Verified
BHP Group
Experimenting- Mining
- 40,863 employees (FY2026)
- ~200 AI point solutions
- Azure and AWS
- Tech EBITDA run rate ~US$500M (Jun 2026)
BHP Group deploys AI inside its mines, plants and ports: machine learning recommendations in Escondida's copper concentrators on Microsoft Azure, computer vision that stops oversized rocks before Pilbara crushers, and AI-led maintenance, run by a Technical function under Chief Technical Officer Johan van Jaarsveld. BHP reports returns for technology as a whole, almost US$500 million of annual EBITDA run rate by Jun 2026, without an AI share, and has published no count of employees using AI tools.
bhp.comTheir approach to AI
- The platform
- BHP Group builds its own operational AI with its assets on Microsoft Azure Machine Learning and AWS, opened an Industry AI Hub in Singapore in May 2025, and is building a Central Data Platform for exploration data (Mar 2026).
- The models
- BHP Group has not named a large language model provider; its disclosed models are machine learning and computer vision models it trains for plant control, ore detection and maintenance, plus generative AI tools for safety questions and old exploration records.
- The reach
- BHP Group has not said how many of its 40,863 employees use AI tools; its published AI work sits in operations, from Escondida's concentrators to Western Australia Iron Ore's conveyors.
- The catch
- BHP Group credits technology initiatives with about US$150 million of EBITDA in FY2026 and an exit run rate of almost US$500 million, but folds AI together with autonomous trucks, leaching and rail technology and gives AI no separate figure.
Snapshot
- Employees
- 40,863
- FY2026, plus an average of 43,446 contractors
- Technology budget
- Not disclosed
- BHP Group reports technology returns, not technology spending
- AI/ML headcount
- Not disclosed
- A dedicated AI team (Jan 2026) and an Industry AI Hub of AI specialists in Singapore (May 2025)
- Who owns AI
- Johan van Jaarsveld, CTO
- Chief Technical Officer since Mar 2024, over Technology, Digital and Exploration; Digital Officer Mikko Tepponen runs the digital strategy
- Platform
- Azure ML and AWS, built in-house
- Azure Machine Learning operations platform by May 2023; Central Data Platform for exploration data in development (Mar 2026)
- Models
- In-house ML and computer vision
- Generative AI in a safety chatbot and in exploration data extraction; no LLM provider named
- Reach
- Not published
- No count of employees using AI tools
- Use cases
- ~200 AI point solutions
- Johan van Jaarsveld, Jan 2026: plants, blasting, conveyors, rail, maintenance and safety
Key findings
BHP Group said technology initiatives delivered about US$150 million of EBITDA in FY2026 and were running at almost US$500 million a year by Jun 30, 2026, aiming for more than US$650 million by the end of FY2027, but the figure combines AI with autonomous haulage, leaching and rail technology and gives AI no separate share.
BHP Group's computer vision system for oversized rocks and foreign objects at Western Australia Iron Ore went live in 2025 after about 1,000 hours of crusher downtime over 3 years, and Johan van Jaarsveld told The Age in Jan 2026 that it adds about $50 million a year for a couple of million dollars of outlay, while BHP's own May 2026 case study presents 0.8 million tonnes and US$46 million a year as targets at scale.
BHP Group has used machine learning recommendations from Microsoft Azure in Escondida's copper concentrators since 2023, and said in May 2025 that AI-powered plant control there had saved 3 billion litres of water and 118 GWh of energy since FY2022.
Chief Technical Officer Johan van Jaarsveld told Semafor in Jan 2026 that BHP Group had built a dedicated AI team and competes more with technology companies than with miners for AI talent, after BHP opened an Industry AI Hub in Singapore in May 2025.
Brandon Craig, BHP Group's CEO since Jul 1, 2026, told the Oct 23, 2026 annual meeting that BHP will combine its operating system with technology more deeply and build talent in digital and AI, while the FY2026 annual report rated delivery of the digital strategy above target on the CEO's scorecard.
Where BHP sits
- Stage
- Experimenting
BHP Group has reported no business return it credits to AI alone since May 2025. The figures on record bundle AI with other technology: about US$150 million of EBITDA from technology initiatives in FY2026 and an almost US$500 million exit run rate (Aug and Oct 2026) also cover autonomous haulage, leaching and rail technology, and more than US$2 billion over 4 financial years (Mar 2026) is credited to digital and analytics. The AI-only figures are not realized business returns in BHP Group's own written disclosures: Johan van Jaarsveld told The Age in Jan 2026 that oversize detection at Western Australia Iron Ore adds about $50 million a year, but BHP called that system a pilot in Mar 2026 and its roughly 1 million tonnes and $50 million uplift potential and a target at scale in May 2026; Escondida's 3% mill throughput gain from AI blast design (Jan 2026) is a throughput measure; and Escondida's AI plant control savings since FY2022 are in water and energy units (May 2025). BHP Group has not reported a company-wide AI rollout with a user count either, only about 200 AI point solutions (Jan 2026), so it sits at Experimenting; a realized AI-credited return in currency, such as a confirmed annual figure for oversize detection, or a published count of employees using AI under a named owner such as Johan van Jaarsveld, would move it to Piloting.
How the stages are defined - Ownership
- Chief Technical Officer Johan van Jaarsveld, on the Executive Leadership Team
- Platform
- Hybrid: AI built in-house with assets on Azure and AWS
- Models
- In-house machine learning and computer vision; no LLM provider named
- Hosting
- Microsoft Azure (Azure Machine Learning at Escondida) and AWS for analytics and ML; detection models wired into site process control
- Reach
- No employee usage count published
How they deploy
AI in the operation first
BHP Group aims AI at its mines, processing plants, rail and ports before office work: Johan van Jaarsveld told Semafor in Jan 2026 that the AI team picks the domains with the biggest, quickest wins, such as maintenance and blasting decisions, and is not yet focused on AI-driven exploration because new mines take so long to develop. BHP's FY2026 results list AI resource characterisation, crushing and screening control and AI-led predictive maintenance among the technology levers behind its EBITDA target.
Digital strategy tied to unit costs
BHP Group put a digital strategy in place at the end of FY2025 to integrate data, AI and automation into daily decisions and lower unit costs, build and scale solutions with frontline teams, and make digital tools easier for everyone at BHP to use, Mikko Tepponen wrote in Oct 2026. Brandon Craig's first annual letter, in Aug 2026, made pairing the BHP Operating System with faster adoption of technology the first of 3 strategic pillars, and the 2026 Annual Report names AI as the lead example.
From single-point tools to pit-to-port systems
BHP Group says most of its technology gains so far come from single-point applications, and Mikko Tepponen wrote in Oct 2026 that the next step is integrating the supply chain from pit to port so the whole system can be optimized together. Tepponen told a Perth conference in May 2026 that a gap remains between the AI the industry builds and the AI that scales, and that BHP's oversize detection system worked because it was co-designed with frontline staff and designed to scale from the outset.
Data foundations before models
BHP Group treats data standards and interoperability as the condition for AI value, Mikko Tepponen said in a Mar 2026 keynote, and is building a Central Data Platform to standardize exploration data from more than 140 years of operations. BHP signed data collaborations with South Africa's Council for Geoscience in Feb 2026 and the Government of Canada in Mar 2026 to digitize legacy geoscience records.
Partners, startups and hiring
BHP Group works with Microsoft and AWS on cloud and machine learning, with AI Singapore on its Industry AI Hub, and through BHP Ventures with startups such as GeologicAI, which raised a US$44 million Series B with BHP Ventures in Jul 2025. Johan van Jaarsveld said in Jan 2026 that BHP has found AI programmers who understand geology hard to hire and pays to keep them.
- Deployment model
- BHP Group runs AI through its Technical function under Chief Technical Officer Johan van Jaarsveld, which covers Technology, Digital, Minerals Exploration, Value Engineering and the centres of excellence for projects, maintenance, and engineering and resources (2026 Annual Report). Digital Officer Mikko Tepponen leads the digital strategy, and BHP builds each solution with the asset that uses it, such as Western Australia Iron Ore's oversize detection and Escondida's concentrator recommendations, before scaling it to other sites. BHP's Industry AI Hub in Singapore, opened in May 2025 with AI Singapore and Enterprise Singapore, works on enterprise-wide problems with local AI partners, and BHP was recruiting AI excellence and data excellence practice leads in May 2025 to set engineering standards for models in production.
- Model providers
- BHP Group has not named a large language model provider. BHP's disclosed AI consists mostly of machine learning and computer vision models it builds with its own data: hourly predictions from Escondida concentrator data on Azure Machine Learning (2023), cameras and machine learning models that flag oversized rocks at Western Australia Iron Ore, and AI-led maintenance models. BHP's FY2026 annual report describes an 'Ask Safety' generative AI chatbot and a voice-to-text hazard reporting app, and Mikko Tepponen said in Mar 2026 that BHP uses generative AI to turn unstructured exploration records into datasets. In Mar 2024 BHP's people analytics team described testing ChatGPT in a sandbox environment to review a leadership framework.
- Data and infrastructure
- BHP Group signed long-term cloud agreements with Microsoft and Amazon Web Services in Jun 2021: BHP's global applications portfolio, including SAP, was to move to Microsoft Azure over 3 years, and AWS was to support data analytics and machine learning, with an AWS Cloud Academy to train staff. BHP's Escondida concentrator work runs on Azure Machine Learning, Azure Synapse Analytics and Azure Data Lake Storage (Microsoft, May 2023). The oversize detection models at Western Australia Iron Ore are integrated into the process control system and stop conveyors automatically (Oct 2026). BHP's FY2026 annual report lists cloud service providers among its cybersecurity exposures and says the Chief Technical Officer holds primary responsibility for technology and innovation risks.
Platforms and internal tools
Azure Machine Learning operations platformBy May 2023
BHP Group's Azure Machine Learning operations platform makes hourly predictions from real-time Escondida concentrator data and turns them into recommendations for operators who adjust flotation and other processing variables; Microsoft said in May 2023 that BHP was extending it to a second Escondida concentrator.
- Models
- In-house machine learning models
AWS analytics and machine learningJun 2021
BHP Group chose AWS in Jun 2021 for its data analytics and machine learning work, citing an AWS solution that uses AI to optimize commodity logistics schedules from mine to market.
Oversize and Foreign Object Detection (OFOD)2025
OFOD is BHP Group's computer vision and machine learning system on diggers and conveyors at Western Australia Iron Ore that detects rocks larger than 200mm and foreign objects and stops the belt automatically so crews can remove them.
- Models
- In-house computer vision models
Central Data PlatformIn development (Mar 2026)
BHP Group's Central Data Platform standardizes and integrates exploration data globally, and BHP pairs it with generative AI tools that convert scanned reports and other unstructured records into usable datasets.
Ask Safety and AI hazard reportingFY2026
Ask Safety is BHP Group's generative AI chatbot that connects and aggregates field leadership data, released in FY2026 alongside a voice-to-text hazard reporting app in Spanish and English.
Vendors and partners
MicrosoftJun 2021
Azure cloud for BHP's applications portfolio; Azure Machine Learning behind Escondida's concentrator recommendations · BHP Group and Microsoft announced the Escondida copper recovery collaboration on May 31, 2023.
Amazon Web ServicesJun 2021
Cloud for data analytics and machine learning
AI Singapore and Enterprise SingaporeMay 2025
Partners in BHP's Industry AI Hub in Singapore
BOTONJun 2026
Conveyor supplier developing AI-based automatic belt alignment and robotic inspection for BHP · The Global Framework Agreement builds on conveyor work at Escondida and commits only to investigating some initiatives, with no timeline.
GeologicAIJul 2025
BHP Ventures portfolio company scanning drill core with sensors and machine learning
Council for Geoscience (South Africa) and Government of CanadaFeb 2026
Collaborations to digitize legacy geoscience and drill core data
Leadership
AI, data and platform leaders
Executives
Board
AI org map
Central Technical function sets the digital strategy; solutions built with each operating asset
BHP Group places AI under Chief Technical Officer Johan van Jaarsveld, a member of the Executive Leadership Team under CEO Brandon Craig, who runs Technology, Digital, Minerals Exploration, Value Engineering and the centres of excellence. Digital Officer Mikko Tepponen leads the digital strategy BHP adopted at the end of FY2025, and the solutions are co-designed with operating assets such as Western Australia Iron Ore and Escondida. BHP's board oversees technology and cybersecurity risk, including AI, through the Risk and Audit Committee.
BHP Group has not published how many data scientists or AI engineers it employs. Johan van Jaarsveld told Semafor in Jan 2026 that BHP had recently built a dedicated AI team to make operations more efficient, BHP's Industry AI Hub in Singapore hires AI specialists with Enterprise Singapore's support (May 2025), and AI and data practice leads were being recruited for Brisbane, Adelaide, Perth or Santiago in May 2025. Much of BHP's AI work is built with asset teams: BHP Group says it develops solutions with its operations so frontline teams own and improve them.
Reporting lines
Chief Executive Officer
Technical function
Johan van Jaarsveld, Chief Technical Officer
Technology, Digital, Minerals Exploration, Value Engineering and the centres of excellence for projects, maintenance, and engineering and resources
Digital
Mikko Tepponen, Digital Officer
Reporting line inferred from the Chief Technical Officer's remit; not published
Group Technology
Thomas Leen, Group Technology Officer
Title and line reported by Digital Nation in May 2025; not published by BHP
Technology Cybersecurity and Architecture
Vice President (not named)
Reports on cybersecurity risk to the Chief Technical Officer (2026 Annual Report)
Industry AI Hub, Singapore
Opened May 2025 with AI Singapore and Enterprise Singapore; reporting line not published
Commercial
Ragnar Udd, Chief Commercial Officer
Sales and marketing, procurement, maritime and Group Business Services; BHP has used AI in procurement
Operating assets
Western Australia Iron Ore, Escondida, Spence, Copper South Australia and BMA run AI solutions built with the Technical function
BHP Group's 2026 Annual Report names the Chief Technical Officer as the executive responsible for technology and innovation risks, with the Vice President Technology Cybersecurity and Architecture reporting to him. BHP has not published where its Digital Officer, group technology officer or Industry AI Hub sit, and the group technology officer title comes from trade press only.
Where the AI workforce sits
Central AI team and Industry AI Hub
Johan van Jaarsveld described a dedicated AI team in Jan 2026; BHP has given no headcount for it or for the Singapore hub.
Not disclosedAsset teams running AI in operations
Data scientists and operators at sites such as Western Australia Iron Ore and Escondida co-design and run the solutions; BHP gives no count.
Not disclosed
Named people
Use cases
BHP Group has named AI use cases in copper processing at Escondida, oversize detection at Western Australia Iron Ore, AI resource characterisation, predictive maintenance, safety reporting, compliance monitoring, procurement and exploration data, but has published no business return it credits to AI alone for any of them. BHP Group's firm-level figures cover technology as a whole: about US$150 million of EBITDA delivered in FY2026 and an exit run rate of almost US$500 million, with more than US$650 million targeted by the end of FY2027, alongside more than US$2 billion over 4 financial years credited to digital and analytics (Mar 2026). The AI-specific figures BHP gives are a mill throughput gain, water and energy savings and loss reductions in targeted applications, and the about $50 million a year Johan van Jaarsveld cited for oversize detection in Jan 2026 appears in BHP's later materials as potential and a target at scale.
| Use case | Area | Status | ROI |
|---|---|---|---|
| Oversize and foreign object detection Cameras on diggers and above conveyors feed computer vision models that flag rocks larger than 200mm and foreign objects, stopping the belt before they reach crushers and ports, after about 1,000 hours of crusher downtime over 3 years. Johan van Jaarsveld told The Age in Jan 2026 that avoided downtime equates to up to 1 million tonnes of extra ore and about $50 million a year, and Mikko Tepponen said in May 2026 that the system, live since 2025, delivers a little under 1 million tonnes a year of uplift potential and cut crusher downtime by 20%; BHP called it a pilot in Mar 2026, and its case study with the Business Council of Australia gives 0.8 million tonnes and US$46 million a year as targets at scale. | Iron ore (Western Australia) | In production | |
| Concentrator recommendations on Azure Machine Learning Machine learning models make hourly predictions from real-time concentrator data and recommend flotation and other settings to Escondida operators, a program BHP Group began with Microsoft and announced in May 2023. BHP's FY2023 annual report credited increased copper recovery at Escondida to AI recommendations, and BHP said in May 2025 that AI-powered plant control had saved 3 billion litres of water and 118 GWh of energy since FY2022. | Copper (Escondida) | In production | |
| AI blast design and ore characterisation Escondida teams apply AI to historic drill data to predict orebody density and design blast patterns that put the strongest blasts against the hardest rock, and use digital twins to adjust blending before ore reaches the semi-autogenous grinding mill. Johan van Jaarsveld said in Jan 2026 that mill throughput rose 3% the day the blast design started, and BHP Group said in Mar 2026 that some granulometry-linked production losses fell by around 70% in targeted applications. | Copper (Escondida) | In production | |
| Rail loading spillage detection BHP Group built a real-time computer vision model on its existing CCTV that detects ore spillage while trains are loaded and stops the loading and raises the chute in under half a second, Johan van Jaarsveld said in Jan 2026. | Iron ore (Western Australia) | In production | |
| AI resource characterisation BHP Group lists more definitive resource characterisation through AI among the levers behind its technology EBITDA in the FY2026 results, without a figure for the lever itself. | Resources and mine planning | In production | |
| AI-led predictive and preventive maintenance BHP Group runs AI-led predictive and preventive maintenance across its mining operations, which Johan van Jaarsveld named in Jan 2026 as one of the domains with the quickest wins and Mikko Tepponen credited with reliability improvements and cost savings in Oct 2026, without a figure. | Maintenance | In production | |
| Steel ball counting with artificial vision Spence replaced a photoelectric steel ball counting system in its concentrator with an artificial vision system that reduced steel ball consumption, one of the BHP Operating System cost improvements BHP Group listed in Aug 2026 at about US$2 million a year; BHP does not describe the system as AI. | Copper (Spence) | In production | |
| Ask Safety chatbot and AI hazard reporting BHP Group introduced the Ask Safety generative AI chatbot, which connects and aggregates field leadership data, and a voice-to-text hazard reporting app in Spanish and English in FY2026; the app geotags reports and gives quick risk assessments with warnings drawn from past incidents, Johan van Jaarsveld said in Jan 2026. | Safety | In production | |
| Generative AI extraction of legacy exploration records BHP Group uses generative AI tools to convert scanned reports and other unstructured geological records into datasets for its Central Data Platform, and said in Mar 2026 that one compilation that took months was done in hours. Mike Henry said in Mar 2025 that machine learning had found copper mineralisation at Oak Dam in South Australia that people had missed. | Exploration | Pilot | |
| Transaction monitoring BHP Group's Group Compliance team monitors transactions with data analytics and AI as part of its anti-corruption framework (2026 Annual Report). | Compliance | In production | |
| Procurement processes and financial forecasts BHP Group said in May 2025 that it had applied AI to make procurement processes more efficient and to improve financial forecasts, without naming the tools. | Procurement and finance | In production | |
| AI belt alignment and robotic inspection with BOTON BHP Group and conveyor supplier BOTON agreed in Jun 2026 to develop AI-based automatic belt alignment, robotic inspection, rip detection and X-ray scanning, building on work at Escondida, with a goal of rolling the technology out across BHP's operations. | Conveyors | Announced |
Adoption
BHP Group has not published how many of its 40,863 employees use AI tools, and its public record names no company-wide AI assistant. Johan van Jaarsveld said in Jan 2026 that BHP had rolled out about 200 AI point solutions across the business and is moving to platforms installed at every site. BHP's AI spread runs through operations: machine learning in Escondida's concentrators from 2023, computer vision on Western Australia Iron Ore's diggers and conveyors from 2025, AI-led maintenance across mining operations, and generative AI tools for safety questions and exploration records by FY2026. BHP's digital strategy, adopted at the end of FY2025, aims to make digital tools easier for everyone at BHP to use, and Mikko Tepponen wrote in Oct 2026 that BHP is now scaling single-point applications across its global operations.
By the numbers
Rollout, first pilot to today
Jun 2021
BHP Group signed long-term cloud agreements with Microsoft, to host its global applications on Azure, and with AWS, for data analytics and machine learning.
May 2023
BHP Group and Microsoft announced AI recommendations for Escondida's copper concentrators on Azure Machine Learning, extending to a second concentrator.
Sep 2023
BHP Group's FY2023 annual report credited increased copper recovery at Escondida to AI recommendations and new machine learning platforms.
Mar 2024
BHP Group's people analytics team described testing ChatGPT in a sandbox to review its leadership framework, one of several generative AI use cases in people management.
2025
BHP Group implemented oversize and foreign object detection on conveyors at Western Australia Iron Ore.
May 2025
BHP Group opened its first Industry AI Hub in Singapore with AI Singapore and Enterprise Singapore, and began recruiting AI excellence and data excellence practice leads.
Jun 2025
BHP Group put its digital strategy in place at the end of FY2025 to integrate data, AI and automation into daily decisions and lower unit costs.
Jan 2026
Johan van Jaarsveld said BHP Group had rolled out about 200 AI point solutions and was moving to a platform-wide approach installed across sites.
Jan 2026
Johan van Jaarsveld said BHP Group had built a dedicated AI team focused on maintenance and blasting decisions.
Aug 2026
BHP Group reported its Ask Safety generative AI chatbot and AI hazard reporting app, and put technology EBITDA at about US$150 million for FY2026 with an exit run rate of almost US$500 million.
Oct 2026
Mikko Tepponen wrote that BHP Group is scaling single-point applications across its global operations and moving to integrated pit-to-port systems.
Economics & controls
Spend and ROI
Technology budget
BHP Group has not published a technology or AI budget in its FY2026 annual report or results presentation. BHP's annual report says BHP Ventures' FY2026 investments included technologies advancing robotics and AI, without amounts.
Reported ROI
BHP Group reports returns for technology as a whole, not for AI alone. Mikko Tepponen wrote on Oct 29, 2026 that technology initiatives delivered about US$150 million of EBITDA in FY2026 and were running at almost US$500 million a year by Jun 30, 2026, with more than US$650 million targeted by the end of FY2027; the FY2026 results slide behind those figures lists AI resource characterisation and AI-led maintenance next to autonomous haulage, leaching and rail technology. Tepponen also said in Mar 2026 that digital and analytics initiatives delivered more than US$2 billion over the past 4 financial years, without a split by AI or by year; Johan van Jaarsveld gave the same roughly $2 billion since 2021 to The Age in Jan 2026 when describing BHP's AI point solutions, and said the next, platform-wide phase targets about $250 million a year in savings or revenue.
AI-specific figures
BHP Group's AI-specific figures measure physical or potential gains, not realized business returns. BHP said in May 2025 that AI-powered plant control at Escondida had saved 3 billion litres of water and 118 GWh of energy since FY2022; Mikko Tepponen said in Mar 2026 that AI models cut some granulometry-linked production losses at Escondida by around 70% in targeted applications; Johan van Jaarsveld said in Jan 2026 that AI blast design lifted Escondida mill throughput 3% and that oversize detection at Western Australia Iron Ore adds about $50 million a year for a couple of million dollars of outlay, while Tepponen called the oversize figure uplift potential in May 2026 and a BHP case study presents it as a target at scale.
Incentives
BHP Group's FY2026 CEO scorecard included delivering the digital strategy targets, which the board assessed as above target, within a performance category rated at maximum (2026 Annual Report).
Headcount
BHP Group had 40,863 employees and an average of 43,446 contractors in FY2026. BHP's 2025 annual report warned that data science, AI and robotics may replace some tasks and change the workforce, and BHP has published no AI-related job reductions.
Governance and risk
Executive accountability
BHP Group's CEO delegates primary responsibility for technology and innovation risks, including cybersecurity, to Chief Technical Officer Johan van Jaarsveld, with the Vice President Technology Cybersecurity and Architecture reporting to him; the board oversees those risks through the Risk and Audit Committee (2026 Annual Report).
Disclosed AI risks
BHP Group's FY2026 annual report lists the risk of failing to adopt and scale AI in time, AI systems deployed without appropriate governance or human oversight, and the responsible, secure and ethical deployment of AI, including workforce disruption. BHP also lists cyber risks from AI-enabled impersonation and deepfakes and from compromised or autonomous AI systems, including frontier large language models and AI agents.
Data and responsible use
BHP Group recruited its AI excellence practice lead in May 2025 to set technical standards that include ensuring business use of AI is fair, unbiased and ethically responsible, and its data excellence lead to cover governance and security of data systems. BHP has not published an AI policy or a list of approved models.
Safety controls
BHP Group uses AI mainly to move safety from human-dependent controls toward engineered ones: the oversize detection system stops conveyors automatically so crews intervene less, and the FY2026 annual report ties AI hazard reporting and the Ask Safety chatbot to early identification of risks. Brandon Craig told the Oct 2026 annual meeting that BHP will move faster to adopt technologies that make work safer.
Recent signals
Oct 29, 2026
Mikko Tepponen wrote that BHP Group's technology initiatives delivered about US$150 million of EBITDA in FY2026, reached an almost US$500 million run rate and target more than US$650 million by the end of FY2027.
Oct 23, 2026
Brandon Craig told BHP Group's annual meeting that BHP will combine its operating system with technology more deeply and build talent in project delivery, digital and AI.
Sep 3, 2026
Albemarle named BHP Group Chief Commercial Officer Ragnar Udd its CEO, effective Feb 1, 2027.
Aug 18, 2026
BHP Group's FY2026 results put the EBITDA run rate from technology at almost US$500 million and named AI resource characterisation and AI-led maintenance among the levers, with no AI share.
Aug 2026
BHP Group's 2026 Annual Report described the Ask Safety generative AI chatbot and a voice-to-text hazard reporting app, and rated digital strategy delivery above target.
Jul 1, 2026
Brandon Craig succeeded Mike Henry as BHP Group's Chief Executive Officer.
Jun 12, 2026
BHP Group and conveyor supplier BOTON signed a Global Framework Agreement to develop AI-based belt alignment and robotic inspection.
May 2026
Mikko Tepponen told the Global Resources Innovation Expo in Perth that BHP Group's oversize detection system offers a little under 1 million tonnes a year of uplift potential, worth almost US$50 million.
Mar 25, 2026
Mikko Tepponen said BHP Group's digital and analytics initiatives delivered more than US$2 billion over 4 financial years and that BHP is building a Central Data Platform for exploration data.
Mar 18, 2026
BHP Group's board named President Americas Brandon Craig as the next CEO, effective Jul 1, 2026.
Jan 22, 2026
Johan van Jaarsveld told Semafor that BHP Group had built a dedicated AI team and competes with technology companies for AI talent.
Jan 5, 2026
Johan van Jaarsveld told The Age that BHP Group has rolled out about 200 AI point solutions, that AI blast design lifted Escondida mill throughput 3%, and that a platform-wide approach targets about $250 million a year.
Quick answers
- How does BHP Group deploy AI?
- BHP Group deploys AI inside its mines, plants and ports: machine learning recommendations in Escondida's copper concentrators on Microsoft Azure, computer vision that stops oversized rocks before Pilbara crushers, and AI-led maintenance, run by a Technical function under Chief Technical Officer Johan van Jaarsveld. BHP reports returns for technology as a whole, almost US$500 million of annual EBITDA run rate by Jun 2026, without an AI share, and has published no count of employees using AI tools.
- What does BHP Group use AI for?
- BHP Group has named AI use cases in copper processing at Escondida, oversize detection at Western Australia Iron Ore, AI resource characterisation, predictive maintenance, safety reporting, compliance monitoring, procurement and exploration data, but has published no business return it credits to AI alone for any of them. BHP Group's firm-level figures cover technology as a whole: about US$150 million of EBITDA delivered in FY2026 and an exit run rate of almost US$500 million, with more than US$650 million targeted by the end of FY2027, alongside more than US$2 billion over 4 financial years credited to digital and analytics (Mar 2026). The AI-specific figures BHP gives are a mill throughput gain, water and energy savings and loss reductions in targeted applications, and the about $50 million a year Johan van Jaarsveld cited for oversize detection in Jan 2026 appears in BHP's later materials as potential and a target at scale.
- Who leads AI at BHP Group?
- AI at BHP Group is led by Johan van Jaarsveld (Chief Technical Officer); Mikko Tepponen (Digital Officer); Thomas Leen (Group Technology Officer). Ownership of AI on the org chart: Chief Technical Officer Johan van Jaarsveld, on the Executive Leadership Team.
- What AI platform does BHP Group use?
- BHP Group builds its own operational AI with its assets on Microsoft Azure Machine Learning and AWS, opened an Industry AI Hub in Singapore in May 2025, and is building a Central Data Platform for exploration data (Mar 2026).
- Which AI models does BHP Group use?
- BHP Group has not named a large language model provider; its disclosed models are machine learning and computer vision models it trains for plant control, ore detection and maintenance, plus generative AI tools for safety questions and old exploration records.
- How many BHP employees use AI?
- BHP Group has not said how many of its 40,863 employees use AI tools; its published AI work sits in operations, from Escondida's concentrators to Western Australia Iron Ore's conveyors.
- What return has BHP Group reported from AI?
- BHP Group has reported no business return it credits to AI alone since May 2025. The figures on record bundle AI with other technology: about US$150 million of EBITDA from technology initiatives in FY2026 and an almost US$500 million exit run rate (Aug and Oct 2026) also cover autonomous haulage, leaching and rail technology, and more than US$2 billion over 4 financial years (Mar 2026) is credited to digital and analytics. The AI-only figures are not realized business returns in BHP Group's own written disclosures: Johan van Jaarsveld told The Age in Jan 2026 that oversize detection at Western Australia Iron Ore adds about $50 million a year, but BHP called that system a pilot in Mar 2026 and its roughly 1 million tonnes and $50 million uplift potential and a target at scale in May 2026; Escondida's 3% mill throughput gain from AI blast design (Jan 2026) is a throughput measure; and Escondida's AI plant control savings since FY2022 are in water and energy units (May 2025). BHP Group has not reported a company-wide AI rollout with a user count either, only about 200 AI point solutions (Jan 2026), so it sits at Experimenting; a realized AI-credited return in currency, such as a confirmed annual figure for oversize detection, or a published count of employees using AI under a named owner such as Johan van Jaarsveld, would move it to Piloting.
Open questions
- How much of BHP Group's technology EBITDA, about US$150 million in FY2026 and an almost US$500 million run rate, comes from AI as opposed to autonomous haulage, leaching and rail technology?
- Has BHP Group's oversize detection system at Western Australia Iron Ore delivered the roughly 1 million tonnes and $50 million a year that Johan van Jaarsveld described in Jan 2026, given that BHP later called the figures potential and targets at scale, and in which currency?
- What do Escondida's AI concentrator recommendations and AI-led maintenance return in copper tonnes or dollars, given that BHP reports water, energy and loss reductions only?
- How many BHP Group employees use generative AI tools such as Ask Safety, and which large language models and cloud services run them?
- How many people work in BHP Group's dedicated AI team and Industry AI Hub in Singapore, and to whom do the hub and the Digital Officer report?
- How much does BHP Group spend on technology and on AI each year?
- Who will speak for AI on BHP Group's Executive Leadership Team after Ragnar Udd, who spoke at the AI hub launch, leaves for Albemarle in Feb 2027?
Sources
BHP Group
- How digital and technology is transforming how we work, Mikko Tepponen (Oct 29, 2026)
- From discovery to decisions: data, AI and the future of mining (Mar 25, 2026)
- Brandon Craig at the 2026 BHP AGM (Oct 23, 2026)
- BHP launches its first Industry AI Hub in Singapore (May 27, 2025)
- BHP and Microsoft use AI to lift Escondida copper recovery (May 31, 2023)
- BHP: Oversize and foreign object detection with AI, Business Council of Australia (2026)
- PDAC 2026 presenter profile: Mikko Tepponen, Digital Officer
Filings and investor materials
- BHP Group 2026 Annual Report, Form 20-F (Aug 2026)
- BHP Group 2025 Annual Report, Form 20-F (Aug 2025)
- Results for the full year ended Jun 30, 2026, presentation (Aug 18, 2026)
- Results for the full year ended Jun 30, 2026, release (Aug 18, 2026)
- Form 6-K, FY2026 results and annual report extracts (Aug 2026)
- Albemarle announces CEO succession plan, Ragnar Udd (Sep 3, 2026)
Partners
News
- iTnews: BHP eyes momentum as it scales digital systems across supply chain (Oct 30, 2026)
- SteelOrbis: BHP says AI solution boosts iron ore output in Western Australia (May 2026)
- Reuters brief via Tiger Brokers: BHP digital officer on data standards and US$2 billion of value (Mar 25, 2026)
- Semafor: Mining giant BHP seeks out AI talent (Jan 22, 2026)
- The Age: AI digs out more profits at mining giant BHP, Johan van Jaarsveld interview (Jan 5, 2026)
- iTnews: BHP sets sights on enterprise-wide AI transformation (May 22, 2025)
- Caixin: BHP chief maps out post-iron ore future, Mike Henry interview (May 2025)
- iTnews: BHP uses ChatGPT to analyse a leadership framework (Mar 4, 2024)
- Bloomberg via Mining.com: AI helps BHP fight declining ore grades at biggest copper mine (Sep 2023)
- iTnews: BHP to shift global applications portfolio to Microsoft Azure (Jun 18, 2021)
Data confidence
- Confirmed
- Stated in a primary or reputable public source.
- Estimated
- Derived or approximated from public reporting.
- Unconfirmed
- Reported, but not yet corroborated.
BHP Group profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 5, 2026
