The tracker was discovered while sorting mail on board, though only after Omroep Gelderland had been tracking the Evertsen for 24 hours, the broadcaster wrote on its website on Thursday. The Dutch Ministry of Defence said it’s taking measures in response, according to the broadcaster.

The tracker incident comes after a Le Monde reported in March it was able to locate a French officer taking a 7-kilometer run around the deck of the Charles de Gaulle while the carrier was at sea, through data from the officer’s connected watch via the running and cycling app Strava.

“You do want to be able to intercept such a tracker,” Rowin Jansen, assistant professor of national security law at Radboud University in Nijmegen, told Omroep Gelderland. “Commercial satellite images are currently released with a delay for good reason. You certainly don’t want to make it easy for terrorists to send a similar package and track a ship’s location in real time. You then run the risk of having missiles fired at you.”

The broadcaster described sending the Bluetooth tracker, a gadget used for example to find keys, to the frigate in an envelope using the military postal service, following online instructions from the MoD on how to send mail to military personnel.

While the ministry checks whether prohibited or dangerous items are sent by mail by X-ray scanning packages, Omroep Gelderland noted that online videos showed envelops not being scanned, so decided to pack the tracker in a postcard, with the gadget going undetected and simply mailed.

“In a large-scale conflict, everyone needs to ask themselves: What can I contribute to the safety of our men and women?” said retired Lt. Gen. Mart de Kruif, as cited by the broadcaster. “So you should no longer rely on existing rules, but on what is necessary. We’re still a bit naive, and that mindset needs to change.”

Omroep Gelderland mapped the route of the tracker from the Dutch naval base in Den Helder to Eindhoven Airport, and then on to the port of Heraklion in Crete, where webcam images showed the Evertsen moored at the quay.

With the frigate departing the port on March 27, the broadcaster said it was able to track the vessel sailing west along the coast of Crete before setting an eastward course. The tracker then went permanently offline 24 hours later near Cyprus, it said.

The MoD has made adjustments in response to the reported incident, including no longer allowing greeting cards with batteries to be sent to the Evertsen, and the ministry will further review the guidelines for military mail, a spokesperson told Omroep Gelderland.

The tracker was found during mail sorting aboard the frigate after it set sail, and while the vessel could be tracked at sea, this would not have posed an operational risk, according to the MoD.

Defence Minister Dilan Yeşilgöz informed parliament about the incident on Thursday evening, Omroep Gelderland said.

The revelation of undeclared remote access devices in American solar farms, first reported by Reuters earlier this month, came less than a month after a power outage that shut down electricity for millions throughout Spain, Portugal, Andorra and parts of France, highlighting the possible fragility of even highly developed and integrated European power grids.

According to unnamed sources cited by Reuters, the communication devices that were embedded in solar farm gear were not shown on schematics and customer information of the products, suggesting they may have been deliberately concealed. The undisclosed devices were reportedly found during a routine disassembly of Chinese-made power inverters, which serve to connect solar farms to the electricity grid, control the flow of power and maintain the all-important grid frequency.

While Iberian authorities have ruled out a cyber attack in the case of the peninsula’s massive blackout, the finding has nonetheless instilled a new sense of urgency in European planning to make the continent’s integrated electrical grid safe and resilient.

Market dominance

Inverters are crucial in linking photovoltaic (PV) power plants, which output DC electricity, to the broader electricity network, which runs on AC. In 2023, 78% of all inverters installed in Europe came from Chinese vendors, with the overwhelming majority being made by Huawei and SunGrow, according to DNV, a risk consultancy. The report was commissioned by SolarPower Europe, an industry advocacy group.

This market dominance can likely be explained by a combination of China’s large manufacturing capacity and the comparatively lower prices of Chinese inverters compared to European ones.

Control over the inverters allows outsiders to simultaneously disconnect generating capacity from the grid, which can cause blackouts. It would also allow them to manipulate voltage and frequency settings to destabilize local grids and to override safety protections like anti-islanding systems.

Both Huawei and Sungrow have documented links to the Chinese government and the country’s ruling Communist Party, including formal ties, participation in government projects, and officials holding high-ranking positions simultaneously in both the companies and the state.

Under Chinese law, Huawei faces mandatory cooperation requirements with intelligence services. The 2017 National Intelligence Law declares that Chinese companies must “support, assist, and cooperate with” China’s intelligence-gathering authorities. As a result of questions over its independence and safety, the electronics giant has already faced restrictions on work on critical communications infrastructure – especially 5G networks – in several countries. It is also front and center in a major investigation currently ongoing in Brussels surrounding bribery of European officials.

Energy sovereignty at risk

“Europe’s energy sovereignty is at serious risk due to the unregulated and remote control capabilities of photovoltaic inverters from high-risk, non-European manufacturers – most notably from China,” said the European Solar Manufacturing Council, an industry association.

This isn’t purely hypothetical, either. In November 2024, some solar inverters in the U.S., U.K. and Pakistan were actually disabled remotely from China. Very little was publicly revealed about this incident and its consequences; investigations later showed that the shutdown may have been the result of an industry dispute, according to Günter Born, a German tech and cybersecurity journalist.

Following the decoupling of Europe’s energy needs from cheap and readily available Russian gas after the invasion of Ukraine, energy sovereignty has become a new priority for capitals across the continent. Renewables have been front and center, building on existing momentum to combat climate change and residual skepticism of nuclear power for its perceived safety shortcomings and high cost.

“Today, over 200 GW of European PV capacity is already linked to inverters manufactured in China – the equivalent of more than 200 nuclear power plants,” said Christoph Podewils, the ESMC secretary general, citing numbers by the DNV consultancy.

“This means Europe has effectively surrendered remote control of a vast portion of its electricity infrastructure.”

The Iberian Peninsula blackout was triggered by a 2.2-gigawatt (GW) loss in electricity generation that occurred within seconds. Previously, DNV had estimated that a loss of 3 GW of generating power could have serious ramifications for the European power grid. A loss of this amount of input can result in cascading effects, leading to a large-scale shutdown of the power grid, just as had occurred in Spain on April 28.

The DNV report identified over a dozen threat scenarios, most of which it considered “high” or “critical” risks even after the full implementation of existing EU cybersecurity measures.

In the first three months of 2025 alone, nearly 68 terawatt hours of electricity were produced by solar panels, data from the energy think tank Ember shows. This marks an increase by almost a third over the same period a year earlier. It amounted to 8.2% of Europe’s electricity production in March, despite the shorter days at that time of year. In summer, around 15% of Europe’s electricity may come from the sun, with some countries – particularly around the Mediterranean – relying on it to make up over a quarter of their total energy mix during the daytime.

According to SolarPower Europe, the industry advocacy group, the installed PV capacity in Europe is expected to exceed 800 GW by 2030.

Wrangling the dragon

The shock of the Spanish power outage – even without it being caused by an attack – has jolted some in Europe into action. But consciousness surrounding a China vulnerability predates the grid collapse.

Lithuania on May 1 implemented new legislation that requires photovoltaic projects above 100kW to use inverters that meet national safety standards. Existing projects must also be retrofitted. This practically outlaws Chinese inverters in the country’s power infrastructure.

Late last year, Estonia’s spy chief Kaupo Rosin said that Chinese technology in critical infrastructure, particularly solar farms, could open Europe up to Chinese blackmail.

Some EU lawmakers are also taking note, with one member of the European parliament submitting an inquiry on the topic of solar inverters and the risk of Chinese influence on May 15.

A 2022 EU directive called NIS2 provides the basis for joint cybersecurity defense measures of critical infrastructure, outlining 18 broad sectors that require particular attention, including electrical grids.

However, the directive primarily applies to very large projects, leaving smaller power generation projects vulnerable. This is particularly relevant to solar, which is not only produced in large farms but also by countless dispersed, smaller photovoltaic projects, such as those on rooftops and factory grounds.

Meloni said Starlink offered advanced technology and, unlike European collaborative satellite program IRIS2, was ready.

“This is a European problem,” she said, adding the European Union “did not get there in time to have public technology able to secure this communication.”

Italy is in talks with Starlink, part of Musk’s SpaceX aerospace business, to use its 6,700 satellite constellation in a reported five-year, €1.5 billion ($1.55 billion) deal.

“Space X has illustrated to the government its technology that allows secure communication nationally and globally which means for us guaranteed communications with our diplomats and military missions abroad,” Meloni told reporters in Rome on Thursday.

She said no decision had been made, and that talks were also being held with other private operators, but said Space X had “the most advanced technology.”

She added, “There are no public alternatives.”

Meloni said, “Tomorrow there will probably be public entities able to guarantee the protection of communication. Today there are not.”

On Monday, Elon Musk said he was “ready to provide Italy the most secure and advanced connectivity.”

The European Union has signed to develop an array of 280 satellites known as ISIS2 which will offer encrypted satellite communications to governments, but it will not be ready until at least 2030.

Meloni said that she considered relying on a private operator to protect sensitive military data less than ideal, but could not wait for IRIS2.

“There are two scenarios and neither is optimal,” she said.

Fabiana di Porto, an expert on Innovation and Regulatory policy at Rome’s Luiss University, said, “Musk’s satellites are cheaper than IRIS2 and there are more of them, but the security aspect is important since we are talking about government communications in the hands of a private operator.”

She also pointed out that Italian state-controlled defense champion Leonardo had a stake in IRIS2.

Meloni said that Microsoft had previously worked on data centers in Italy, adding, “no-one complained.”

She said, “Is the problem with Space X that it is private, or is it the political ideas of Elon Musk?”

Handing control of Italy’s military satellite communications to the U.S. tech tycoon has prompted sharp criticism in Italy from opposition politicians given Musk’s recent use of his social media platform X to lambast governments in the U.K. and Germany.

Meloni has forged a friendship with Musk, recently calling him a “genius.”

“I value investments overseas through one lens – national interest – and not through friendships or political ideas,” she said, adding, “I have never talked to Elon Musk of this project.”

The milestone was significant in that it represented a shift in the Defense Department’s approach to connecting military forces across operational domains, known as Combined Joint All Domain-Command and Control, or CJADC2.

Whereas DOD leaders had previously focused — and struggled to show progress — on the ultimate goal of full interoperability, this time they honed in on identifying capabilities to address operators’ most urgent problems and workflows and vetting them through a regular C2 exercise series known as Global Information Dominance Experiments, or GIDE.

Now, the Defense Department’s Chief Data and AI Officer Radha Plumb wants to expand on that work in two ways: targeting more of the department’s C2 capability needs and developing an enduring process through which companies can propose solutions and DOD can evaluate them.

Plumb spoke with reporters July 16 on the sidelines of DOD’s first-ever industry day aimed at helping companies understand its near-term advanced C2 requirements and how they can showcase potential solutions through GIDE.

“We want to bring in more technologists and technology solutions to compete in those venues, so that we can get a better option set of solutions,” she said.

The department to date has conducted 11 GIDE events, each with a particular theme or focus — which is generally classified — connected to a capability gap that the military services and combatant commands have identified. Plumb’s office leads these campaigns, running one every three months.

In the past, the team has reached out to a limited pool of companies to see what solutions they might have to those gaps and then integrated those into the experiments. For example, during its experimentation campaign to validate the CJADC2 baseline capability set last year, the department worked with more than 20 firms.

For future GIDEs — and possibly even the next event in September — Plumb hopes to incorporate tools from a broader set of companies, to potentially include some firms her office engages with during the industry day.

Plumb’s office will use its acquisition hub, called Tradewinds, to communicate its requirements to industry, solicit capabilities and select the most promising to participate in experimentation events.

“If they perform well and they meet a sort of scaled capability need, then we can go through our normal kind of acquisition processes to get broader scaled deployments of them,” she said. “But we’ll also have the prototype vehicle acquisition vehicle to continue that testing and experimentation series.”

Plumb would not discuss what specific GIDE technologies the department will include in its next experimentation set, but operators are generally looking for things like visual interfaces and tools that allow them to integrate data faster and more efficiently to inform decision making.

“We evaluate the technology’s performance both on the overall success of it, but then on whether that candidate technology is the right thing to solve that capability need for the that particular area,” she said.

The experimental applications DOD will end up buying and scaling through this process represent just one category of tools the department wants to integrate into its advanced C2 ecosystem, known as the Open Data and Applications Government-owned Interoperable Repositories, or Open DAGIR.

Plumb announced in late May the department will use Open DAGIR to not only vet experimental tools, but to buy data infrastructure as well as more mature, enterprise-level licenses for applications that will be used by a broad swath of DOD users.

“Open DAGIR ensures the Department can leverage the innovative solutions from the world-class software developers in both the traditional and nontraditional industrial base to create capabilities for our warfighters and decision makers,” she said in a May 30 press release. “We aim to give industry front-row access to both our data and our users to develop relevant and timely software for decision advantage.”

At the TechNet Cyber conference presented by the Armed Forces Communications & Electronics Association International in Baltimore last month, Randy Resnick, director of the Zero Trust Portfolio Management Office, said tagging and labeling, the practice of assigning metadata and identifiers to pieces of data, has been a long-term challenge for the department.

“They’ve been apparently working on this for 12 or more years —15 years — and I think it’s time enough to do something,” he said.

By way of an update on these efforts, Resnick said he approved three pilot programs in conjunction with the U.S. Department of Defense Chief Data and Artificial Intelligence Office and the Department of Homeland Security to develop a plan that would allow for all the necessary conversions and interpretations to process any data tagging and labeling standard in an understandable,repeatable way.

The goal is to have a successful demo of a schema by the end of the calendar year. Resnick also set a deadline of October for an internal working group to brief his office on a solution, even a partial one.

“We’re not looking for perfection,” he said at TechNet. “We have to start implementing something, and then it’ll grow over time as people agree to more tags and more labels. It’s got to be flexible enough to allow for growth.”

In a January study by Defense Innovation Board, researchers found “data access remains the central enterprise-level obstacle to the sharing and use of data for the warfighter.” Part of that is because military departments are “haphazardly” placing data leaders throughout the organization while top-level tech leaders are struggling to enforce their position as a unifier. Other persistent issues like a lack of uniform guidance, sustained funding, workforce gaps and technical silos also make progress on broader zero-trust difficult.

For now, there remain roadblocks that separate the DoD from the data economy it wants, but Resnick is realistic about these challenges and said a solution in development is better than nothing at all.

“That’s the type of solution that I personally am looking for, because that’s what the department needs,” he said.

The aircraft, a modified Bombardier business jet outfitted with Northrop Grumman’s Battlefield Airborne Communications Node payload, or BACN, is lesser known than the Air Force’s stalwart fleet of bombers, fighters and cargo jets. Now, with U.S. troops out of Afghanistan, the E-11 has taken on a new behind-the-scenes role supporting the multinational mission to deliver humanitarian aid to Gaza.

Since March, the 430th Expeditionary Electronic Combat Squadron, the only unit in the Air Force to fly the aircraft, has supported more than 30 of the service’s airdrops in Gaza while operating out of an undisclosed location in U.S. Central Command. Pilots with the squadron who spoke with Air Force Times in an interview declined to discuss specific missions, but offered brief insights into their work in the region.

The Air Force has participated in 40 airdrops alongside coalition partners since March. Those missions were paused for more than a month as Israel’s invasion of the city of Rafah began in early May. The U.S. flew its most recent airdrop June 9.

“The BACN mission provides the Air Force, as well as joint and international partners, line-of-sight and beyond-line-of-sight interoperability, and that’s between dissimilar platforms with varying electronic capabilities,” 430th EECS Commander Lt. Col. R. Clayton “Vector” McCart said. “We capture, translate, combine and, ultimately, relay and extend legacy and modern comms and data links.”

That means flagging potential threats to the cargo aircraft ferrying shipments of food, any course corrections or other important details from mission planners and other troops around the area.

In a region as large as CENTCOM, airdrops can be derailed by a number of variables, from weather to other flight safety issues. The BACN aircraft, which carries two pilots, can quickly relay real-time changes coming from the Combined Air Operations Center in Qatar to mission partners on the front lines.

“We can monitor progress in a way that other platforms may not be able to see … as far as locations of the assets, the weather calls, etc. Weather plays a big factor in the airdrops off the coast of Gaza,” Capt. Britton “Rad” Ellington, an E-11A pilot who previously flew the service’s now-retired E-8C Joint Surveillance Target Attack Radar System, or JSTARS, aircraft, said.

With a range of more than 6,900 miles and a typical sortie time of around 10 hours, the aircraft can take over where satellite phones fail, and interpret transmissions between the military’s myriad communications systems that can’t otherwise connect. While AFCENT declined to say how many of the service’s seven E-11As are operating in theater, the 430th acts as a data lilypad for more than 1,000 hours every two weeks.

“We’re just not relegated to an airdrop. For instance, we could do that as a part of a larger expectation from CENTCOM leadership, and just move around the battlespace. … We’re not really tethered to any one location,” McCart said.

The E-11A is in the process of establishing its new home at Robins Air Force Base, Georgia, within the 18th Airborne Command and Control Squadron. It’s one of four missions replacing JSTARS planes at Robins over the next several years; the squadron is expected to become fully operational by 2027.

The 430th EECS, which deploys out of the Middle East, and doubles as a training unit, will shift its training mission to Robins.

“Every sortie is a training sortie, in the sense that we learn every day that we fly,” McCart said.

Published recently, the “Vision for the IC Information Environment: An Information Technology Roadmap” articulates the pressing need for transformation within the IC, emphasizing how the current strategic environment is vastly different from that of September 2001.

Today, we face adversaries ranging from peer nations to non-state actors, each seeking to challenge our national interests and security. This dynamic environment demands that our information technology infrastructure not only keep pace but also provide a strategic advantage.

The roadmap aims high and purposefully so. It outlines five key focus areas, each designed to fortify, assure, enable, enhance, and accelerate the mission through a comprehensive IT strategy. These areas include:

— Fortify the Mission with a Reliable and Resilient Digital Foundation.

— Assure the Mission with Robust Cybersecurity.

— Enable the Mission with Modern Practices and Partnerships.

— Enhance the Mission with Data-Centricity.

— Accelerate the Mission with Advanced Technologies and Workforce Readiness.

One of the most exciting aspects of this roadmap is Focus Area 4: Enhance the Mission with Data-Centricity. This section highlights the importance of managing and securely using data effectively to expedite mission outcomes and maximize intelligence value. Of particular note, it stresses three main objectives:

Realizing End-to-End Data Management: To achieve data-centricity at scale, the IC must govern and manage data cohesively, at every point of the data lifecycle. This involves comprehensive data management planning, which aligns complex data lifecycle management activities with critical mission architectures. Whether you’re gathering sensor data or video footage in a remote location or extracting insights from that data in an air-gapped facility, such plans will increase the adoption of IC Data Services and applied advanced analytics and AI.

Implementing a Data-Centric Architecture: Valuing data as a strategic tool requires establishing common data standards, models, services, and enterprise digital policies so that everyone is handling mission-critical information in a cohesive way, aligned with the appropriate protection and sharing of each data object. This will facilitate a decentralized data ecosystem, enabling seamless data exchanges and fostering advanced AI/Machine Learning capabilities. Ultimately, this architecture will streamline data sharing and collaboration within the IC and with external partners.

Transitioning Sensitive Data Silos to Data-Centric Enclaves: Sensitive data within the IC is often siloed, making it difficult for analysts to access and utilize information across different enclaves. By transitioning to data-centric enclaves, the IC can break down these barriers, allowing authorized users to discover and access relevant data more efficiently and effectively.

As a whole, this focus area encapsulates the IC’s shift from traditional processes to a data-centric approach. The report details, “Timely, accurate, well-informed insight is key to delivering enhanced mission outcomes. The IC must shift from an organization- and system-centric paradigm to one that is data-centric; preserves organizational equities, authorities, and rights; implements legal/compliance frameworks; and enforces security.”

Centering the data changes everything

Let’s give credit where it’s due; this vision for data-centricity is transformative. Aiming to shift the IC from a paradigm of isolated data silos to an integrated, agile, and efficient data environment, this transition is crucial for enabling analysts to derive actionable insights swiftly and accurately, thereby enhancing the overall intelligence process. By implementing a data-centric architecture, the IC can ensure that data is not just a byproduct of operations, but a core asset that drives decision-making and operational effectiveness.

One of the most critical aspects of this transformation is bringing data-centric security down to the data object level using technologies like the ODNI’s Trusted Data Format and the recently approved Zero Trust Data Format by the CCEB. By doing so, the IC can ensure that data remains secure and can be effectively shared and utilized across various platforms and with international partners and allies. This granular level of security is essential for maintaining the integrity and confidentiality of sensitive information while promoting interoperability and collaboration.

The “Vision for the IC Information Environment: An Information Technology Roadmap” provides a robust framework for modernizing the IC’s IT infrastructure. By prioritizing data-centricity specifically, the IC can unlock the full potential of its data assets, enabling more informed decision-making and more effective operations, which will ultimately prove to be critical for maintaining a strategic edge in a data-driven world and ensuring the United States remains protected.

The roadmap’s vision is ambitious, but with commitment and collaboration, it is well within reach. Together, we have the opportunity, talent, and determination to close IT gaps and open new mission horizons, securing our nation’s future.

Shannon Vaughn is general manager of federal at Virtru, a data security company.

At the TechNet Cyber conference presented by the Armed Forces Communications & Electronics Association International in Baltimore, leaders from the Pentagon discussed capabilities for public, private and hybrid networks. Officials acknowledged there’s a natural appetite for the most exclusive, secure networks in the national security space. And sometimes there is no wireless network infrastructure available in remote warfighting locations far from population centers.

So as the services determine appetite for private networks that offer more control over information sharing, the DoD is guiding them to use open radio access networks, or ORAN, said Juan Ramírez, the director of the 5G Cross-Functional Team at DoD.

“I think what industry wants to hear is there’s actually going to be requirements that come out that … necessitate an open RAN architecture,” he said at the conference. “So you’ll start to see those come out in the next couple of years, pending budgets.”

Certainly, private networks aren’t the only way to go. In fact, sometimes that’s not the best solution, said Lt. Col. Benjamin Pimentel, who leads the Camp Pendleton 5G experiment for Expeditionary Advanced Base Operations.

“Think about when we deploy in a theater,” he said. “A lot of countries that we go to or locations that we go to already have roads and bridges, and it’d be silly to then go and build my own private roads and my own private bridges separate and apart from that to get where I need to go. If those roads and bridges meet my transportation requirement, and they’re not going to fall under the weight of a ‘seven ton,’ we’re going to drive over it.”

But, somewhere like the first island chain, for example, may not have adequate coverage to put up sensors for long-range precision fires. In cases like those, he said, it would make more sense for units to bring private capabilities.

Given China’s rising aggression and U.S. efforts to deter it in the Taiwan Strait, what Pimentel described is the type of environment where current threats seem to colocate.

Regardless, to ensure there is connectivity wherever the need is, Ramirez said the department is looking at ORANs, which allow multiple vendors to operate as one network and provide more flexibility to scale.

ORAN is something the DoD has been pushing aggressively to explore as it simultaneously journeys toward more standard 5G adoption on military installations and “smart bases.”

Ramirez said the department is hopeful it will get additional support from Congress via future defense spending bills that will backup forthcoming requirements with dollars.

The Pentagon’s 2024 budget requested $143 billion in research, developing and testing of emerging technologies including 5G, but also artificial intelligence. Much of the spending in recent years has been for prototyping, and though the Office of the Secretary of Defense has the lion’s share, Ramirez said his office is offering direction to the services for them to budget for 5G.

“We think that pursuing ideas like [ORAN] advanced by the ORAN Alliance all the way to fully open-source code … provides the feature velocity the DoD needs and the ability to innovate quickly,” said Pimentel.

The DoD has an extensive wealth of data. In today’s sensor-filled theaters, commanders can access text, images, video, radio signals, and sensor data from all sorts of assets. However, each data type is often analyzed separately, leaving human analysts to draw — and potentially miss — connections.

Using AI frameworks for multimodal data analysis allows different data streams to be analyzed together, offering decision-makers a comprehensive view of an event. For example, Navy systems can identify a ship nearby, but generative AI could zero in on the country of origin, ship class, and whether the system has encountered that specific vessel before.

With an object of interest identified, data fusion techniques and machine learning algorithms could review all the data available for other complementary information. Radio signals could show that the ship stopped emitting signals and no crew members are using cell phones. Has the vessel gone dark to prepare for battle, or could it be in distress? Pulling in recent weather reports could help decide the next move.

This enhanced situational awareness is only possible if real-time analysis happens at the edge instead of sending data to a central location for processing.

Keeping AI local is critical for battlefield awareness, cybersecurity, and healthcare monitoring applications requiring timely responses. To prepare, DoD must adopt solutions with significant computing power at the edge, find ways to reduce the size of their AI/ML models and mitigate new security threats.

With most new AI tools and models being open, meaning that the information placed into these technologies is publicly available, agencies need to implement advanced security measures and protocols to ensure that this critical data remains secure.

Pushing processing power

Historically, tactical edge devices collect information and send data back to command data centers for analysis. Their limited computing and processing capabilities slow battlefield decision-making, but they don’t have to. Processing at the edge saves time and avoids significant costs by allowing devices to upload analysis results to the cloud instead of vast amounts of raw data.

However, AI at the edge requires equipment with sufficient computing power for today and tomorrow’s algorithms. Devices and sensors must be able to operate in a standalone manner to perform computing, analysis, learning, training, and inference in the field, wherever that may be. Whether on the battlefield or attached to a patient in a hospital, AI at the edge learns from scenarios to better predict and respond for the next time. For the Navy crew, that could mean identifying what path a ship of interest may take based on previous encounters. In a hospital, sensors could flag the symptoms of a heart attack before arrest happens.

Connectivity will be necessary, but systems should also be able to operate in degraded or intermittent communication environments. Using 5G or other channels allows sensors to talk and collaborate while disconnected from headquarters or a command cloud.

Another consideration is orchestration: Any resilient system should include dynamic role assignments. For example, if multiple drones are flying and the leader gets taken out, another system component needs to assume that role.

Shrinking AI to manageable size

A battlefield is not an ideal environment for artificial intelligence. AI models like ChatGPT operate in climate-controlled data centers on thousands of GPU servers that consume enormous energy. They train on massive datasets, and their computing requirements increase exponentially in operational inference stages. The scenario presents a new size, weight, and power puzzle for what the military can deploy at the edge.

Some AI algorithms are now being designed for SWAP-constrained environments and novel hardware architectures. One option is miniaturizing AI models. Researchers are experimenting with multiple ways to make smaller, more efficient models through compression, model pruning, and other options.

Miniaturization has risks. A trained model could undergo “catastrophic forgetting” when it no longer recalls something previously learned. Or it could increasingly generate unreliable information — called hallucinations — due to flaws introduced by compression techniques or training a smaller model pulled from a larger one.

Computers without borders

While large data centers can be physically walled off with gates, barriers, and guards, AI at the edge presents new digital and physical security challenges. Putting valuable, mission-critical data and advanced analytics capabilities at the edge requires more than protecting an AI’s backend API.

Adversaries could feed bad or manufactured data in a poisoning attack to taint a model and its outputs. Prompt injections could lead a model to ignore its original instructions, divulge sensitive data, or execute malicious code. However, defense-in-depth tactics and hardware features like physical access controls, tamper-evident enclosures, along with secure boot and trust execution environments / confidential computing can help prevent unauthorized access to sensitive equipment, applications, and data.

Still, having AI capabilities at the tactical edge can provide a critical advantage during evolving combat scenarios. By enabling advanced analytics at the edge, data can be quickly transformed into actionable intelligence, augmenting human decision-making with real-time information and providing a strategic advantage over adversaries.

Steve Orrin is Federal Chief Technology Officer at Intel.

On June 20, Deputy Defense Secretary Kathleen Hicks approved the strategy, and the department announced it on June 25, coinciding with the 2024 TechNet Cyber conference presented by the Armed Forces Communications & Electronics Association International in Baltimore.

“Fulcrum describes ‘what’ the DoD must achieve with respect to advancing IT for the warfighter and ‘why’ it matters,” according to a department statement.

The strategy offers specifics on broad, sweeping goals by the Pentagon to communicate faster and more securely with warfighters and allies. To date, the government has issued a number of strategic plans and frameworks to double down on cybersecurity and accelerate the military’s development of tools like AI, but this plan takes that a step further by prioritizing user experience and scalable, agile investments.

“It’s called ‘fulcrum’ because it sits at the nexus between our national security strategy, our strategic management plans — our really ‘big thinking’ strategies,” said Leslie Beavers, principal deputy chief information officer, on Tuesday.

With so many tools coming to market from different vendors and small businesses also hoping to break in, the strategy gives leaders guideposts for how to assess what tools and internal resources are needed.

Critically, it also devotes a category to developing the workforce, specifically by broadening the DoD Cyber Workforce Framework to focus more intently on roles for data, AI and software engineering.

The DoD has long said this is an established need.

“There is a recognized shortage of skilled cyber personnel that could potentially impact operational readiness across the Department and put national security at risk,” according to the department’s 2023 Cyber Workforce Strategy. “Despite the vast expansion of cyber educational and experiential opportunities, the nation’s cyber talent pipeline remains limited.”

As in the private sector, government, too, is struggling with too few tech practitioners that make new cybersecurity practices difficult to implement. At the same time, much of the technical guidance handed down by experts in tech shops is written for the actual programmists, not senior executives, so there needs to be an available reserve of workers, DoD officials said at the conference.

“Leadership is a force multiplier and to outpace [adversaries] we have to maximize the talent from our broad array of partners, and recruit and retain our own talent,” said Gen. Timothy D. Haugh, commander of U.S. Cyber Command and director of the National Security Agency, at the conference.

Finally, the new strategy also seeks to streamline governance and enhance use of data to backup decision-making and identify cost savings to deliver joint warfighting capabilities faster amid high-tempo, multi-domain operations.

“We must challenge legacy approaches and move move toward rapid innovation for them,” said Haugh. “We cannot succeed when our existing processes move slower than the rate of innovation change.”