11AM Hour: UFC's Dana White, Tech Takes a Tumble & Buy the Dip in Netflix? 7/17/26

**Carl Quintanilla** (0:06)
Good Friday morning, welcome to Squawk on the Street. I'm Carl Quintanilla with Sara Eisen here at Post 9 of the New York Stock Exchange. Busy session today, a lot of cross currents, tech taking a tumble, though the NASDAQ is off the lows of the day. We'll talk about why Alphabet's putting pressure on the sector, and how a new AI model out of China is stoking investor fears.

**Sara Eisen** (0:23)
Then UFC President and CEO Dana White joins us at Post 9 We'll discuss sports betting, the growth of the USC, and their recent White House event, of course.

**Carl Quintanilla** (0:31)
Netflix shares are plunging on the heels of disappointing profit guidance. We'll talk to one analyst about why she says to buy the weakness here.
As for the markets, well off the session lows, NASDAQ 100 was down 2.7% earlier this morning. Shaved that by more than a full point. Russell definitely outperforming overall equal weight over the cap weighted S&P. Ten year, a little bit of relief as we got the two year down to 411 That'll take you back to July 6th.

**Sara Eisen** (0:57)
Yeah, still pretty ugly in tech world, especially if you're communication services or information technology. The chip names, the hyperscalers, we'll start with that big move in tech. Alphabet weighing on the sector this morning as well. While Apple gets an upgrade, McKenzie Segalos has more. McKenzie.

**Mackenzie Sigalos** (1:13)
Sara, this is really a tale of two mag seven names that have reversed their trading positions in just two months, based on really opposite bets about how to win in AI.
Let's start with Alphabet, which is dragging down the triple Q this morning. Google's next flagship Gemini Pro model is reportedly months behind schedule and still falling short in coding. And that comes as China's moonshot AI rolls out Kimi K3, a lower cost model already outperforming Anthropix Claude Fable 5
I've spoken with people familiar with operations at Google DeepMind, and a consistent message is that this is not a lack of talent, money or compute, it's an execution problem. My sources describe strong researchers trapped inside a slow, highly political organization with overlapping projects, too many people involved in decisions, and multiple product teams weighing in before anything ships. Apple, meanwhile, is increasingly winning as the anti-CapEx AI trade. HSBC upgraded the stock to buy this morning and raised its target to $366, arguing that Apple can monetize AI across 2.5 billion active devices without funding the massive infrastructure buildout itself. Apple spending 2.5 percent of sales on CapEx versus 39 percent for hyperscalers like Google. That lighter spending model lets Apple lean on Google's models in cloud capacity, essentially using Gemini to drive a new hardware replacement cycle and faster services growth, all while keeping the customer relationship and preserving cash for buybacks. But Apple, it is still fighting its own AI talent war. The FT reporting that they've sent legal warnings to 40 former employees now at OpenAI as it builds its trade secrets case. But the market doesn't appear concerned by Apple's AI talent exits, instead really rewarding them for controlling distribution without carrying Alphabet's cat-bix burden, hence this record run in Apple shares. Guys?

**Carl Quintanilla** (3:03)
Meantime, Mac, there's this piece in the FT about Apple reportedly sending legal letters to former employees who are now at OpenAI, another signal that they are taking this IP issue pretty seriously.

**Mackenzie Sigalos** (3:15)
They are, and in those letters, according to FT's reporting, it's directing them to preserve documents and asking that they meet with Apple's lawyers, really as it builds up this trade secret case. But what I will say, Carl, is that Apple frequently sues former employees, especially over accusations that they've downloaded Apple files on their way out the door. Apple's able to see that because it controls employee devices. And when you look across the last decade, Apple's referred at least three former employees who allegedly took files to the FBI for criminal charges. They sued an employee, went to Chinese rival Oppo. So this isn't out of character for them. And we're talking about, at least according to that lawsuit, 400 Apple employees now working at OpenAI.

**Carl Quintanilla** (3:56)
Keeping an eye on that and on all of Sam Altman's tweets lately. Mac, thanks. I'm McKenzie Segalos. Also in the tech space today, China's Moonshot AI, unveiling its latest AI model that ramps up competition with OpenAI and Anthropic. Our Kate Rooney watching that one and answer the question. Maybe, Kate, if we're in this Kimi moment as opposed to a deep seek moment.

**Kate Rooney** (4:16)
Yeah, exactly.
Yeah, similar maybe to the deep seek moment, Carl. Yeah, the model out of China that we're talking about has been causing new jitters in the tech space. Chinese startup Moonshot behind this one, it claims to be closing the gap right now with some of the leading US frontier labs. It says that it surpasses versions of OpenAI and Anthropics models on some benchmarks, but it still does trail the very best versions on overall performance. Bank of America, talking about this in a note this morning, calls it a step-change gain for flagship Chinese models, raises the capability ceiling as they put it for China. Wells Fargo also chiming in on this, calls it an incredible feat, talks about stocks having their next deep seek moment today. That new jerk reaction we're talking about, they say, is questioning whether all of this capex is worth it when China is just replicating all of this for a fraction of the cost. The news is weighing more on the picks and shovels in AI. Take a look at Nvidia and some of the chip names as evidence of that. It is a more muted reaction though than what we did see with deep seek. The threat from Chinese models does seem to be a bit more priced in at this point. It's less of a surprise. The bare case when we talk about this has been, really, if you build these AI models on the cheap or if that's possible, why go out and spend the hundreds of billions of dollars on compute that we're doing here in the US? Kimmy also brings up new questions about open weight models and if companies here in the US would move towards more of a build it yourself ethos. The bull case, though, is that models don't always tend to see widespread adoption, it's probably too soon to tell. Cheaper AI can often mean more AI if you've heard of Jevons Paradox, that's sort of the argument there. Plus the AI labs are wising up to this, they are getting a lot more efficient as well. Wells Fargo also did pour some cold water on this being any sort of a cost breakthrough points out that cost per intelligent task is just under a dollar and that is only slightly cheaper when you look at OpenAI's ChatGPT 5.5. This is also a geopolitics discussion as well. Former White House AIs are David Sachs. Talking about this on Twitter, X this morning says, the performance of Kimi K3, that model, is concerning. He also criticizes some of the growing calls for AI regulation here in the US. He says, this is how you lose the AI race. The rest of the world won't play by our rules if we bog ourselves down, guys.