The oil analogy is everywhere
We have probably all heard it in one conference keynote or another: “Data is the new oil,” sometimes accompanied by “and AI is the new internal combustion engine.” The intent is clear enough. Data is framed as an asset as economically consequential and transformative today as oil was in the early 20th century, while AI is cast as the force driving demand for data, much as cars once propelled the growth of the oil and gas industry. The analogy runs deeper than a throwaway metaphor. The IT sector has even adopted the language of oil and gas: We mine data, possibly from a data reservoir. That results in crude data that flows through pipelines to undergo refining, before putting it to productive use.
Why the oil metaphor breaks down
IT and AI are abstract businesses. A strong metaphor like this seems to make IT and AI easier to grasp. Yet, the oil analogy sends us down the wrong path, especially when thinking about business models and growth. The IT and AI sectors do not make money, or scale, in the way the oil and gas industry does. Oil is a consumable. Once an engine burns it, it is gone, leaving the user needing more. Data, by contrast, is not consumed. It can be copied and replicated at virtually zero cost. AI doesn’t “run on” data in the way an engine runs on fuel. AI runs on energy — which is why we increasingly discuss data centers in terms of their power demand. What AI does with data is not consumption but transformation: It processes data into other forms of data.
Data is not a commodity
There is another major difference: Oil is a commodity. There are different grades of crude, to be sure, but they serve the same purpose, and the fuels derived from crude are highly standardized. Data, especially processed data, is different. It behaves more like a specialty commodity. You often need many distinct types and qualities of data to produce a single, highly specific, actionable insight. Some data is more valuable than other data, and that value is tied to usefulness and scarcity. In commodity markets, where the product is consumable and standardized, success comes from maximizing production and market share. Calling data “the new oil” implicitly suggests that building a data or AI business means hoarding as much data as possible and maximizing processing capacity. That can be a profitable strategy, but it is not the only one, as I will show.
A better metaphor: data as metal
A more useful metaphor is to compare data to metals. Like metals, data comes in many types and can serve many different purposes. Raw data can be mined and, like metal ore, reduced before it becomes useful. And like metal, data is not consumed but used, reused, and recycled. When combined with other data, it can become more valuable, much as metal alloys are engineered for specific applications. In this metaphor, AI is the foundry and the forge: the machinery used to reduce, melt, mix, and shape data into the desired material and form — at a cost of substantial amounts of energy.
From a metal ore to a specialty alloy
There are commodity metals businesses that thrive by maximizing production and market share. But there are also precious metal and specialty materials businesses that succeed through either access to scarce and particularly useful metals or the capability to design highly specific, high-performance alloys. The same is true for data and AI. Companies like Google and OpenAI operate primarily in the commodity data business, thriving on vast quantities of raw data, enormous processing capacity, and the largest possible user bases. At Lux Research, we operate in the precious metals and specialty alloys business. We mine rare and hard-to-access data and combine it with commodity data to create decision-grade insights that help you catalyze, plan, and execute innovation with precision and resilience. So, stop thinking about data as the new oil. Data is the new metal. Success depends not on how much you extract, but on what you can make from it.