Larry Ellison's Keynote on Oracle's Vision and Strategy From Oracle AI World 2025, Explained

While OpenAI grabs headlines with ChatGPT breakthroughs and launches like Sora, there's another company quietly building the physical backbone of the AI revolution, and it's not who you'd expect.

Oracle has emerged as the infrastructure sugar-daddy behind Stargate, the $500 billion AI data center initiative first announced at the White House in January 2025 alongside President Trump. In September, the partnership doubled down with five new massive data center sites across Texas, New Mexico, and Ohio, bringing total planned capacity to nearly 7 gigawatts (enough electricity to power roughly 5 million homes).

The flagship Abilene, Texas campus is already operational with thousands of NVIDIA chips humming away, and according to OpenAI's finance chief, "No one in the history of man built data centers this fast."

So how did a database company traditionally known for enterprise software become the linchpin of AI's future? In a recent keynote, Oracle Chairman and CTO Larry Ellison laid out his audacious vision for why controlling AI infrastructure is the real game... and why the next wave of AI will be more valuable than anything we've seen so far. We watched Larry Ellison's hour and a half keynote on Oracle's vision for AI going into 2026.

‍

Below, we've got three ways to explore his vision:

• ‍Quick Overview: A fast summary of Ellison's two-phase AI revolution theory and Oracle's master plan.
• Key Moments with Timecodes: Jump straight to the most interesting parts of the keynote with our curated timestamps.
• Deep Dive: The full breakdown of how Oracle plans to modernize healthcare, banking, agriculture, and public safety using AI agents powered by your private data.

Here's What Larry Covered In Brief

Larry Ellison laid out the two major phases of the AI revolution:

• Phase 1: AI Training (The Dawn). This is what we’re living through now. Companies like OpenAI and Google are spending "vast fortunes" to train massive AI models on publicly available internet data. Ellison calls this the "largest, fastest-growing business in human history."
• Phase 2: AI Reasoning (The Real Value). This is what's next. Instead of just knowing about everything on the internet, AI will be used to reason on private, proprietary data, which he says "will be an even more valuable" market.

The big challenge is letting AI "think" with your private data without actually seeing or storing it. Oracle's solution is a new platform that lets AI models like GPT-5 securely access your private company data to create “AI agents” that don’t just answer questions, but automate entire business ecosystems.

The secret sauce is combining two key pieces of tech. First is Retrieval Augmented Generation (RAG), which lets an AI use real-time information—like your latest sales numbers—without having to be retrained. Second is Oracle’s AI Database, which converts all your company’s information into a special "vector" format that the AI can instantly search and understand, all while keeping your data private.

Here’s what these AI agents can actually do:

• For Healthcare: An agent can read a patient’s chart, check the latest medical research, and cross-reference complex insurance rules to recommend the best treatment that is also fully reimbursable.
• For Sales: An agent can identify which customers are most likely to buy a new product and automatically draft a sales email, complete with the three best customer references for that specific buyer.
• For Finance: A hospital’s AI agent could connect directly to a bank’s AI to secure a loan based on its upcoming insurance payments, automating the financing process.

This is part of a bigger trend Ellison explains as "vibe coding." Instead of writing code line-by-line, developers can now just describe the application they want in plain English. Using its new APEX AI Code Generator, Oracle plans to completely rebuild the Cerner healthcare platform—a job that took 25 years—in just three.

This vision expands to tackling some of the world's biggest problems by merging AI with robotics, biotech, and IoT.

Here’s a look at what’s coming:

• Healthcare in your home: AI-powered IoT devices will monitor patients at home as effectively as in a hospital, allowing for quicker, safer discharges. Even ambulances are getting an upgrade, becoming fully connected mobile clinics that stream data to the ER in real time.
• Disease detection, supercharged: A single blood test, analyzed by an AI-powered gene sequencer, could soon provide early detection for cancer and infectious diseases by scanning for circulating tumor DNA and pathogen DNA. A global network of these devices could serve as a pandemic early warning system.
• Farming without fertilizer: Ellison revealed AI-designed crops that are revolutionary. Think corn that pulls its own nitrogen from the atmosphere (goodbye, polluting fertilizer) and wheat that not only yields 20% more food but also captures atmospheric CO2 and converts it into a harmless mineral in the soil.
• The end of high-speed chases: Autonomous drones will revolutionize public safety. They can instantly detect forest fires with infrared cameras, find lost hikers, and follow fleeing vehicles, making dangerous high-speed pursuits obsolete.

The first AI wave was about building a massive, shared brain. The next, more lucrative wave will be about giving that brain a secure, temporary library card to access your company's most valuable secrets.

The Key Moments From the Talk

Here are the unique points of view, predictions, and insights from Larry Ellison's keynote on the future of AI.

The Dawn of the AI Era

• (1:43) Two Phases of AI: We're in the first phase, which is building and training enormous AI models. The second, more impactful phase will be using these models to solve humanity's biggest problems.
• (2:06) AI Models Mimic the Brain: Modern "multimodal" AI models are built like the human brain, with different neural networks specialized for different tasks like vision, language, and reasoning.
• (3:37) Fastest Growing Business in History: The business of training AI models is the largest and fastest-growing in human history, bigger than the industrial revolution.
• (4:53) The Real Opportunity: The truly world-changing opportunity isn't just training models, but using these "remarkable electronic brains" to tackle major challenges.
• (6:42) AI is a Tool, Not a Replacement: AI won't replace all human endeavors. Instead, it's an incredible tool that will make us much better scientists, engineers, surgeons, and more.
• (9:08) This Isn't a Bubble: While some companies may falsely label themselves as "AI companies" (similar to Pets.com in the dot-com era), the underlying AI technology is the highest value we have ever seen, by far.

AI, Data, and Oracle's Strategy

• (5:50) The Private Data Problem: Models trained on public internet data are limited. To reach their peak value, they must be able to reason on private data, which is Oracle's key focus.
• (28:36) The "Big Gotcha" of AI: People want to keep their data private but also want AI to reason on that private data. Oracle's AI Data Platform is designed to solve this contradiction.
• (31:26) How to Use Private Data Safely (RAG): The solution is a technique called Retrieval-Augmented Generation (RAG). Oracle's AI Database "vectorizes" private data—even from competitors like Amazon's cloud—making it securely available for AI reasoning without retraining the model on it.
• (35:48) Actionable Takeaway: Oracle used its own RAG technology to analyze its private customer data to predict which customers would buy new products. It then deployed an AI agent to automatically email those specific customers with customized references.
• (55:04) Oracle's Unique Position: Oracle is the only cloud company that provides both the scaled AI infrastructure for training models and the high-level enterprise applications to automate entire industries and ecosystems.

Future-Forward Predictions and Tangents

• (12:29) Robots Will Learn from YouTube: Robots will learn complex tasks like playing the piano by watching internet videos at high speed, mastering a piece by Chopin in about 5 seconds.
• (13:23) AI Surgeons Will Be Better Than Humans: Robots will be superior surgeons not because they're smarter, but because they have microscopic vision and superhuman hand-eye coordination, allowing them to cut between individual cancerous and healthy cells.
• (18:48) The Scale of AI Brains: We aren't building a 20-watt "meat computer" like the human brain; we're building a 1.2 billion-watt AI brain, powerful enough to light up a city of one million homes.
• (42:24) Hyper-Productivity: Oracle will use AI code-generation to rebuild the entire Cerner healthcare software suite—a project that took humans over 25 years—in just three years.
• (45:41) Automate the Entire Ecosystem: To truly innovate, you can't just improve one part of a system (like a hospital). You must automate the entire ecosystem (regulators, payers, patients, banks), just as Elon Musk had to build a charging network for Teslas to succeed.
• (1:00:30) The End of Identity Theft: Biometric AI will make passwords completely obsolete, eliminating identity theft and credit card fraud.
• (1:10:49) Pandemic Early Warning System: New metagenomic testing devices will be able to sequence everything in a blood sample. This will enable extremely early cancer detection and identify any pathogen instantly, creating a global early warning system for future pandemics.
• (1:15:13) Robotic Farming for the Future: Indoor robotic vertical farms are essential for feeding a growing world population. They will reduce water consumption by 90% and provide fresher, more nutritious food grown near cities.
• (1:22:34) Managing Climate Change for Free: We can use AI to engineer crops like wheat to perform biomineralization, a process that converts atmospheric CO2 into inert calcium carbonate (rock). This would allow us to actively and precisely manage the CO2 in our atmosphere at virtually no cost.
• (1:25:37) Eliminating Fertilizer Pollution: AI can help engineer crops to pull nitrogen directly from the atmosphere, completely eliminating the need for polluting nitrogen-based fertilizers.

The Deep Dive

In a sweeping vision for the future of artificial intelligence, Oracle Chairman and CTO Larry Ellison declared that the industry is at the cusp of a second, far more valuable phase—one that will pivot from training on public information to reasoning on private corporate data. While the current generative AI boom has captivated the world and created a market he calls "the largest, fastest-growing business in human history," Ellison argues it’s merely the foundation for a much larger transformation. The real revolution, he contends, will be unleashing AI on the proprietary data that companies have spent decades collecting, all while guaranteeing its privacy.

Phase One: The Dawn of an Era Built on Public Data

Ellison pinpoints the release of OpenAI’s ChatGPT 3.0 as the moment the AI revolution began "in earnest." It was the turning point when these complex systems "started sounding a little bit like us," moving from niche academic projects to tools with human-like conversational abilities. This first phase, which we are still in, is defined by AI Training.

"A series of companies are spending vast fortunes training these AI models on publicly available data on the internet," Ellison explained. He described this as the "Dawn of the AI Era," a period characterized by the creation of enormous multimodal models. These models, like OpenAI's GPT-5, Anthropic's Claude 4, and Google's Gemini, are analogous to the human brain. They are not monolithic structures but rather a collection of specialized neural networks working in concert.

He drew a direct parallel: just as the human brain has a visual cortex for processing sight and other areas for language, modern AI models use different neural networks for different tasks.

• Convolutional Neural Networks (CNNs) simulate the visual cortex, detecting edges and colors.
• Vision Transformers (ViTs) handle more complex tasks like holistic scene understanding and facial recognition.
• Transformer Networks, the foundation of Large Language Models (LLMs), are responsible for language and reasoning but require massive computational power.

These models are "multimodal," meaning they can process and understand a vast array of data types—text, images, audio, and video—scraped from the entirety of the public internet. The sheer scale of this effort, Ellison claims, has created a business growing faster than the railroads or the industrial revolution. It has sparked an arms race for computational power, leading to the construction of gigantic data centers filled with hundreds of thousands of GPUs.

However, Ellison is clear that this massive undertaking is just the beginning. The true value of these powerful "electronic brains" isn't in their creation, but in their application.

Phase Two: The Future is Reasoning on Private Data

"The one that will truly change the world isn't the creation of the models themselves," Ellison stated. "What will change the world is when we start using these remarkable electronic brains to solve humanity's most difficult and enduring problems."

This marks the transition to Phase Two: AI Reasoning. And the most valuable reasoning, according to Ellison, won't happen on public data, but on the private, high-value data that enterprises hold. "AI models reasoning on private data will be an even more valuable [business]," he predicted, calling it a "whole new world."

This introduces the central challenge of the next AI wave: the privacy paradox. Companies possess decades of invaluable information—financial records, customer histories, manufacturing processes, patient data, and proprietary research. Applying AI to this data could unlock unprecedented efficiencies and insights. An AI could diagnose medical images with superhuman accuracy, optimize a supply chain in real-time, or generate code perfectly tailored to a company's internal software architecture.

But there’s a “big gotcha,” as Ellison puts it: "AI models do not get trained on your private data... for some reason, people want to keep their private data private." No CEO will risk sending their company's crown jewels to an external model where it could be absorbed, leaked, or used to train a competitor's AI.

Oracle’s Answer: The Secure Data Platform

This is the problem Oracle has set out to solve. Ellison announced the New Oracle AI Data Platform, a system designed to bridge this gap. The concept is to allow an AI model to access and reason on private data without ever moving or compromising it.

The platform uses a technique known as Retrieval-Augmented Generation (RAG), where the AI model is given secure, temporary access to a company's private database. The model combines its vast general knowledge from public data with the specific context from the private data to generate an answer. For example, it can answer a question by consulting both Wikipedia and a company's internal sales reports from the last quarter.

"We give you the ability to add your private data to the model's library of information and knowledge," Ellison explained. "The model can reason across not just public data, but also private data, while keeping your private data private."

This platform is model-agnostic, meaning customers can bring their preferred AI model—whether it's from OpenAI, Anthropic, Google, xAI, or Meta. Oracle's role is to provide the secure infrastructure that makes this private reasoning possible. Ellison emphasized that because "most of the world's high-value data is already in an Oracle database," the company is uniquely positioned to enable this next phase.

The Unprecedented Scale of the AI Build-Out

To power both the ongoing training phase and the coming reasoning phase, the infrastructure build-out is happening at a mind-boggling scale. Ellison revealed that Oracle is constructing a "1.2 Billion-Watt AI Brain"—a massive data center for OpenAI in Abilene, Texas.

To put that number in perspective, 1.2 billion watts is enough electricity to power one million four-bedroom American homes. This single data center requires the energy of a "pretty good size city." It will eventually house half a million of NVIDIA's most advanced GPUs. Building such a facility goes far beyond simply erecting buildings; it requires creating an entire ecosystem of power generation, including on-site natural gas turbines, power transmission grids, liquid cooling systems, and ultra-fast networks.

This monumental investment is not unique to Oracle. Ellison pointed out that the "smartest engineers I know"—listing Elon Musk, Mark Zuckerberg, and Sam Altman—are all investing their personal fortunes to build similar capabilities. This, he argues, is the ultimate validation that AI is not a fleeting bubble like pets.com, but the most important technology in human history.

From Artificial Intelligence to Artificial Perception and Beyond

Expanding on the capabilities of these systems, Ellison suggested "Artificial Intelligence" is almost a misnomer. A better term might be "Artificial Perception," as these models can now see, hear, and even "smell" (detect chemicals) with incredible acuity.

He sees this leading to a future where AI robots become indispensable partners. An AI robot, he explained, could learn to scramble eggs or play a Chopin piece on the piano in five seconds simply by watching a YouTube video at high speed. Their true power, however, will be in high-stakes fields like medicine.

"AI robots will be much better surgeons than the best doctors," he declared. The reason isn't that they are "smarter," but that their physical capabilities are superior. A robot's vision is inherently microscopic, allowing it to see individual cells without an external device. This gives it the ability to perform a procedure like Mohs surgery, for removing skin cancer, with a level of precision impossible for a human hand, cutting perfectly between the last layer of cancerous cells and the first layer of healthy tissue.

Ultimately, Ellison's message was one of immense opportunity. The first wave of AI, built on public data, was about creating the tool. The second, more profound wave will be about securely applying that tool to the world's most important private data. For businesses, the implication is clear: the data you already own is about to become your most powerful asset in the age of AI. The challenge—and the opportunity—lies in finding a way to let the machine think with it.

The Foundation: A True Full-Stack AI Cloud

Oracle's central argument is that to truly unlock AI's potential, you need to control every layer of the technology stack. While competitors like Microsoft, Amazon, and Google provide powerful infrastructure and access to AI models, Ellison contends they are missing a critical component: deep, industry-specific enterprise applications. Oracle is unique because it operates at both ends of the spectrum.

This "full-stack" approach consists of several interconnected layers:

1. AI Infrastructure: Oracle has built some of the world's largest and most powerful AI training datacenters, providing the raw compute power necessary to train and run massive models.
2. Top-Ranked AI Models: Through its Oracle Cloud Infrastructure (OCI), the company offers access to the industry’s most advanced multimodal AI models, including OpenAI's GPT-5, Google's Gemini 2.5, xAI's GROK 4, and Meta's Llama 4. This gives customers a choice of the best "brains" for their specific tasks.
3. The AI Data Platform: This is the core of Oracle’s strategy. It's the bridge that allows the powerful AI models to securely interact with a company's most valuable asset: its private data.
4. AI-Powered Development Tools: With tools like the new APEX AI Code Generator, Oracle is changing how applications are built.
5. Connected Application Suites: Finally, Oracle brings it all together with its Fusion, Health, and Banking suites, designed to automate complex processes across entire industries.

The "Secret Sauce": Keeping Private Data Private with RAG and Vector Databases

The biggest hurdle for enterprise AI has always been data privacy. No company wants to send its sensitive customer lists, financial records, or trade secrets to a third-party AI to be used for training. Oracle's solution to this is built on two key technologies: Retrieval Augmented Generation (RAG) and its new AI Database.

Retrieval Augmented Generation (RAG) is a technique that allows an AI model to answer questions using information it wasn't trained on. Think of it like an open-book exam. Instead of relying solely on its pre-existing knowledge, the AI can "look up" answers in a specific set of documents you provide. This means it can access real-time information, like today's stock prices, the latest medical research, or a new government regulation, without that data ever becoming part of the model's core training.

To make this work, Oracle has transformed its flagship database into an AI Database with powerful vector capabilities. "Vectorization" is the process of converting any type of data—text documents, images, audio files, database entries—into a numerical representation (a vector) that AI models can easily understand and compare. The Oracle AI Database can automatically vectorize a company's entire trove of private data and create a searchable index.

Here’s how it works in practice:

1. A user asks a question, like, "What are our sales numbers in the northeast region for last quarter?"
2. The Oracle AI Database instantly converts this question into a vector.
3. It then searches its index of vectorized private data to find the documents with the most similar vectors—in this case, the relevant sales reports.
4. This retrieved information is then packaged with the original question and sent to a large language model (like GPT-5).
5. The AI model uses the provided data to formulate a precise answer.

Crucially, the private data is only used for the query itself and is never absorbed into the model or exposed to the outside world. As Ellison put it, the platform "accesses your private data while keeping it private." This ability extends beyond just the Oracle ecosystem; the vector embeddings can point to data stored anywhere, including other databases and even competitors' clouds like Amazon S3.

From Answering Questions to Taking Action: AI Agents and "Vibe Coding"

While retrieving information is powerful, Oracle's vision extends to generating action. This is where AI Agents and the new APEX Application Code Generator come in. An AI agent is an autonomous program that can perform complex, multi-step tasks. Instead of just answering a question, an agent can execute a workflow.

For example, a sales manager could ask, "Which of my customers are most likely to buy our new product, and can you help me reach out to them?" An AI agent could:

• Analyze all private customer data (purchase history, support tickets, communications).
• Identify the top prospects based on buying patterns.
• Search for the three most relevant customer references for each prospect.
• Draft a personalized email to each prospect, including the tailored references.
• Present the draft emails to the sales manager for approval before sending.

This level of automation is made possible by a new development paradigm Ellison referred to as "Vibe Coding." Driven by the new APEX AI Code Generator, developers can now declare their intent in plain English. You can simply state, "I need an application that tracks customer orders and sends a notification when an order ships," and the AI will generate the entire application—the user interface, the database connections, and the business logic.

This isn't just for simple prototypes. Oracle is using this technology for its most mission-critical software. The company claims this new declarative AI language improves developer productivity by a factor of 10. They are so confident in this approach that they have committed to rebuilding the entire Cerner healthcare platform—a project that took over 25 years of manual coding—in just three years using AI-generated code. The applications created this way are, by design, stateless, reliable, scalable, and secure, solving many of the problems that plague hand-written legacy code.

Modernizing the World, One Ecosystem at a Time

Ellison invoked what he called "Musk's Law"—the idea that modernizing an entire ecosystem is often easier and more effective than trying to improve it one piece at a time. To build a successful electric car company, Tesla couldn't just build the car; it had to build the entire ecosystem, including the battery factories and a global charging network.

Oracle is applying this same logic to enterprise software, with healthcare as its primary battleground. The goal is not just to automate a hospital, but to automate and interconnect the entire healthcare ecosystem. This includes the individual patient, the provider (hospital or clinic), the payer (insurance company), pharmaceutical companies, the banks that provide financing, and the government agencies that regulate it all.

The Oracle Health Suite is designed to orchestrate this complex dance. Imagine an AI agent that facilitates the entire patient journey. It assists a doctor by using RAG to pull the latest medical literature relevant to a patient's condition. It then cross-references the proposed treatment plan with the latest, ever-changing reimbursement rules from the payer (whether a private insurer or a national system like the UK's NHS). The system can then propose the "Best Care at the Highest Reimbursement Level Achievable," ensuring the patient gets optimal treatment and the provider gets paid. It can even go a step further and connect the provider to a bank to secure a short-term loan against those verified reimbursement receivables, solving a critical cash-flow problem for many clinics.

This ecosystem-level thinking is being applied across industries. Oracle is developing AI solutions for everything from preventing identity theft with biometric logins to designing new types of corn that pull nitrogen from the atmosphere, eliminating the need for fertilizer.

By providing every layer of the AI stack—from the datacenters to the models to the application suites—Oracle is making a bold bet that the future of AI isn't about selling tools, but about delivering fully automated, interconnected solutions that solve the biggest problems for the world's biggest industries.

Reinventing Healthcare from the Hospital to the Home

At the core of Ellison’s vision is a radical overhaul of the healthcare system. He argued that hospitals, despite their necessity, are dangerous places where patients are exposed to some of the "nastiest pathogens." The solution, he proposed, is to get people home faster and safer, a goal achievable through AI-powered remote monitoring.

"We can build these IoT medical devices where we can monitor you at home as well as we can monitor you in the hospital," Ellison explained. This involves a network of sensors that continuously track vital signs and other health metrics, with AI systems analyzing the data in real time to alert medical professionals to any potential issues. This constant stream of information doesn't just stop at the front door. Ellison revealed that Oracle is even developing AI-equipped, fully connected ambulances. In an emergency, these vehicles would function as mobile clinics, allowing EMTs to collaborate directly with ER doctors via audio, video, and digital data streams before the patient even arrives.

The diagnostic process itself is also poised for an AI revolution. Ellison shared a personal anecdote about breaking eight ribs in a motorcycle accident. During his MRI, the medical staff focused solely on counting the broken bones, ignoring the vast amount of other data the scan produced. "There was all this other data that that MRI produced; no one looked at it," he recalled. "That's always the case." AI, he argued, changes this paradigm. Instead of being limited by what a human is specifically looking for, AI can examine the entirety of a scan—be it an X-ray, CT, or MRI—to find anomalies and patterns that doctors might miss, leading to more accurate diagnoses and better patient outcomes.

Perhaps the most groundbreaking healthcare application Ellison discussed is metagenomic testing. He described a future where a single blood test, processed by an IoT-connected gene sequencer, could offer early detection for both cancer and a wide range of pathogens. The AI would analyze all the DNA in a blood sample, identifying not only the host's DNA but also "circulating tumor DNA" (ctDNA) from nascent cancers and the genetic signatures of any invading bacteria, fungi, or viruses.

This technology isn't limited to individual health. Ellison envisions a global network of these sequencers testing not just blood, but also wastewater and soil samples. Such a network would act as a powerful pandemic early warning system, capable of detecting novel pathogens like COVID-19 long before they cause a global crisis.

A Greener Future Through AI-Powered Agriculture

Ellison also set his sights on agriculture, an industry ripe for disruption. He introduced the concept of robotic AI greenhouses—massive, automated facilities that can grow food more efficiently and sustainably. "Growing indoors produces better-tasting, more nutritious, lower-cost food," he stated. These greenhouses reduce water use by an astounding 90% and, by being located near urban centers, drastically cut the CO2 emissions associated with transporting food over long distances. AI vision systems monitor the crops, deciding which specific plants are ready for harvest each day, a method that can produce a five times higher yield per acre compared to traditional farming, all while preserving natural habitats.

The innovation goes beyond just the method of farming. Ellison revealed that AI is being used to design entirely new types of crops. He highlighted two remarkable examples:

• AI-Designed Wheat: A new strain of wheat that not only increases grain yield by 20% but also actively combats climate change. Through a process called biomineralization, the plant is engineered to convert atmospheric CO2 into calcium carbonate (CaCO3), a stable mineral, effectively sequestering carbon in the soil.
• AI-Designed Corn: A variety of corn that eliminates the need for nitrogen fertilizer. It's designed to pull nitrogen directly from the atmosphere, a trick some plants like soybeans can do naturally. This would dramatically reduce the environmental damage caused by nitrogen runoff, which pollutes rivers and oceans.

In case you're interested what types of companies are working on these types of projects, we pulled this list of them.

Robotic AI Greenhouses & Vertical Farming

• Iron Ox: https://ironox.com
• Plenty: https://www.plenty.ag
• 80 Acres Farms: https://www.80acresfarms.com
• Bowery Farming: https://www.boweryfarming.com
• AppHarvest: https://www.appharvest.com
• Gotham Greens: https://www.gothamgreens.com
• Eden Green: https://www.edengreen.com
• Swegreen: https://www.swegreen.com

Nitrogen-Fixing Crops (AI-Designed Corn)

• Pivot Bio: https://www.pivotbio.com
• BioConsortia: https://bioconsortia.com
• Joyn Bio: https://joynbio.com
• Azotic Technologies: https://azotictechnologies.com
• Heritable Agriculture (Google X spinout): https://heritableag.com

Metagenomic Testing / Multi-Cancer Early Detection

• GRAIL (Galleri Test): https://grail.com
• DELFI Diagnostics: https://delfidiagnostics.com
• Geneseeq (CanScan): https://www.geneseeq.com
• Freenome: https://www.freenome.com
• Exact Sciences (CancerSEEK/Thrive): https://www.exactsciences.com
• Foundation Medicine: https://www.foundationmedicine.com

Enhanced Rock Weathering / Carbon Sequestration

• Everest Carbon: https://www.everestcarbon.com
• Groundwork BioAg: https://www.groundworkbioag.com
• CREW Carbon: https://crewcarbon.com
• Silicate: https://silicate.earth

Automating for a Safer, More Secure Society

Security, both personal and public, was another key theme. Ellison dismissed passwords as an "insane" and fundamentally broken security model. The future, he asserted, lies in biometrics. By using AI to recognize a person's face, voice, or fingerprint, we can create truly fraud-proof credit cards and secure computer logins. This wouldn't just be a convenience; it would have economic benefits, as reduced fraud would allow banks to lower credit card interest rates for consumers.

The vision extends to public safety with the use of autonomous drones. Ellison outlined several use cases that could transform emergency response and law enforcement. Drones equipped with infrared cameras can patrol vast areas and provide immediate detection of forest fires, allowing for a much faster response. They can also be deployed to search for lost hikers or to follow fleeing vehicles, eliminating the need for dangerous, high-speed police chases that endanger officers and civilians alike.

To support this, Oracle has developed systems like an autonomous air traffic control system for drones and an "RFID Specimen Vault" that allows drones to securely transport medical samples from a clinic to a laboratory, ensuring chain of custody and improving privacy.

Ellison's presentation was a powerful reminder that the true potential of AI may not be in creating more convincing digital content, but in solving the tangible, physical-world problems that have plagued us for generations. By modernizing foundational industries, he argued, AI reasoning can create a safer, healthier, and more sustainable world for everyone.