Using Technology to Drill Smarter and Power the Future

The oil and gas industry is one of the largest sectors in the world, valued at $6.10 trillion in 2024 and projected to surpass $8.79 trillion by 2034. The sector also generated more than $4.2 trillion in estimated global revenue last year.

Powering everything from our morning commute to the factories that make our stuff, oil is the backbone of modern society. It heats our buildings, generates electricity, and keeps industries humming. Beyond energy, oil serves as a raw material for products we rely on daily. Medicines that keep us healthy, plastics in everything we touch, and chemicals that make countless products possible all trace back to oil. Oil literally moves the world.

Oil, however, is a non-renewable fossil fuel that has a significant environmental impact, including greenhouse gas (GHG) emissions and climate change. It also presents geopolitical challenges due to the concentration of oil reserves in a few countries, thus threatening global energy security.

Oil Is Finite: What It Means for Production & Prices

Being a non-renewable resource, crude oil cannot be replaced naturally at the rate that it is consumed. This is because the finite and rapidly depleting resource is formed over millions of years from the decomposition of organic matter, plants, and animals, under extreme heat and pressure.

It is a hydrocarbon resource, which means an organic compound made exclusively of carbon and hydrogen atoms. Hydrocarbons make up the basis of oil, natural gas, and coal, which are naturally occurring substances found in rock in the earth’s crust. 

The total amount of oil on Earth is limited. And as oil is extracted and used, the available reserve decreases, creating concerns about future supply.

Not to mention, not all the oil in a reserve is pumped out. A reservoir typically recovers anywhere from a few percent to over 50% of its oil, depending on reservoir pressure, the quantity of dissolved gas, rock properties such as porosity and permeability, and the recovery techniques employed.

Primary recovery, which is the initial extraction, often yields only about 20% of the oil. Enhanced recovery methods like waterflooding or gas injection can significantly increase the total yield. Companies use computer simulations to model reservoir behavior, evaluate different production scenarios, devise injection strategies, and estimate oil recovery over time.

While limited, new reserves are also constantly discovered and extracted using new and improved technology that allows for more oil recovery from existing fields.

In order to extract crude oil, drilling and pumping techniques bring it from underground reservoirs to the surface. First, tools are used to locate and assess reservoirs. Once a viable one is found, drilling techniques create a well from which oil is pumped out, often using structures like oil rigs.

Oil rigs drill deep holes into the earth to create wells that extract petroleum. However, when drilling reveals insufficient hydrocarbons to be profitable for the operator, that’s called a dry well.

Interestingly, an oil well can run dry even when measurements are pointing to there still being oil. To determine if a well contains oil, multiple methods are employed, ranging from large-scale regional surveys to direct sampling at the drill site. 

Before starting any drilling, operators use geological and geophysical surveys to identify areas with the potential to hold hydrocarbons. Geologists study surface features, subsurface structure, and rock types to locate potential oil and gas reserves.

Other techniques include gravity surveys, which measure subtle changes in the Earth’s gravity, and magnetic surveys, which measure magnetic anomalies. Geologists also analyze surface soils and vegetation to find trace amounts of hydrocarbons that may have seeped up from deeper reservoirs.

Among these methods, seismic surveys are particularly important. They use sound waves that travel through the Earth to reveal the location of oil deposits and provide estimates of a reserve’s approximate size.

Still, it’s pretty common for an oil well to go dry just after a portion of the expected oil has been extracted. 

Why ‘Dry’ Wells Happen—and How 4D Seismic Fixes It

Finding oil, a critical and limited resource, is complex. Moreover, the negative environmental impact of oil and gas extraction adds another layer of urgency: drilling needs to be smarter and more efficient. This requires more accurate measurements of how much oil a reservoir actually contains.

A group of researchers from Penn State University developed more precise calculations of how much oil a given well will actually produce. Their work addresses a key question: why does a well go dry even when seismic scans indicate there’s still oil underground?

“We actually tested … data from the North Sea. You know, they started drilling in 2008 and based on their estimation … they could produce oil for 20 years, 30 years. But unfortunately, after two years, there was nothing. Their well is dry. They just got confused. Where is the oil? Gone? The big issue actually is the complexity of the geology in the reservoir.”

– Study author Tieyuan Zhu, a geophysicist from Penn State

So, Zhu, along with his students and postdoc fellows, decided to study more details about the data from sound measurements than previously used.

This meant the team needed more computing power as well as large memory to store parts of the problem in the computer’s processors, avoiding time-consuming and costly trips back to data storage.

The solution was PSC’s flagship, National Science Foundation-funded Bridges-2 supercomputer, which enables data-intensive research by integrating new technologies for converged, scalable HPC, machine learning, and data analysis.

It has over a thousand powerful CPUs in its hundreds of regular memory nodes that provide the speed for general-purpose computing and data analytics. The supercomputer could also provide the memory, as each of its CPU nodes contains 256 GB to 512 GB of RAM, which is 8 to 16 times that of an advanced gaming laptop.

Additionally, it has Extreme Memory (EM) nodes, providing 4TB of shared memory, and GPU nodes for exceptional performance and scalability for deep learning and accelerated computing.

With this computational power, the researchers used Bridges-2 to add a time dimension to seismic measurements and analyze how oil suppresses the loudness of sound traveling through it.

The team’s initial analysis found hidden rock structures in oil reserves to be responsible for preventing the extraction of all the oil within them. To tackle practical, large oil fields, the researchers are currently working on scaling up their system.

The researchers first reported¹ their results in the journal Geophysics last year and then again this year with more extensive² results.

Swipe to scroll →

Hidden Subsurface Structures: The Real Reason Wells Underperform

Oil isn’t just sitting in pools underground, ready to be pulled out. It actually gets soaked into porous sedimentary rock and then migrates through it toward the earth’s surface. But when oil gets trapped under less-porous cap rock, a reservoir is formed.

This is where sound comes in. Solid rock transmits sound at a higher speed than rock saturated with oil. By measuring how much oil slows down sound as it travels through rocks, experts can identify oil reserves.

These seismic methods create 3D images of where the oil-soaked rock is sitting, much like a medical ultrasound does of your muscles and internal organs.

Despite these capabilities, wells drilled based on those sophisticated images often produce less oil than expected. That’s because the 3D imaging isn’t capturing the full picture. Key information is missing.

The research team suspects that imaging the same reserves at different times would produce a 4D animation that creates a more accurate picture. Using more aspects of the seismic data in their analysis would also provide a better understanding of what’s happening.

Previously, oil reserves were detected based on the longer amount of time it took for sound to move through them. Now, the researchers added the amplitude of the signal, how oil damps out loudness, to the time data.

To do all the calculations at speed and temporarily store different components of the problem in its memory, the team used Bridges-2. 

The supercomputer was used in two phases. The first phase involved parallelizing the research code and making it more practical. The second phase involved implementing the code in the field data.

“PSC guaranteed me a hundred thousand computing hours, and also the memory to store my data, my field data … That just cannot be achieved with our local (resources).”

– Zhu

All this expanded analysis and repeated measurements paid off. The Penn State scientists found that when mapped out just by time, in a single measurement, the images didn’t capture the structures within the oil reserve. 

That is because some of these structures, like a layer of more solid rock within the reserve, wouldn’t affect the speed of the sound enough to be detected, but they would prevent the pumping out of the oil from below it.

In some cases, drilling a bit deeper can resolve this problem and access the rest of the oil in the well.

However, the researchers have applied their approach to a rather limited geological area, about 9 square miles. At this point, the work from Penn State scientists is only a proof of concept. The team is now focused on expanding their computations to more nodes, which will allow them to create accurate maps for much larger areas.

The team has another option to scale up their work, which they may explore, and that’s using Bridges-2’s extreme memory nodes, which have 4,000 gigabytes (GB) of RAM each.

From Depleted Wells to Energy Storage: CAES & Geothermal

Oil has been a dominant energy source for over a century. However, the world is now shifting away from fossil fuels to renewable energy sources like wind and solar, driven by concerns over climate change, air and water pollution, and habitat destruction caused by non-renewable resources.

Renewable energy sources are intermittent in nature and require better ways to store energy for later use. Interestingly, depleted oil and gas wells may provide a solution for this challenge.

These wells are actually a significant source of natural geothermal heat, and a study³ by Penn State researchers from earlier this year found that that heat can be utilized to boost the efficiency of compressed-air energy storage (CAES) by 9.5%, enabling more of the stored energy to be recovered and turned into electricity.

“This improvement in efficiency can be a game changer to justify the economics of compressed-air energy storage projects. And on top of that, we could significantly avoid the upfront cost by using existing oil and gas wells that are no longer in production. This could be a win, win situation.“

– Study co-author Arash Dahi Taleghani

Repurposing depleted oil and gas wells can also help reduce the negative impacts of orphaned wells. These are wells no longer maintained by their owners because they’re not economically viable.

Without oversight, these wellheads may leak toxic substances like methane, which has a warming impact 84 times greater than CO2 over a 20-year period. They also release substances like hydrogen sulfide, arsenic, and benzene that seep into local air, water, and soil systems, creating significant pollution problems.

There are at least 29 million abandoned wells internationally, according to a 2020 estimate by Reuters. 

A report⁴ from earlier this year, meanwhile, estimates the total number of abandoned oil and gas (AOG) wells to be 4,499,000, with 3,557,000 wells located in the USA. Moreover, they estimate that methane emissions from almost 4.5 million wells worldwide totaled ∼0.4 million tons (Mt) in 2022, which is equivalent to 10.5 Mt of CO2 over a 100-year timescale.

Not all orphaned wells are documented, though. In fact, many aren’t even listed in formal records and have no known operators.

To address this problem, researchers from the Department of Energy’s Lawrence Berkeley National Laboratory used modern tools⁵, including sensors, laser imaging, drones, and AI, to find these undocumented orphaned wells (UOWs).

According to lead study author Fabio Ciulla, a postdoctoral researcher at Berkeley Lab:

“While AI is a contemporary and rapidly evolving technology, it should not be exclusively associated with modern data sources. AI can enhance our understanding of the past by extracting information from historical data on a scale that was unattainable just a few years ago. The more we go into the future, the more you can also use the past.“

In their research, they scoured four counties of interest that had much early oil production and found about 1,300 potential UOWs. Twenty-nine have been verified using satellite images, while field surveys verified another 15.

This AI-driven mapping and verification work is part of a larger initiative to address undocumented orphaned wells. The Consortium Advancing Technology for Assessment of Lost Oil & Gas Wells (CATALOG) program is a collaboration to improve methods for finding wells, detecting and measuring methane, screening wells for their condition, prioritizing wells for plugging, and creating inexpensive tools for broad use.

“There’s a requirement now to quantify emissions before and after plugging an oil and gas well. Both because you want to make sure the plugging is done right, and you also want to quantify the impact of the program itself on our climate mitigation strategies – particularly for methane emissions, which can cause global warming impacts more quickly than carbon dioxide.“

– Scientist Sebastien Biraud, who leads the CATALOG project at Berkeley Lab

Investing in Smart Energy Exploration

When it comes to smart drilling, Baker Hughes (BKR +1.2%) is known for leading energy technology services. The company uses advanced sensors, cloud computing, digital twins, and AI for drilling optimization. It also provides methane detection and emission reduction technologies.

Baker Hughes (BKR +1.2%)

With a market cap of $47.8 billion, BKR shares are trading at $48.50, up 18.24% so far this year. The company is delivering an EPS (TTM) of 2.93 and a P/E (TTM) of 16.58. Baker Hughes pays a dividend yield of 1.90%.

As for its financial positioning, the company reported a revenue of $6.9 billion for Q2 2025. Attributable net income was $701 million. GAAP diluted EPS for the quarter came in at $0.71, and adjusted diluted EPS was $0.63.

“We delivered strong second-quarter results, with total adjusted EBITDA margins increasing 170 basis points year-over-year to 17.5% despite a modest decline in revenue. This performance reflects the benefits of structural cost improvements and continued deployment of our business system, which is driving higher productivity, stronger operating leverage, and more durable earnings across the company.”

– CEO Lorenzo Simonelli

During this period, the company reported a record backlog of $3.5 bln for its Industrial & Energy Technology (IET) sector that provides services for power-generation applications across the energy industry.

It also generated $510 million in cash flows from operating activities, while free cash flow was $239 million. Baker Hughes returned $423 million to its shareholders in 2Q25, including $196 million through share repurchases.

Latest Baker Hughes (BKR) Stock News and Developments

Conclusion

Crude oil remains one of the world’s most important energy sources, as it forms a significant portion of the global economy. However, it’s a finite resource, which could pose challenges for the future.

And as the easy-to-reach oil runs out, companies are drilling deeper than ever before, and only highly advanced technology can break this deadlock. This requires high-performance computing, advanced seismic analytics, sensors, data science, and AI. These tools are changing how we find oil, extract it, and even repurpose old wells.

So, the goal isn’t just to extract more oil from the ground; it’s also to do so with less environmental damage. And in some cases, these same technologies can help transition depleted wells into clean energy storage solutions.

Click here to learn if algal biofuel is the next energy revolution.

References

<sup>1. Xing, G., & Zhu, T. (2024). Advancing attenuation estimation through integration of the Hessian in multiparameter viscoacoustic full-waveform inversion. Geophysics, 89(5), r429. Published 1 September 2024. https://doi.org/10.1190/geo2023-0634.1
2. Kim, D., & Zhu, T. (2025). Why do seismic attenuation models enhance time-lapse imaging? A 2D viscoacoustic full-waveform inversion case study from the Volve field. Geophysics, 90(4), b193. Published 1 July 2025. https://doi.org/10.1190/geo2024-0793.1
3. Zhang, Q., Taleghani, A. D., & Elsworth, D. (2025). Underground energy storage using abandoned oil & gas wells assisted by geothermal. Journal of Energy Storage, 60, 115317. Published 8 January 2025. https://doi.org/10.1016/j.est.2025.115317
4. Lei, T., Chen, X., Ma, S., Jing, L., & Guan, D. (2025). A global inventory of methane emissions from abandoned oil and gas wells and possible mitigation pathways. National Science Review, 12(7), nwaf184. Published July 2025. https://doi.org/10.1093/nsr/nwaf184
5. Ciulla, F., Santos, A., Jordan, P., Kneafsey, T., Biraud, S. C., & Varadharajan, C. (2024). A Deep Learning Based Framework to Identify Undocumented Orphaned Oil and Gas Wells from Historical Maps: A Case Study for California and Oklahoma. Environmental Science & Technology, 58(50). Published December 2024. https://doi.org/10.1021/acs.est.4c04413</sup>