IOT News
Purdue's Semiconductor and Manufacturing Research: Technologies to Watch
Published by IOT Indiana Research & Editorial Team
- The foundation: what Purdue has publicly established
- Watchlist: seven technologies and why they matter to Indiana
- Deep dive: advanced packaging
- Deep dive: processing and memory together
- Deep dive: physical AI
- How a company without a lab can engage
- A hypothetical example: a machining supplier and a capstone team
- Questions to settle before starting any collaboration
- Realistic expectations
- Signals to follow over the next year
- The bottom line
Universities rarely make headlines the way factories do, but they decide what those factories will be able to build a decade from now. In Indiana, Purdue University sits at the center of two stories that matter to industrial IoT: the state's push into semiconductors and the growth of physical artificial intelligence, the field that puts AI into robots, vehicles and machines. This week the university and Google Public Sector are hosting a summit in Indianapolis focused on moving physical AI from pilot projects to proven results, which makes this a good moment to take stock.
This article is a watchlist, not a forecast. It describes what Purdue has publicly said about its programs, flags the research directions most likely to matter to Indiana manufacturers, and explains how a company without a research department can get involved. Where an item is my own judgment about relevance, the text says so.
The foundation: what Purdue has publicly established
Purdue's semiconductor work is not new. The university's chief semiconductor officer, Mark Lundstrom, has said its leadership dates to the 1930s. The Birck Nanotechnology Center, which opened in 2005 in the Discovery Park District, includes a cleanroom of about 25,000 square feet that Lundstrom has described as a facility only a handful of universities in the country have. Semiconductors are one of four pillars of the university's Purdue Computes initiative, alongside its work in computing and AI.
Several recent developments connect that foundation to industry:
- SK hynix. The company plans a roughly $3.87 billion advanced packaging and R&D facility in Purdue Research Park. Reporting from 2024 said the university was finalizing a research partnership agreement with SK hynix, with collaboration expected on topics including combining processing and memory and generative AI models.
- Workforce programs. Purdue runs semiconductor degree and training efforts, partnerships with Ivy Tech Community College and the federally supported SCALE program, which links defense-related microelectronics careers with education and has reached thousands of K-12 students across several states.
- International ties. Purdue participates in the UPWARDS Network, a U.S. and Japan workforce and research partnership that includes Tokyo Electron.
- Physical AI. The Institute for Physical Artificial Intelligence covers areas from robotics to autonomous vehicles. A multiyear partnership with Google Public Sector, announced in January 2026, covers cloud infrastructure, joint research and AI skills training.
Watchlist: seven technologies and why they matter to Indiana
| Technology | What it is | Why Indiana manufacturers should care |
|---|---|---|
| Advanced packaging | Stacking and connecting multiple chips in a single package, as in high-bandwidth memory | Anchor employer SK hynix is building here; supply chains need precision parts and process monitoring |
| Processing and memory integration | Moving computation closer to or inside memory to cut energy and time spent moving data | Could reshape AI hardware, including the compute used at the factory edge |
| Physical AI and robotics | AI systems that perceive and act in the physical world | Applies to inspection, material handling and flexible automation |
| Microelectronics workforce methods | New training formats for technicians and engineers | Eases a shared hiring constraint across industries |
| Sensors and nanotechnology | Smaller, cheaper, more capable sensing devices | Cheaper condition monitoring and environmental sensing |
| Cybersecurity of connected equipment | Securing devices, supply chains and networks | Large customers increasingly require supplier security evidence |
| Data and AI skills at scale | AI competency requirements for graduates | A future workforce more comfortable with analytics |
The first, second and fourth items rest on what Purdue and SK hynix have said publicly. The relevance judgments in the right-hand column are mine, and companies should test them against their own situations.
Deep dive: advanced packaging
For years, progress in computing came mostly from shrinking transistors. Increasingly it comes from how chips are assembled together. High-bandwidth memory stacks several memory dies vertically and links them with very short connections, which lets AI accelerators read data far faster than traditional memory permits. That is the market SK hynix is entering in West Lafayette.
Packaging is a manufacturing problem as much as a chip design problem. It depends on precise alignment, controlled heating and pressure, clean environments and extensive inspection. Academic research in this area tends to focus on materials that handle heat, bonding methods that improve yield and testing approaches that find hidden defects. For Indiana's suppliers, this means opportunities in precision tooling, process monitoring and environmental control, and a need for people who understand both mechanical manufacturing and electronics. Our article on where industrial IoT fits in SK hynix's plant covers the supplier angle.
Deep dive: processing and memory together
Moving data between memory and processors consumes a large share of the energy and time in AI computing. Research that brings computation closer to memory aims to reduce that cost. Whether specific approaches reach commercial products is uncertain, and many promising lab results do not. The practical lesson for manufacturers is to watch how future AI hardware changes cost and power. Edge devices that run models on the factory floor could become cheaper and more efficient, which would widen the set of problems worth solving with local AI.
Deep dive: physical AI
Physical AI is where software intelligence meets motion, force and safety. Purdue's institute spans robotics and autonomous vehicles, and the university has framed the current challenge as moving from pilots to scaled, measurable results. For factories and warehouses, relevant applications include:
- Flexible assembly and picking. Robots that handle varied parts with less custom fixturing.
- Autonomous mobile equipment. Vehicles that navigate plants and warehouses and adapt to changing layouts.
- Inspection. Systems that perceive defects and learn from corrections.
- Human and machine collaboration. Safer interaction around moving equipment.
The hardest problems are reliability, safety certification and handling rare situations. Research partners can help companies assess where the technology is ready and where it is not.
How a company without a lab can engage
- Start with Purdue Manufacturing Extension Partnership. It works with small and mid-size manufacturers on practical improvement and can connect you to university expertise.
- Use the industry partnership channels. Purdue's Office of Industry Partnerships helps companies find collaborators, sponsor projects and recruit.
- Offer a real problem, not a vague interest. Researchers respond to specific challenges with data, such as a defect that resists detection or a process that is hard to monitor.
- Consider student projects. Senior design and capstone projects can address bounded questions at low cost, and they build relationships with future hires.
- Join consortia and events. Regional events and industry groups provide connections and early visibility into emerging tools.
- Be clear about intellectual property early. Agree on ownership and publication terms before work starts. University and corporate expectations differ.
A hypothetical example: a machining supplier and a capstone team
This scenario is invented to show how a small collaboration can work. A 50-person machine shop makes housings for equipment used in cleanroom environments. Its customers want proof that parts stay clean and dimensionally stable through washing and packaging, but the shop has no good way to correlate its cleaning process with particle counts.
- The owner brings the problem and three months of washing and inspection records to a university contact through an industry partnership office.
- A senior design team takes on a bounded question: which wash parameters most affect residual particles, using a few low-cost sensors and a simple test fixture.
- The students present results and a prototype monitoring setup. The shop keeps what it finds useful, hires one of the students and asks a faculty member whether a larger study is worthwhile.
The project will not create a new technology. It does give the shop evidence, a stronger sales conversation and a relationship with people who understand its field. That is a realistic first outcome from a university collaboration.
Questions to settle before starting any collaboration
- What decision will the results inform, and who in your company owns that decision?
- What data can you share, and what must stay confidential?
- Who owns inventions or software that come out of the work?
- What will you do if the result is negative? Negative findings are still findings.
- Who on your side will spend time with the team each week?
Realistic expectations
University research moves on different timelines than factory operations. A discovery may take five to ten years to reach products, and many never do. Companies should treat Purdue as a source of talent, early insight and problem-solving capacity rather than a vendor of ready solutions. Expect to invest staff time. Expect some projects to produce knowledge rather than immediate savings.
Equally, treat announcements carefully. Plant schedules shift, research partnerships evolve and funding changes. The details cited here come from Purdue publications and news coverage from 2024 and 2026, and readers should check current sources before making decisions.
Signals to follow over the next year
- Progress of SK hynix's construction and the growth of the surrounding supplier cluster in Purdue Research Park.
- Outputs of the Purdue and Google Public Sector partnership, including joint research and training programs.
- Technician and engineer program growth, including Ivy Tech collaboration.
- New corporate research agreements announced by Purdue's industry offices.
- Federal and state funding decisions affecting semiconductors and manufacturing research.
The bottom line
Purdue gives Indiana something many states lack: a major research university with real depth in semiconductors and an expanding focus on physical AI, located next to one of the state's largest industrial investments. The technologies to watch are advanced packaging, processing and memory integration, physical AI, sensing and workforce innovation. Companies that bring a concrete problem, a bit of data and some patience can turn that proximity into an advantage.
Frequently asked questions
Related articles
IOT News
