# Ahmad Jubran

> CEO of Gewan AI. Engineer, entrepreneur, operator. Two decades building technology across the US and Middle East. Building agentic AI systems, deploying capital, delivering at scale.

## About

Ahmad Jubran builds and invests in AI-native companies. He runs Gewan AI, an AI studio headquartered in Abu Dhabi that builds agentic systems, proves them in live operations, and deploys them for enterprises. His work spans three domains: building AI systems, deploying capital into AI ventures, and delivering at enterprise scale.

## Core Thesis

Three shifts are happening simultaneously:

1. **The Economic Shift** — The cost of creation has collapsed. AI inference costs dropped 99% in a year. What took teams of 50 now takes 5 with the right tools.
2. **The Organizational Shift** — The org chart is 150 years old. AI changes what one person can do. Design for judgment, not headcount.
3. **The Operational Shift** — Digital transformation mechanized what existed. AI enables the previously impossible. Redesign operations around intelligence.

At the intersection: "We don't make faster horses. We build engines."

## Writing

- [Building Engines: My Thesis](https://ahmadjubran.com/blog/building-engines): The full thesis on the future of work, capital, and technology
- [The Economic Shift](https://ahmadjubran.com/blog/the-economic-shift): How AI collapses the cost of creation and what it means for business models
- [The Organizational Shift](https://ahmadjubran.com/blog/the-organizational-shift): Why org charts built for the industrial era fail in an AI-native world
- [The Operational Shift](https://ahmadjubran.com/blog/the-operational-shift): Rebuilding operations around AI-native workflows
- [AI at the Energy Chokepoint](https://ahmadjubran.com/blog/ai-energy-chokepoint): How a Middle East oil shock ripples through global compute, data centers, and the case for nuclear
- [We Still Need Engineers](https://ahmadjubran.com/blog/we-still-need-engineers): AI is a great programmer but not a good engineer — why judgment, systems thinking, and taste still matter
- [The Era of the Entrepreneur](https://ahmadjubran.com/blog/era-of-the-entrepreneur): AI unlocks entrepreneurship by collapsing the cost of starting, building, and scaling
- [ARK Big Ideas 2026: The Convergence Is Here](https://ahmadjubran.com/blog/ark-big-ideas-2026-the-great-acceleration): Executive analysis of ARK Invest's annual technology forecast
- [The Intelligence Cost Collapse](https://ahmadjubran.com/blog/the-intelligence-cost-collapse): What happens when thinking becomes free
- [Stablecoins and the New Financial Rails](https://ahmadjubran.com/blog/stablecoins-and-the-new-financial-rails): Why this isn't a crypto story anymore
- [When AI Meets the Physical World](https://ahmadjubran.com/blog/when-ai-meets-the-physical-world): Robotics, energy, and the next trillion-dollar markets
- [Building for the Convergence](https://ahmadjubran.com/blog/building-for-the-convergence): Why thinking in systems beats thinking in sectors
- [Blog Index](https://ahmadjubran.com/blog): All published essays

## Reports

- [AI at the Energy Chokepoint](https://ahmadjubran.com/reports/ai-energy-chokepoint): How a Middle East oil shock ripples through global compute, data centers, and the case for nuclear. 18-page analysis with interactive flashcards and key points.
- [Building Engines: A Thesis](https://ahmadjubran.com/reports/building-engines-thesis): The complete thesis on the future of work, capital, and technology — combining the Economic, Organizational, and Operational shifts. 36 pages.
- [ARK Big Ideas 2026 Briefing](https://ahmadjubran.com/reports/ark-big-ideas-2026): Executive briefing on ARK Invest's Big Ideas 2026 — AI cost collapse, stablecoins, autonomous systems, convergence. 14 pages.
- [Reports Index](https://ahmadjubran.com/reports): All downloadable reports

## Interactive Library

### Algorithms to Live By
Interactive guide to computational thinking — learn by running simulations, not reading formulas.
- [Book Overview](https://ahmadjubran.com/library/algorithms-to-live-by)
- [Uncertainty](https://ahmadjubran.com/library/algorithms-to-live-by/uncertainty): How randomness shapes decisions — coin flip simulations and confidence intervals
- [Expected Value](https://ahmadjubran.com/library/algorithms-to-live-by/expected-value): Measuring decisions before seeing results — bet simulators and long-run outcomes
- [Distributions](https://ahmadjubran.com/library/algorithms-to-live-by/distributions): Normal vs power law — why the shape of uncertainty matters
- [Explore vs Exploit (Intro)](https://ahmadjubran.com/library/algorithms-to-live-by/exploration-vs-exploitation): The fundamental tradeoff between trying new things and sticking with what works
- [Optimal Stopping](https://ahmadjubran.com/library/algorithms-to-live-by/optimal-stopping): The 37% rule — how long to search before committing
- [Explore vs Exploit](https://ahmadjubran.com/library/algorithms-to-live-by/explore-exploit): Multi-armed bandit algorithms — Epsilon-Greedy, UCB1, Thompson Sampling
- [Sorting](https://ahmadjubran.com/library/algorithms-to-live-by/sorting): Bubble sort vs insertion sort vs merge sort — racing algorithms side by side

### How Neural Networks Learn
Interactive guide to the mechanics of machine learning — every weight, gradient, and activation is computed live.
- [Book Overview](https://ahmadjubran.com/library/neural-networks)
- [The Perceptron](https://ahmadjubran.com/library/neural-networks/perceptron): A single neuron draws a decision boundary — click to place data points and watch it learn
- [Activation Functions](https://ahmadjubran.com/library/neural-networks/activation-functions): ReLU vs Sigmoid vs Tanh — how nonlinearity shapes what networks can learn
- [Gradient Descent](https://ahmadjubran.com/library/neural-networks/gradient-descent): Watch a ball roll down a loss surface — learning rate controls everything
- [Loss Functions](https://ahmadjubran.com/library/neural-networks/loss-functions): MSE vs Cross-Entropy — why how you measure error changes how the network learns
- [Backpropagation](https://ahmadjubran.com/library/neural-networks/backpropagation): Watch error flow backward through a network diagram — gradients light up each edge
- [Weight Initialization](https://ahmadjubran.com/library/neural-networks/weight-initialization): Xavier vs He vs zeros — why the starting point determines if a network learns or dies
- [Optimizers](https://ahmadjubran.com/library/neural-networks/optimizers): SGD vs Momentum vs Adam racing on contour plots — watch momentum carry through valleys
- [Deep Networks](https://ahmadjubran.com/library/neural-networks/deep-networks): Draw patterns on a grid and watch activations propagate through layers
- [Overfitting](https://ahmadjubran.com/library/neural-networks/overfitting): Drag a complexity slider and watch the U-shaped test error curve appear
- [Regularization](https://ahmadjubran.com/library/neural-networks/regularization): Dropout, L2, early stopping — side-by-side comparison against unregularized baseline
- [Putting It All Together](https://ahmadjubran.com/library/neural-networks/putting-it-together): Full network visualization — watch weights initialize, activations flow, gradients propagate, and the network learn XOR

## Contact

- Website: [ahmadjubran.com](https://ahmadjubran.com)
- Email: ahmad.jubran@gewanai.com
- Company: [gewanai.com](https://www.gewanai.com)
- LinkedIn: [linkedin.com/in/ahmadjubran](https://www.linkedin.com/in/ahmadjubran)