Shadowing Practice: What's next for mainframes and AI? - Learn English Speaking with Video

A trillion web transactions daily.

More than 300 billion inference requests in a single day.

A million plus transactions in a second.

The modern mainframe can handle all that and more with ease.

Welcome to AI Academy.

I'm Lou Folia, a writer and journalist.

And I'm here with Christian Jacoby, IBM Fellow and CTL of Systems Development at IBM.

On average, Visa processes around 8 ,500 global transactions every second.

A mainframe which can handle one million transactions per second would be able to process more than 100 times Visa's volume.

So, Christian, to you, what makes the mainframe remarkable?

And what makes them key to AI's next big breakthrough?

Is it the hardware?

Is it the software?

Is it somewhere in between?

It's really the hardware and the software together.

When we design the next generation of the mainframe, we're enhancing the performance, the scalability, availability features.

But we're doing this in conjunction as a hardware team.

We're working closely with the firmware team and software team to make sure

that we end up with a stack optimized end -to -end solution that clients can use to run their workloads.

IBM has been thinking about AI on mainframes for a long time now, right?

This isn't something we've recently come to.

Ten years at least.

And the IBM Telum processor includes an on -chip AI accelerator.

Why is that important, having the accelerator on chip?

And how does it enhance AI performance?

Yeah, as you're running workloads on the processor, of course, a lot of the data resides within the chip.

And within the chip, we can communicate between different components of the chip, literally in nanoseconds, right?

Nanoseconds is a tiny sliver of time, of course.

Versus if you go off -chip and go to a separate card, you're talking microseconds.

And if you go to a different system, you might be talking milliseconds, right?

And so by putting AI acceleration directly onto the processor chip, we could achieve the lowest possible latency for some of those traditional models.

And then now with the Spire acceleration card that goes into the IO subsystem,

we can run generative AI also with very low latency without needing to cross system boundaries and go into a separate system.

It matters how big that slice of the second is.

It matters big time.

Absolutely.

When you're running millions of transactions every second

and the latency requirements from a business perspective is that transactions need to finish within a handful of milliseconds,

then you really don't want to waste milliseconds and traveling the data between here and some other system for AI inferencing.

Let's drill in on that a little bit.

So what makes mainframes so pivotal for AI?

Their hardware capabilities, their software capabilities, What is it that sort of makes it perfectly suited for AI?

Yeah, so clients are running mission -critical central workloads for their corporations.

They run the central ledger of a bank or they run insurance claims processing on mainframes.

That's where a lot of the data resides.

That's where the transactions get processed.

And of course, you can use AI to increase the sort of business value

of each transaction by checking for fraud on an insurance claim

or on a banking transaction and being able to integrate that directly into the workload, detect fraud, for example,

in the instant of time where the transaction is getting processed is of huge value to clients compared to,

for example, doing it in a cloud a few minutes or an hour after the transaction has already completed.

By doing it directly in the mainframe box itself, you cut out all that latency and that's how you get to the milliseconds of latency for, for example, fraud detection.

You already have all the data, all the transaction processing happening in the mainframe system by doing the AI directly in the processor for traditional models

or in the I .O subsystem on a Spire accelerator for generative models.

You can maintain low latency, get the good throughput and performance that the mainframe is known for,

but also preserve the security capabilities by not needing to send the data to a separate system.

When you're sending the data out into the world, it's slowing things down and it's creating security risk.

Correct.

Exactly.

You're sending it somewhere else that inherently takes time and you're moving the data out of the highly secure mainframe environment.

So now it gets exposed to different attack surfaces that hackers could exploit.

What should organizations consider when choosing AI -ready infrastructure?

And how do mainframes address those needs better than other platforms?

When clients embark on AI journeys in the context of the mission -critical workloads, We often talk to them about the readiness of the data,

the readiness of the models.

Do they have the capabilities in the data science departments to kind of train their own data into the models

that they will be using?

And then also look at, you know, what latency requirements do you have?

What throughput requirements?

How many inferences per second do you need to run in that workload?

And by working with clients through those considerations early on,

we can make sure that clients are actually successful in deploying AI solutions, not just developing them, and then kind of getting surprised about how hard it is to scale them to actual deployment.

So we're in a room that has seen a lot of mainframe history.

You spent the bulk of your career working on mainframes.

What excites you about the future of mainframes?

We've been already developing the next generation and the generation after that.

We're always working on three mainframe generations at the same time.

So it's your job to think about what the future's going to look like and build a machine that serves it.

And that's exciting and a lot of fun.

Could you talk a little bit about some of the use cases that may have surprised you?

Yeah, I'll give you an example.

There's one client who's actually using mainframes to analyze imagery of seagrass, satellite images.

Like seagrass, like on the beach seagrass?

Yes, yeah.

And they see beach grass dying in certain areas, and they use that information

that they obtained from the AI models to then do ecological measures to protect more of the seagrass.

So when you start designing something, and then you see where it goes, like protecting seagrass, like how cool is that for you to see the expanse?

That is super exciting.

Like we have, when we design the hardware and the software, we have use cases in mind that we're optimizing for.

And then once it's in the hands of our customers, their innovation starts and how they can use all these capabilities for use cases we would have never imagined.

And that is one of the most fascinating parts to see how this sometimes evolves.

Yeah, totally.

So that's the future.

But even today, just think about it.

Every time you pay using a credit card, withdraw cash from an ATM, make reservations with any airline,

file your taxes or access your medical records or basically do a thousand other things, it's likely you are being served by a mainframe.

Now, imagine a day without mainframes.

Mainframes are without a doubt integral to enterprise IT.

And today, with new integrations and enhancements, they are redefining their role in modern IT.

100

So, to answer the question we started with, what makes the mainframe unique?

101

It's the hardware and the software and everything in between.

102

It's how the end -to -end stack is optimized from the Silicon App for mission -critical workloads.

103

And I think that we could both agree that that's what makes the mainframe so remarkable.

104

And that's what makes it great for AI.

📺 Same Channel

Crypto Agility Explained: Protect Data from Quantum Computing Threats

Business EnglishUniversity StudentsWorking Adults

✨ Recommend Video

114 sentences

About This Lesson

Ready to boost your English speaking practice? Dive into this engaging conversation about mainframes and AI to sharpen your listening and speaking skills. You’ll practice following technical discussions, picking up industry terms, and mimicking natural dialogue flow—perfect for shadow speech exercises. By the end, you’ll feel more confident tackling complex topics while improving your pronunciation and fluency.

Key Vocabulary & Phrases

Mainframe: A powerful computer used for large-scale, mission-critical tasks (e.g., "Modern mainframes handle millions of transactions per second").
Inference requests: AI processes that use trained models to make predictions (e.g., "300 billion inference requests daily").
Latency: The time delay in data processing (e.g., "Low latency is crucial for real-time fraud detection").
Generative AI: AI that creates new content (e.g., "Aspera cards run Generative AI with low latency").
Mission-critical: Essential to a business’s operations (e.g., "Mainframes manage mission-critical banking systems").

Practice Tips

Use the shadowing technique to master this video’s pace and tone! Start by listening to short 5-second clips, then repeat them immediately, matching the speakers’ rhythm and emphasis. Focus on words like "nanoseconds" and "milliseconds"—notice how they’re stressed. The dialogue has a conversational flow, so mimic the natural pauses between questions and answers. For extra practice, use a shadowing site to record yourself and compare it to the original. This helps you catch subtle intonation changes. Remember, even 10 minutes a day of shadowing will boost your fluency quickly. You’ve got this—every repeat brings you closer to sounding more natural!

Grammar in this video

The structures the speaker uses most, with the exact words from the video:

Structure	In the video
Present perfect have/has + past participle — a past action that still matters now	has already completed · has seen · We've been
Passive voice be + past participle — the focus is on what happens, not who does it	is known · being served · is optimized

What is the Shadowing Technique?

Shadowing is a science-backed language learning technique originally developed for professional interpreter training and popularized by polyglot Dr. Alexander Arguelles. The method is simple but powerful: you listen to native English audio and immediately repeat it out loud — like a shadow following the speaker with just a 1–2 second delay. Unlike passive listening or grammar drills, shadowing forces your brain and mouth muscles to simultaneously process and reproduce real speech patterns. Research shows it significantly improves pronunciation accuracy, intonation, rhythm, connected speech, listening comprehension, and speaking fluency — making it one of the most effective methods for IELTS Speaking preparation and real-world English communication.

Shadowing technique: read the full step-by-step guide →

Shadowing Practice: What's next for mainframes and AI? - Learn English Speaking with Video

Crypto Agility Explained: Protect Data from Quantum Computing Threats

AI, Machine Learning, Deep Learning and Generative AI Explained

Inside The Hidden Bottleneck In The AI Chip Boom

System Design: Why is Kafka fast?

Fireside Chat with Jensen Huang & Juju Chang - 2026 Annual Gala

Gemma 4 Just Released - Whats In? In 2026!

how AI broke the SWE Career Ladder

The Engineering that Runs the Digital World 🛠️⚙️💻 How do CPUs Work?

The Global Power Play

We Let $200 GPT-Pro Think for 7 Days on Unsolved Math

The AI Atrophy Problem How CIOs Fight It

NVIDIA GTC 2025 - Keynote Highlights by NVIDIA CEO Jensen Huang

About This Lesson

Key Vocabulary & Phrases

Practice Tips

Grammar in this video

What is the Shadowing Technique?