Fundamental
Fundamental builds Nexus, an AI model that predicts outcomes from enterprise tabular data.
On the record
Checkpoint: First public accuracy comparisons between Nexus and TabH2O or TabICLv2 successors on shared benchmarks
The call
Fundamental has funding, founders, and early customers, but sits in a prediction layer that rivals and cloud platforms can copy or bundle quickly.
Fundamental built software that uses AI to make predictions from business spreadsheet data, landing major customers and raising a lot of money. The problem is that Google, H2O, and others released similar tools within months, and the big data companies (Databricks, Snowflake, Amazon) can bundle prediction into platforms they already sell, which is hard to compete against. The company could turn things around if it moves beyond basic prediction to help customers actually make decisions and follow rules, creating something harder to copy than the prediction part alone.
Why
01
AI pressure: H2O.ai's TabH2O and open models like TabICLv2 and Google TabFM. Rival tabular prediction models launched within three months of Nexus, threatening to commoditize the exact prediction task Fundamental sells. Could AI do the core job? Yes, likely.
02
Bigger players: Databricks, Snowflake, and AWS SageMaker. These platforms already hold the enterprise data and could bundle native tabular prediction, cutting out a standalone vendor like Fundamental. Could a giant copy it? Yes, easily.
03
What protects it. DeepMind founder pedigree, $255M in funding, and live seven-figure Fortune 100 contracts buy time to build a stickier product layer.
What would change this call
Databricks or Snowflake ships a native tabular foundation model inside their data platform. Would show the core hyperscaler threat materializing, undercutting Fundamental's standalone value.
AWS launches a SageMaker native tabular prediction feature. Would show Fundamental's AWS partner is also its likely eventual replacement.
Fundamental announces a workflow or decisioning layer product with vertical integrations. Would show Fundamental moving up the stack to build defensibility beyond raw prediction.
AWS launches a SageMaker native tabular prediction feature. Would show Fundamental's AWS partner is also its likely eventual replacement.
Fundamental announces a workflow or decisioning layer product with vertical integrations. Would show Fundamental moving up the stack to build defensibility beyond raw prediction.
Key numbers
- $255M raised across seed and Series A
- $1.4B post-money valuation reported by TechCrunch
- H2O.ai shipped rival TabH2O three months after launch
Named risks
- Core tabular prediction layer commoditizing fast via open models like TabPFN and TabICLv2.
- AWS is both distribution partner and potential future competitor with overlapping SageMaker products.
- No workflow or decisioning layer shipped yet to create switching costs beyond prediction accuracy.
- No revenue disclosure or independently verified customer ROI data as of July 2026.
- Valuation reported inconsistently ($1.2B vs $1.4B) from the same February 2026 announcement.
Rivals, as read in July 2026
| Rival | Direction | Evidence |
|---|---|---|
| H2O.ai | contesting | Launched competing tabular model TabH2O at Dell Technologies World, May 19, 2026 |
| Google Research (TabFM) | contesting | Zero-shot tabular foundation model cited as top performer on 2026 benchmarks |
| TabICLv2 (Inria SODA, open model) | gaining on | Reported to outperform tuned gradient-boosting baselines, July 2026 benchmark |
| AWS, Databricks, Snowflake | gaining on | Already sell overlapping ML products (SageMaker Autopilot, Forecast) and control data platforms Fundamental depends on |
| DataRobot, Google Cloud AutoML Tables | contesting | Legacy predictive-AI incumbents explicitly positioned against by Fundamental as of July 2026 |
Questions people ask
- Is Fundamental profitable in 2026?
- No profitability figures have been disclosed. Fundamental raised a $255M Series A on February 5, 2026, and a third-party aggregator estimate of ~$17M revenue is unconfirmed and lacks a primary source, so no verified financial performance data exists as of July 2026.
- Who founded Fundamental and when?
- Fundamental was founded in October or December 2024 by CEO Jeremy Fraenkel and co-founder Gabriel Suissa, both DeepMind alumni. The company exited stealth on February 5, 2026, launching its Nexus tabular prediction model alongside a $255M funding announcement.
- How much funding has Fundamental raised?
- Fundamental has raised $255M total, disclosed February 5, 2026: a $30M Seed round and a $225M Series A led by Oak HC/FT, with Valor Equity Partners, Battery Ventures, Salesforce Ventures, and Hetz Ventures participating. No follow-on round has been reported through July 2026.
- Could AWS or Databricks replace Fundamental's product?
- AWS already sells overlapping products including Forecast, Fraud Detector, and SageMaker Autopilot, and is Fundamental's current distribution partner, a classic frenemy setup. Databricks and Snowflake also control the data infrastructure needed to bundle native tabular prediction, with no structural barrier stopping them as of mid-2026.
- Who is Fundamental's biggest competitor?
- H2O.ai launched a directly competing tabular foundation model, TabH2O, at Dell Technologies World in May 2026, just three months after Fundamental's own launch. Google Research's TabFM and open academic models like TabICLv2 also compete in the same tabular foundation model category.
- Is Fundamental's Nexus model actually novel?
- Nexus is described as a deterministic, non-transformer-based Large Tabular Model trained on billions of tabular datasets for tasks like fraud detection and underwriting, announced February 5, 2026. Independent benchmarking from July 2026 shows open models like TabICLv2 already matching or beating tuned gradient-boosting baselines.
Timeline
- 2026-07-03Independent benchmark shows open model TabICLv2 outperforming tuned gradient-boosting baselines
- 2026-06Fundamental headcount reaches 85, up 347.4% year over year
- 2026-05-19H2O.ai unveils competing tabular foundation model TabH2O at Dell Technologies World
- 2026-05-19Fundamental announces Tel Aviv hiring push for dozens of engineers to expand Nexus
- 2026-02-26Commentary flags tabular prediction layer as commoditizing, moat not yet secured
- 2026-02-05Fundamental exits stealth, announces $255M Series A and AWS partnership
- 2025-06Fundamental headcount reaches 19 employees
- 2024-10Fundamental founded by Jeremy Fraenkel and Gabriel Suissa
Revision history
29 July 2026
First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources
- Fundamental exits stealth as $1.2 billion unicorn, armed with $255 million for tabular (2026-02-05)
- Fundamental raises $255M Series A with a new take on big data analysis (2026-02-05)
- Fundamental Announces $255M in Funding and Publicly Launches its Most Powerful Large Tabular Model (2026-02-05)
- Management and Employees List
- Fundamental - Headcount, Funding & Open Jobs (2026-07-09)
- AI unicorn Fundamental expands Israel operations as it scales enterprise AI push (2026-05-19)
- Fundamental raises $255M Series A for enterprise data AI model
- H2O.ai Unveils tabH2O (2026-05-19)
- Tabular Models Benchmark: Performance Across 19 Datasets 2026 (2026-07-03)
- Large Tabular Models Are Here. Are They Ready for Insurance? (2026-02-26)
- DataRobot vs H2O.ai vs Google AutoML Compared (2026-07-17)
- Top AI Companies by Funding, Industry, and Business Function (2026-04-30)