Funding Cuts Stall Critical Chip Security Research

As part of the Trump administration batch to reduce government spending, the US Department of Defense fell more than $ 580 million of financing, including $ 360 million in grants, some of which were linked to universities. These brake cuts were pumped to project projects throughout the country, including research focusing on national security, which President Trump considered non -alignment with national priorities.

Among the projects that were placed on an unspecified contract, it was a multi -year study, as it was funded at a value of $ 3 million at Cornell University through the National Defense Licensing law (NDAA) to examine cyber weaknesses in the chain of supplies of American conductors. It was one of the more than 75 DOD projects in Cornell, who got a stop.

The research came amid the growing geopolitical tensions and the global male arms race. The chips are necessary for both the infrastructure of the prosecution and the military systems. Without understanding risks – such as design and IP design attacks – some experts warn that stopping this work may leave the supply chain of chips open to sabotage, which constitutes threats to American economic and national security.

The research led Sarah Krebis, a government professor at Cornell and the director of the university’s technological policies institute. In conversation with IEEE SICTRUMKREPS explained what its team started studying, how the project stopped, and why the supply chain for the chips is more urgent than ever. The following has been released for length and clarity.

Sarah Krebis on:

How did this project happen?

Sarah Kribes: About two years ago, Micron arrived at the Cornell Technology campus. They were building a $ 100 billion manufacturing facility near Cerekios and were concerned about the security of the entire supply chain. They said they do not have the mental frequency range or internal experience to know where the weaknesses exist, and therefore they were unable to know how to treat them.

Cornell Tech said it was a multidisciplinary project that they could not deal with on their own, so they reached us. My work focuses on emerging technologies – AII, semiconductors – with national security focus, which Micron was more interested in, especially with chips that ultimately end with DOD as a customer. We collected one office devices around the project, and in the end we got to NDAA 2025. We finally received the grant in December 2024, and the work was set to run for the 2025 calendar year.

Why did you feel this extreme urgent?

KrepsWe have started this soon, launched the mutation of artificial intelligence. I have worked in the artificial intelligence space since 2018, but it has been completely accelerated. Political geography of artificial intelligence. If there is a goal to be a global leader in artificial intelligence, we must secure the infrastructure inherent in it.

Part of the problem is that the infrastructure is not very exciting. It is easy to take a matter of it. But if you want to be a pioneer in artificial intelligence, meet national security needs, and achieve local manufacturing goals, you need a flexible supply chain. Our early work has shown that we are not there yet.

What is the scope of research, and what you are hoping to achieve?

KrepsWe were pressing a lot of work in a one -year window. There were three phases, each with the main weaknesses in the semiconductor supply chain.

The first stage included the appointment of material networks and information networks: defining the critical contract, suffocation and repetition points to assess the flexibility of the supply chain. The material aspect follows the consequences of manufacturing, while the aspect of the information looks at data flows, safety protocols, and alternative communication methods to prevent theft or turmoil.

The second stage was brought in the corner of cybersecurity. We used to conduct the penetration and test of the main manufacturing stages such as WeFer FAB and Lithography to determine the security gaps and high -risk contract based on the test results.

The third stage in the fall would merge the results by simulating the network and table exercises. This would design how the disorders infiltrate, analyzing the double weaknesses of materials and information flows, and mitigation strategies tested. We would have attended the stakeholders in the industry, the government and the academy to verify the authenticity of the risk scenarios and the recommendations of the form of form. In the end, we were debtor to the congress with an unknown version and classification of these results.

We have also started proving the concept-appearance in similar use and abuse of cases, such as Israeli emerging attacks-Cubs and Stuxnet. The idea was to understand what the threat environment actually seemed by referring to examples in the real world of the types of risks that we were organizing. The basis for the classification of the threats we were developing.

Kreps: We got something that stopped in the middle of March, so we were still in the scaling phase of the project-where the team gathered together, building the laboratory, and preparing devices. We have spent a few hundred thousand dollars on advanced servers and our account to build the internet test scope.

One of the early ideas was the lack of different parts of the supply chain that really understood their weaknesses, not to mention knowing the shape of their supply chain. If you do not know who suppliers in the three layer are, you cannot know what the weaknesses are. We were supposed to visit Micron in Boise a week. But we could not go.

They are clearly attacked every day, and they do not have experience to evaluate these threats and reduce them systematically.

She walks to me through some electronic risks in the real world in the supply chain.

KrepsWe have an amazing set of potential attacks that can come from this. At the source, in the design stage, you can enter IP theft or Trojan horses. Your opponent can put something harmful in the program in the front end can lead to cloning or lathe chips.

In the manufacturing stage, which is often abroad – Taiwan, South Korea, China – Mabek itself can be at risk. You can imagine changing doping levels or including hidden functions without knowing the designer. This may destroy or manufacture low return or chips that do not work.

During the post-FAB test, contractors can make accurate-inverted symbols in fixed programs or testing environments that are difficult to discover after publishing. In the tools phase, you have risks of harmful programs in design programs that can affect chips. These tools are often updated in this field, which becomes another attack.

In packaging and distribution, harmful updates can be included in the last mile. Then there is an idea that the inert code can allow the slide to be disabled in the future – the Kill -Kill key risks. These are all real concerns.

What is your reaction to reducing financing, and how does this affect your team?

Kreps: Frankly, I was blind. In the first weeks of Dog, we felt as if we had escaped a bullet. We felt that the project seemed essential for national security, and that anyone who cares about the country will see that stopping this type of work shoots at the foot of the foot. But by mid-March, we fell into the intersection of political repercussions in higher education-a billion dollars were withdrawn from Cornell.

In the project, PhD students had to stop working immediately because their support is linked to the grant. We had some of the help of officials, university students, all of which went away. I really thought about the financing and training pipeline from universities to industry.

Students often continue these programs to work in Spacex, Micro and ASML. These industries – which are now very essential for the American economy and national security – will lose the training pipeline. I think this is a real problem. Federal grants make this training possible. If you cut it, you only break the search; You create a deficit in national security and space that will appear within three to five years.

Is there any opportunity to re -finance?

KrepsWe are working on a few different angles. But this was an executive decision, so it is not something that Cornell can change unilaterally. Colombia, for example, has received a list of the conditions they need to fulfill to restart their financing. We didn’t get anything like that, which made it more difficult.

Are you optimistic that the project will resume?

KrepsDepends on today. Sometimes we seem to make progress. At other times, the struggle between the Supreme ED and the executive authority feels without a solution. We are doing what we can – move forward in theses, and keep the infrastructure in its place – hoping that there will be ready to operate the project. So, if we get green light and when we get the green light, we will be ready.