We live in a world where many devices around us — from phones and autonomous vehicles to planes and factory equipment — are continuously collecting, analyzing, and transmitting data.

Despite advancements in sensor technology, many organizations still struggle to capitalize on the immense volumes of data being generated. At Archetype AI, we believe that the key to unlocking this value lies in the strategic application of AI, which can transform raw real-world data into actionable insights. Join us as we explore how the convergence of sensors and AI can help us better understand the world around us.

‍

‍

The AI Data Dilemma: Not Enough or Too Much?

In recent months, many developers and GenAI early adopters have been discussing the scarcity of high-quality training data. “The generative AI boom of the past few years has led to tensions with the owners of that data — many of whom have misgivings about being used as AI training fodder, or at least want to be paid for it. As the backlash has grown, some publishers have set up paywalls or changed their terms of service to limit the use of their data for A.I. training,” writes NYT.

Sites like Reddit and Stack Overflow have started charging AI companies for data access, while publishers like The New York Times have taken legal action against companies like OpenAI and Microsoft for using their content without permission. In the meantime, in June 2024, Forbes accused Perplexity of plagiarizing a news article in its beta Perplexity Pages feature. Wired also accused Perplexity of illicitly scraping its website.

Companies such as OpenAI, Google, and Meta have gone to great lengths to gather data, including making deals with publishers, but ongoing data restrictions pose a threat to the steady supply of high-quality data and, as a result, the quality of GenAI tools.

‍

The Overlooked Opportunity: Real World Sensor Data

While there are certain restrictions on data from the internet, experts often overlook another source of data — real world data gathered from sensors used in manufacturing, agriculture, automotive, and almost every other modern industry.

The number of sensors is so high that this wealth of connected devices is often described by the term “trillion sensor economy.” It refers to a predicted future state where trillions of sensors are deployed globally, connecting the physical world to the internet and enabling massive data collection and analysis. This vision is an extension of the Internet of Things (IoT) concept, but on a much larger scale: analysts believe that the number of sensors worldwide will grow into the trillions, creating an unprecedented level of connectivity and data generation.

“We are birthing a “trillion-sensor economy” in which everything is being monitored, imaged, and listened to at all times. In this future, it’s not “what you know,” but rather “the quality of the questions you ask” that will be most important,” writes entrepreneur futurist Peter H. Diamandis.

Here are some examples of how sensors are leveraged today:

• In manufacturing, telemetry based on sensor data is used to report information about equipment, and according to the IDC report, IoT devices worldwide are expected to generate nearly 80B zettabytes (ZB) by 2025. Moreover, a 2019 IBM report stated that a typical factory at the time generated one terabyte of production data each day, and 90% of this data wasn’t leveraged.
• In the world of autonomous vehicles, a single car generates at least 25 GB of data per day. When Waymo tested the Jaguar I-Pace in late 2019, the SUV's powerful sensors generated a stream of information so large that an hour of driving produced more than 1,100 gigabytes of data, enough to fill 240 DVDs, Wired reports.
• In aviation, sensors on an aircraft collect over 300,000 parameters, with engine data being especially crucial. A commercial aircraft like the Boeing 737 generates 20 terabytes of engine data per hour. Some planes are equipped with up to 10,000 sensors per wing, measuring load, strain, ice accumulation, etc. Sensors are also placed across the fuselage, inside the cabin, and on the air.

The world has a wealth of data from sensors that capture unique physical world properties to expand human perception and world understanding. However, this data is highly fragmented and deployments are siloed, meaning that only small amounts of it end up getting used to solve narrow use cases.

The promise of Big Data — a huge buzzword in the early 2010s — was never realized. In 2020, NewVantage Partners published the Big Data and Executive Survey concluding the decade of the big data boom: it demonstrated that only 26.8% of firms have a data culture and only 37.8% of them are data-driven.

‍

From a Trillion Sensors to a Trillion Dollars

Currently, many organizations don’t have the right tools to work with this data or draw insights from it. We believe that AI will allow companies to get insights out of terabytes of real world sensor data. At Archetype AI, we are building Newton, the first universal AI foundation model that understands the physical world, to help enterprises access the sensor data insights that were previously obscured by data fragmentation.

The challenge lies in the diversity of sensor types and data formats. A lot of sensor data currently goes unused because of the complexity of interpreting the sensor’s complex signals. Our solution to this challenge is a foundation model that is sensor-agnostic. Newton is capable of interpreting data from a wide variety of sensor types, is adaptable to different industries and applications, and is scalable to handle the massive influx of data from the trillion sensor economy. In the video below, you can see one of many use cases for Physical AI, such as monitoring package handling, especially for fragile goods like vaccines:

The economic implications are significant. The value comes not just from the sensors themselves, but from the data they generate and the insights derived from that data. While we haven't yet reached a trillion sensors (this number is still in the tens of billions), the future impact on the global economy is undeniable. In 2020, McKinsey analysts wrote: “The potential economic value that IoT could unlock is large and growing. By 2030, we estimate that it could enable $5.5 trillion to $12.6 trillion in value globally, including the value captured by consumers and customers of IoT products and services.”

As we’re anticipating the changes the trillion sensor economy will bring, we’re investing in Physical AI to unlock its full potential. By making sense of the vast amounts of real world data, we're not just solving the AI data dilemma – we're working toward a better understanding of the world around us.

‍