The Rundown - Deep Dive: Why Agentic AI is the Next Phase of the AI Boom

Episode Date: March 21, 2026

In this deep dive, Zaid breaks down the next chapter of the AI boom: agentic AI. He explains how AI is evolving from simple chatbots into autonomous agents, and why that could be a major boost for Nvidia and help justify the massive capex spending from Big Tech. We also look at how companies from JPMorgan to FedEx are already deploying AI agents, and what this shift could mean for the broader economy, the future of work, and your job.

Welcome back to the rundown for another weekend deep dive. Today, we are talking about the next phase of the AI boom, AI agents. For the past couple of years, the AI boom has been about training large language models, you know, the chat GPs and the claws of the world. But now the focus is shifting from simply talking to AI to having AI actually do things for you. We are entering the era of AI agents and the implications for the economy, the stock market, and your job are massive. So in today's episode, we'll break down what Agentic AI actually is,

why Nvidia is betting the house on it, and what this means for the AI bullcase moving forward. We got a great one for you today. Let's dive in. The best way to understand Agentic AI is comparing it to what came before it, which was generative AI. I'm talking about the chat GPT moment. With generative AI, people would type in a prompt

and the AI would give you an answer. Now, people could generate text, images, code, even video, all with a simple prompt. You could ask the AI to create a workout plan for you or rewrite an email or write a history essay or a business strategy, and the AI would give you an answer within seconds. When this technology first dropped, it was a huge breakthrough and the adoption exploded overnight. In fact, Chad GPT grew to 100 million monthly active users within two months of its launch. I mean, that is the fastest service to reach 100 million users ever. Today, chat GPT has more than 900 million weekly users with about 50 million people paying for premium features.

I am one of those people that pays for chat GPT. And look, it's not just chat GPT anymore. Other tools have popped up like Anthropics Claude and Google's Gemini, all of which have hundreds of millions of users. So these tools are amazing and very popular, but they do require the user to be in the driver's seat. You know, you ask it a question, it gives you an answer, but the AI can't actually do anything in the real world. You know, it can't log into your email and reply to all the unread messages. It can't file your taxes. It can't manage your calendar.

It's just good at generating outputs like text, images, and video. But what agenic AI does is it takes it a step further. Instead of the AI just answering your question, AI agent can break down a task into steps. It can use tools. It can interact with software and try to complete a job with minimal human supervision. The way I think of it is like generative AI is like, you know, it can read memos, it can send emails,

It can book meetings. It can follow up with clients. It can update the CRM. It can compare vendors and file expense reports all with minimal human input. Now, these tools are relatively new, but pretty powerful already. Anthropic has clawed code. OpenAI has code eggs. And then there's OpenClaw, which has gone insanely viral. So we're at a point now where AI agents can do real economic transactions, even going as far as sending and receiving money. We're actually going to talk more about that in a bit. So this is leading to a transformation of the AI industry. In fact, NVIDIA's CEO Jensen-Wan thinks that we're at an inflection point. So let's talk about what this means for the AI economy and why Jensen-Wong thinks that the rise of AI agents could add an extra $500 billion to Nvidia's revenues. So we're starting to see the rise of agentic AI. So let's talk about why investors are paying attention.

See, for the last few years, AI has been about training large language models. That required tens of thousands of GPUs. It consumed enormous amounts of energy. these giant data centers pack with GPUs ran 24-7 for weeks or months at a time to train these models. Companies like OpenAI, Anthropic, Meta, and Google have been spending tens of billions of dollars on training AI models for the last few years. But now the focus is shifting from training AI models to actually using them. That is called inference. Inference is when an AI model answers questions from a user and generates output.

Infference is also what powers AI agents to execute their tasks. A very simplified analogy that I can think of is AI training is like someone going to business school. And then AI inference is like hiring that person to do a job and answer questions. And this shift to inference is impacting the AI infrastructure spending. According to research from Gartner, global capital spending on inference infrastructure is going to surpass spending on training for the first time ever this year. And this gap is expected to widen fast. By 2029, companies are projected to spend nearly twice as much on.

inference compared to training. The thing is, these AI models are now so good that there's less urgency to train new ones. Instead, there's more focus on getting the most out of these existing AI models. Now, the other side effect of the rise of a genetic AI and the focus on inference is that this is going to exponentially increase the demand for compute. See, if you think about it, a normal AI chatbot gives you one response. But an AI agent needs to think through the task, call multiple tools, check its own work, retry if something fails, pull an outside data, and keep running in the background to complete the work. All that requires a lot more tokens and a lot more computing power.

And it also requires a different type of hardware. Most people know about Nvidia's GPUs, they're really good at training AI models. Well, inference and agentic workloads are different. That requires CPUs working alongside GPUs to get the maximum performance. The CPU acts as a brain that can, can think and execute multi-step tasks. It can call external tools like APIs and databases and you know manage memory better. And that's why invidia is now making CPUs themselves alongside their GPUs. Invita became the most valuable company in the world because their GPUs were the best

at training AI models. The company's data center revenue, which represents their AI Chips business, has grown nearly 13x since the launch of chat GPT. But Nvidia's CEO Jensen Wong understands that the future is agentic AI. At the recent, NVIDIA GTC developer conference, Jensen introduced the new Vera CPU, which is specifically designed to handle these agenic workloads. He also introduced a new central processor and an AI system built on the technology from GROC, which is a startup to company basically acquired for about $20 billion back in December of last year. Now, just to be clear here, I'm not talking about Elon's GROC chatbot. The GROC that I'm talking about ends with a queue, and they make chips

specifically for inference. So, Nvidia is now positioning themselves to be the leader when it comes to inference. By the way, Jensen was talking a big game at the GTC conference. He now expects Nvidia to make over $1 trillion in revenue from the Blackwell and Vera Rubin chips alone through 27. That's up from last year's forecast of $500 billion by the end of 2026. So the company is expecting the AI boom to continue for at least another year. You know, I'll be honest with you guys. I was kind of skeptical of all the CAPEX spending from these big tech companies. You hear these gigantic $100 plus billion capEx numbers and you're like, is it actually going to pay off?

But with the rise of AI agents, I can kind of see how that's going to require exponentially more computing power. Because it looks like every company in the world is going to be utilizing these agents to transform their work. So let's talk more about how some of these companies are using AI agents and what the future could look like. Look, I already know some of you guys are wondering, Are AI agents legit or is it just another overhyped piece of technology like the Metaverse or NFTs?

Well, it's still pretty early, but some companies have started to test these tools. The Wall Street Journal just had a report talking about how FedEx is building out an army of AI agents to work alongside their human workforce. By 2028, FedEx expects to have AI integrated into more than half of its core operational workflows. In fact, the company is already using AI to help packages, clear customs at a quicker pace. So people are also saying that AI agents could change the way that we all shop. So, you know, instead of browsing through Amazon for a product, imagine telling your AI agent that you want to buy a specific item with a certain budget, and then the AI agent

goes out and does that for you. Maybe the AI agent will find all the items that meet your criteria and then ask you to pick one. Now, personally, I like spending four hours comparing products and reading all the Amazon reviews before buying a $30 rice cooker, but I guess if an AI agent could do that, it would save me a lot of time. It's less fun, but, you know, more efficient. All the jokes aside, according to Morgan Stanley, agentic shoppers could reach up to $385 billion in U.S. e-commerce sales by 2030, which would represent around 20% of the e-commerce market share. I'm still a little skeptical about that, but I can see the potential. Now, banks are also testing AI agents. There was a

report from MIT Sloan that said that J.P. Morgan is testing AI agents for fraud detection, financial advice and loan approvals. But what I find to be most interesting is the role that stable coins could play here. In fact, Coinbase is looking to build the infrastructure that would allow AI agents to make payments and carry out tasks. The bet here from Coinbase is that agents will use stable coins on Coinbase's blockchain to automatically make transactions

in the digital world. So I find that to be very interesting. I can kind of see the promise there. By the way, if you guys want to learn more about this, I actually talk to the CFO of Circle about the role that stable coins might play. play in agentic commerce. I'm not going to lie, that conversation had me pretty hyped about the future of stable coins. That interview will be posted on Sunday. It might already be up by the time

you'll listen to this podcast. So definitely go check that out if you want to learn more. Big picture here, though, is that AI is now getting to the point where it can actually do stuff beyond just rewriting emails or helping students cheat on essays. Agents can do the work of humans. And if this tech continues to develop at this rapid pace, I mean, this could change the way entire industries work. And that's starting to make people feel pretty uncomfortable. So let's talk about it. CEOs and CFOs might be salivating at the idea of AI agents boosting productivity and profit margins, but there's a growing wave of anxiety about what this all means for everyday workers.

In fact, we're already starting to see that AI might be impacting entry-level jobs. The unemployment rate for recent college graduates is 5.6%. And if you broaden that out and look at the unemployment rate for younger workers between the age of and 27, that's at 7.8%. Now, historically speaking, that's actually not that bad, but the rate has been climbing since the post-COVID bottom in 2023. It could be a sign that companies are starting to replace entry-level workers with AI. Anecdotally speaking, whenever I talk to college kids,

the biggest problem that I hear is the lack of entry-level jobs. In fact, recently, the CEO of Service Now said that he thinks the unemployment rate for recent graduates could spike to as high as 30% in the next couple of years because AI agents will be able to handle tasks typically done by an entry level or mid-level worker. So that's a pretty big concern. Young people might not be able to get jobs. And then you have some companies announcing big layoffs and blaming it on AI. In February, Block, formerly known as Square, announced that they were laying off half their workforce, about 4,000 people. And the CEO of that company, Jack Dorsey, straight up said that it wasn't

because the business was struggling, but because of the effectiveness that teams are seeing from using AI tools. And look, Block is not the only one. Amazon cut 16,000 jobs in January, Atlassian cut 10% of its workforce. And then there was a report in Reuters in mid-March that meta is considering laying off 20% of its workforce, which would be about 15,000 jobs. So you have all these tech companies laying off tons of people. And then these kind of layoff announcements can create like a domino effect, causing other companies to lay off their staffs to stay lean and competitive. Now, look, the reality is a lot of these tech companies are probably just course correcting from overhiring following the post-COVID boom. And they're just using AI to kind of

cover their mistake and paint a better story for investors. But we can't just ignore the impact that AI could have on jobs. One of my favorite tech analysts is Ben Thompson. He writes a great newsletter called Stratecoree. And in a recent post, he had a very interesting framing of the situation. He wrote that the positive impact of AI in large organizations won't necessarily be eliminating jobs, but instead, it might be about replacing what he called the hard to manage and motivate human cogs in the organizational machine with AI agents, because these AI agents do what they're told and work 24-7 until the job is done. You know, AI doesn't burn out. AI doesn't ask for raises, and they can potentially get the work done at a faster rate than humans. I guess we'll

have to wait and see. The rising unemployment for younger workers is a bit worrisome, though, I got to say. So what's my take here? Well, I think I'm starting to drink the AI agent hype juice, all right? I've been playing around with some of these tools like OpenClawe and Claude Code. And while I'm still trying to figure them out, even in their current state, you can see the potential. And the thing is, these tools are only going to get better. From an investing perspective, the rise of AI agents is making me more bullish on the AI trade, especially the big tech companies and Nvidia.

The shift in inference is going to require exponentially more compute. So all the AI CAPEX from big tech companies is likely. justified. And with Nvidia leaning into CPUs and inference chips, it's hard not to be bullish. I mean, look, long-time listeners know, I'm an optimistic person. And when you look back at every major technological shift in history, it's created some anxiety about jobs, but the new tech has always led to new industries and new opportunities and overall economic growth. I think the same thing will happen with AI. Now, I'm not going to lie, it does feel like things are moving really, really fast. I feel like every week there's a new tool coming out of Open AI and Anthropic that's

destroying industries. But I also think that's a lot. think that companies tend to move pretty slowly, especially large companies. In fact, according to the consulting firm Gartner, over 40% of businesses' AI agent projects will be canceled by the end of 2027 because of unclear business value or inadequate risk controls and escalating costs. And let's talk about the cost part real quick because not enough people are talking about it. Now, running these AI agents is not cheap. These AI agent tools don't just have like a monthly flat fee costs. It's not like signing up for chat GPD pro. These agents consume token and the more complex the task, the more tokens are used. And that compute bill adds up pretty fast. That's the biggest

thing that I've learned from using these tools myself. I run out of tokens pretty fast. So in the near future, it still might be cheaper having a human do a task compared to an AI agent. But look, there's no doubt that the cost of computing is going lower and there's no doubt that AI is going to change the way that we work. But I'm just not ready to panic over it just yet. In the meantime, I will be holding on to my Nvidia stock as a hedge of potentially losing my job to an AI agent. Well, all right, guys, that's it for today's weekend, deep dive. Hope you guys enjoyed today's episode.

Let me know what you guys think about the rise of AI agents. Are you drinking the hype juice like me? Does all this make you less worried about an AI bubble? Do you think that AI agents will replace a ton of jobs and lead to a big rise in unemployment? Drop your thoughts in the comments on Spotify and YouTube. And while you're at it, if you have an extra five seconds, consider giving us a five-star rating. All that engagement really does help us out,

and it helps other people find the show. Thank you guys so much for listening, watching, and commenting. Shout out to Mike and Connor for all the work behind the scenes. And we'll see you guys back here tomorrow. Rosen lasagna, medium power, 15 minutes. Sounds like Ojo time. Let's play. Feel the fun with Play-O-Jo.

The online casino with all the latest Slashino. lot and live casino games. What you win is yours to keep with no wagering requirements, instant payouts, and no minimum withdraws. Hey, I just won. Woohoo! Feel the fun! Play, oh Joe! Honey, forget about the lasagna. Let's celebrate! 19 plus Ontario only, please play responsibly. Concern about your gambling or that of someone close to you. Call 16-531-2600 or visit conexontera.ca.ca.com.com.