What I Saw at TechEx: Why AI Is Bigger—and Closer—Than We Think

by | Jun 17, 2025 | Events , Computing

juliet correnti-1Post by Juliet Correnti, CEO, Radeus Labs: 

Last week I attended TechEx 2025, a conference billed as “seven events in one.” TechEx truly delivered a cross-section of everything AI—from infrastructure to ethics to practical tools for business.

As someone who spends most days thinking about secure systems and simulation hardware, I came home with a bigger takeaway: AI isn’t just for data scientists and Silicon Valley startups. It is set to have massive impacts across industries. The road ahead is exciting, and I’m ready to help turn these tools into real-world impact.



Speech is the New Frontier

200One of the standout themes at the event was voice AI. Not just speech-to-text or smart assistants, but a true merging of natural language models with speech recognition—bridging the gap between how we think and how we talk. Think less Siri, more Star Trek Next Gen's "Computer".

It’s fast. It’s intuitive. And it’s already here—being tested in drive-thrus, deployed in customer service, and embedded in data centers.

But with this progress comes risk. One of the sessions I attended dove into the biases embedded in training data—particularly for voice models. The way we speak, our dialects, educational backgrounds, and even accents can shape how (or if) we’re understood by AI. As business leaders, we’ll need to think about inclusivity and accuracy as much as innovation.


TechEx RecapAgentic AI and the Rise of Autonomous Decisions

Another term that caught my attention was “agentic AI.” It describes AI that doesn’t just respond—it acts autonomously, makes decisions, and learns continuously. It’s no longer about giving commands; it’s about defining boundaries and letting systems take action.

That shift—toward AI acting like a teammate rather than a tool—has massive implications. It’s what makes voice AI usable. It’s what powers automation in everything from logistics to law. And it’s why trust and control are the next battlegrounds in AI adoption.


Why Security (and Cost) Should Still Be Center Stage

For all the innovation on display, security felt under-discussed. Many cloud-based AI companies at TechEx assumed security was “baked in,” but I couldn’t help asking: Is it? When your models, data, and customer interactions live in the cloud, how are you protecting IP, compliance, and reputational risk?

I spoke with a rep from Deloitte who reinforced a growing trend: on-prem AI isn’t going away. Not just for security reasons, but because of cost. Training and retraining models—especially at enterprise scale—can rack up staggering cloud bills. On-site infrastructure may offer both control and cost-efficiency in ways we haven’t fully valued yet.


A Wake-Up Call for Business Owners

Some of the biggest companies in the world—Visa, TransUnion, Microsoft—shared that they, too, are still figuring this out. Despite their resources, their AI use cases remain early-stage. And yet, the message was clear: businesses that empower employees to experiment with AI tools, run internal hackathons, and continuously retrain will be the ones that move ahead.

One quote from the show stuck with me:

“In the future, work will be humans teaching AI to take their jobs—once they see their valued role in the total workforce ecosystem. Not before.”

This doesn’t have to be scary. If your business values lean processes, like we do at Radeus, and if you’re constantly trying to drive out waste and increase impact—AI is a natural fit. It's not about replacing people; it's about redeploying talent to higher-value work.


We’re Headed to Vegas Next! Meet Our Team at the AIAA Aviation Forum, July 21–25

aiaaIf you’re serious about where aerospace, defense, and next-generation computing intersect, you should consider attending the AIAA Aviation Forum in Las Vegas this July

It’s the largest event for aviation research, development, and technology, bringing together:

  • Top aerospace leaders, engineers, and innovators from government, academia, and industry
  • Discussions on everything from advanced air mobility to digital engineering and AI integration
  • Insights into emerging trends around autonomy, cybersecurity, sustainable aviation, and defense readiness
  • Hundreds of technical sessions for anyone building, securing, or optimizing aviation systems

We'll be there at Booth 723, showcasing Radeus Labs’ digital flight recorders (DFRs), AI-ready servers, and secure peripherals for simulation and training environments. If you're involved in mission-critical computing or defense tech, stop by—or you can make an appointment to meet us there.

We'd love to talk shop—and swap ideas about where AI is headed next. We’re not just watching the future of AI and aviation. We’re building it!

Blog

See Our Latest Blog Posts

I/ITSEC 2025: The Trends in Simulation Worth Paying Attention To

itsec 2025-radeus team-ck tan leading technolgy

After a year of AI dominated headlines, I/ITSEC 2025 brought the conversation back to infrastructure. This show was about systems, sustainability, and the real engineering work that keeps training and simulation programs running

And for Radeus Labs, this year felt different; busier, more technical, and packed with the right people in the right conversations.

Here’s what stood out.

MTBF: What It Actually Means and How to Use It Correctly

You're evaluating GPU computing platforms for a mission-critical deployment. Vendor A quotes an MTBF of 100,000 hours. Vendor B claims 150,000 hours. The choice seems obvious: go with Vendor B for 50% better reliability, right?

Not so fast.

If you're making hardware decisions based on MTBF comparisons alone, you're likely making decisions based on incomplete, or worse, misunderstood, information. Mean Time Between Failures remains one of the most widely cited yet most profoundly misunderstood metrics in reliability engineering. And for defense, aerospace, and mission-critical computing environments where failure isn't just inconvenient but potentially catastrophic, this misunderstanding carries real consequences.

Let's set the record straight on what MTBF actually tells you, what it doesn't, and how to use it properly alongside other reliability tools.

SC25: High Performance Computing Meets the AI-Driven Future in St. Louis

A few weeks ago, the Radeus Labs team joined thousands of HPC professionals, researchers, and developers in St. Louis for SC25, the International Conference for High Performance Computing, Networking, Storage, and Analysis. What we found was a conference at an inflection point, where traditional supercomputing meets the explosive demands of AI infrastructure, and where the future of connectivity is being built in real-time.