NEXUS addresses this by utilizing a deterministic architecture that avoids the "hallucinations" common in transformer-based models. The company has already secured seven-figure contracts with Fortune 100 organizations and established a strategic partnership with Amazon Web Services (AWS), allowing customers to deploy NEXUS directly within their existing cloud environments using encrypted model weights.
The emergence of Fundamental highlights a critical pivot in the enterprise AI landscape: the realization that general-purpose LLMs are fundamentally ill-equipped for high-stakes predictive analytics. Traditional LLMs process numbers by tokenizing them into smaller chunks—treating "2.3" as three separate tokens—which destroys the model's inherent understanding of numerical magnitude and distribution. Fraenkel noted that while tools like Microsoft Excel integrations for LLMs are useful for writing formulas, they operate at the "text layer" rather than the "predictive layer." NEXUS, by contrast, is built to predict the next row of data, whether that involves forecasting equipment failure in a factory or identifying credit card fraud in milliseconds.
From an analytical perspective, Fundamental is challenging the decade-long dominance of "old-school" machine learning algorithms like XGBoost and Random Forest. These traditional methods require an "army of data scientists" to perform manual feature engineering—the process of selecting which variables a model should prioritize. NEXUS utilizes a foundation model approach, having been pre-trained on billions of real-world tables. This allows it to identify latent patterns across rows and columns that human analysts might overlook, effectively replacing months of manual pipeline construction with a single line of code. This "zero-shot" capability for tabular data represents a shift from bespoke, siloed models to a generalized intelligence layer for structured data.
The strategic partnership with AWS is particularly telling of the current market dynamics under the administration of U.S. President Trump, where domestic technological sovereignty and enterprise efficiency are high priorities. By allowing NEXUS to run as a "seller of record" on the AWS Marketplace, Fundamental bypasses the typical friction of enterprise procurement. Furthermore, the ability to deploy fully encrypted models within a customer’s own Virtual Private Cloud (VPC) addresses the growing regulatory and security concerns regarding data residency. In an era where U.S. President Trump has emphasized the protection of American intellectual property and infrastructure, Fundamental’s focus on securing the "predictive layer" of domestic industries—from water systems in Flint to PPE supply chains—aligns with broader national interests in resilience.
Looking forward, the success of Fundamental will likely trigger a "Tabular AI" arms race. While OpenAI and Anthropic have focused on the breadth of human knowledge, the next frontier of value creation lies in the depth of proprietary corporate data. If NEXUS can consistently deliver even a 0.5% increase in prediction accuracy for financial institutions, it could translate into billions of dollars in saved capital. As the industry moves toward 2027, expect to see a divergence in the AI market: one path continuing toward creative and conversational AGI, and another—led by firms like Fundamental—perfecting the deterministic, high-speed engines that automate the world’s most complex decision-making processes.
Explore more exclusive insights at nextfin.ai.
