The AI Race Gets a Massive Power Shift

This podcast is sponsored by Google. Hey folks, I'm Amar, product and design lead at Google DeepMind. Have you ever wanted to build an app for yourself, your friends, or finally launch that side project you've been dreaming about? Now you can bring any idea to life, no coding background required, with Gemini 3 in Google AI Studio. It's called Vibe Coding and we're making it dead simple. Just describe your app and Gemini will wire up the right models for you so you can focus on your creative vision. Head to ai.studio slash build to create your first app. Today on the AI Daily Brief, a major shift in the global AI race, and before that in the headlines, Google readies its AI glasses for next year. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, Gemini, KPMG, Robots and Pencils, Blitzy, and Rovo. To get an ad-free version of the show, which starts at just $3 a month, go to patreon.com slash aidailybrief, or you can subscribe on Apple Podcasts. If you're interested in sponsoring the show, you can send us a note at sponsors at AIDailybrief.ai. Now, we are very much in the GPT 5.2 waiting room right now. Anticipating the model will come out perhaps even before you're listening to this episode. And the last thing that I wanted to share before we dive in is that for the next two weeks, I'll be running an experiment bringing back an AIDailybrief newsletter. We're going to try an evening publication that basically lets you know what we're going to cover the following day, kind of a tomorrow's news today type of vibe. We'll see how people respond to this and whether we continue it into next year or try something different. For now, if you want to check it out, go to AIDailyBrief.ai. You can click newsletter and there'll be a link to subscribe. With that, let's dive in. Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. Google is preparing to launch a full range of AI-enabled smart glasses next year. The product line will include audio-only glasses as well as a pair with an in-lens display that can show information like directions and translation. In addition, Google unveiled a prototype called Project Aura, which will offer a full-screen experience similar to Apple's Vision Pro. Unlike the Vision Pro, Project Aura is still packaged in a relatively lightweight glasses format rather than a full headset. While Google said that the product line will begin rolling out from next year, they did not specify which models would be available. The smart glasses will run Android XR, which is Google's new operating system for smart glasses. Samsung actually already released the first Android XR product over the summer with their Galaxy XR headset. And apps that have already been developed for that platform can be installed across Google's product line as well. The platform is also cross-compatible with Android smartphone apps, so the new devices can integrate into the existing Google ecosystem. The glasses can even connect to iOS apps, so you can tether the glasses to an iPhone and access the Gemini app as well as Google's full suite of apps. Google's director of product management for XR, Justin Payne, said, The goal is to give this ability to have multimodal Gemini in your glasses to as many people as possible. If you're an iPhone user and you have the Gemini app on your phone, great news. You're going to get the full Gemini experience there. Now, of course, the elephant in the room is Google Glass, a tech flop that was so big it has almost become iconic. Now, of course, Google Glass was more than a decade ago. And this time, there's no avant-garde sci-fi design, just a normal-looking pair of sunglasses with a thicker frame. In fact, they're actually tapping people who know how to make cool-looking glasses to help. The glasses will actually be designed in conjunction with Gentle Monster and Warby Parker. Now, among all the AI wearables, Meta's Ray-Bans are basically the one form factor that people have actually enjoyed and that has been a success. So it's not at all surprising to see Google jump into this space. So is it possible that glasses will be the first ubiquitous wearable form factor? I guess that is something that we will have to wait till next year to know. But moving back to the present, Claude Code is coming to Slack and people seem pretty excited about it. Building on the existing Cloud integration, users can now spin up an entire coding session within Slack to work on bug fixes and feature requests directly. Now, of course, if you have heard me harp on how 2026 is the year of context, the big benefit of using Cloud directly in Slack is the ability to tap into the context of work conversations, and the Cloud Code integration takes that idea to the next level. TechCrunch actually suggested that this is evidence of a broader shift, writing, AI coding assistants are migrating from IDEs into collaboration tools where teams already work. While the integration is still just a research preview, it does suggest that Anthropic is looking for a way to cut down on context switching and build their agents directly into the platforms where they're most useful. In a quick follow-up from Monday's show, Apple's chip leader says rumors of his departure were greatly exaggerated. Apple's senior VP of hardware technologies, Johnny Shruji, told staff on Monday that he's not going anywhere for the time being. In a memo to his division, he wrote, Bloomberg's Apple insider Mark Gurman had reported that Truji had recently told CEO Tim Cook that he was looking to move on. This was potentially a huge problem for Apple as the M-class chips pioneered under his leadership are absolutely class-leading. After a spate of executive departures, it is good for Apple that this one doesn't appear to be happening, although it seems like there's probably still more going on behind closed doors. Lastly today, IBM is making a big $11 billion acquisition to boost their AI business. IBM is committed to buyout public data infrastructure firm Confluent. The deal was struck at $31 a share, a 34% premium to their last price before the announcement. The offer is all cash and builds on IBM's five-year partnership with Confluent. Confluent provides an open-source data streaming platform, giving enterprises the ability to connect and govern their data in a way that's more accessible to AI agents. Earlier in the year, Salesforce and ServiceNow made similar acquisitions to fill their respective gaps in the AI data layer. By way of example, TireMaker Michelin is using Confluent's platform to make their real-time inventory and supply chain data discoverable by AI. Instacart has adopted Confluent to drive their fraud detection and product database systems. For IBM, the acquisition marks a commitment to building out their high AI product In a statement CEO Arvind Krishna said IBM and Confluent together will enable enterprises to deploy generative and agentic AI better and faster With the acquisition of Confluent IBM will provide the smart data platform for enterprise IT purpose for AI Analysts at Bloomberg Intelligence said that the deal could, quote, significantly improve IBM's AI portfolio and subsequently its software unit sales growth. The Monster deal caps off a record-breaking year in AI M&A, where over a trillion dollars worth of deals have closed this year. For now, though, that is going to do it for the headlines. Next up, the main episode. All right, let's talk about the signal versus the noise in enterprise AI. The challenge right now isn't just about what's possible, it's about what's practical. That's the entire focus of the You Can With AI podcast I host for KPMG. Season one, cut through the hype to focus on deployment and responsible scaling. Season two goes a level deeper. We're bringing together panels of AI builders, clients, and KPMG leaders to debate the strategic questions that will define what's next for AI in the enterprise. Six episodes packed with frameworks you can actually use. Find You Can with AI wherever you get your podcasts. Subscribe now so you don't miss the new season. Today's episode is brought to you by Robots and Pencils. When competitive advantage lasts mere moments, speed to value wins the AI race. While big consultancies bury progress under layers of process, Robots and Pencils builds impact at AI speed. They partner with clients to enhance human potential through AI, modernizing apps, strengthening data pipelines, and accelerating cloud transformation. With AWS-certified teams across US, Canada, Europe, and Latin America, clients get local expertise and global scale. And with a laser focus on real outcomes, their solutions help organizers work smarter and serve customers better. They're your nimble, high-service alternative to big integrators. Turn your AI vision into value fast. Stay ahead with a partner built for progress. Partner with Robots and Pencils at robotsandpencils.com slash AI Daily Brief. delivers 80% plus of the development work autonomously, while providing a guide for the final 20% of human development work required to complete the sprint. Public companies are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding pilot of choice to bring an AI-native SDLC into their org. Visit blitzy.com and press get a demo to learn how Blitzy transforms your SDLC from AI-assisted to AI-native. Meet Rovo, your AI-powered teammate. Robo unleashes the potential of your team with AI-powered search, chat, and agents, or build your own agent with Studio. Robo is powered by your organization's knowledge and lives on Atlassian's trusted and secure platform, so it's always working in the context of your work. Connect Robo to your favorite SaaS app so no knowledge gets left behind. Robo runs on the Teamwork Graph, Atlassian's intelligence layer that unifies data across all of your apps and delivers personalized AI insights from day one. rovo is already built into jira confluence and jira service management standard premium and enterprise subscriptions know the feeling when ai turns from tool to teammate if you rovo you know discover rovo your new ai teammate powered by atlassian get started at rov as in victory o.com one of the things that i've learned about the ai daily brief audience is that as much as you guys are interested generally in AI and what it means for the world, you're definitely more interested on average in what AI means for you and your careers and your companies. I think this is a totally reasonable stance to take, and it's why I try to slant the coverage towards news that is practical and relevant. For example, today I'm gearing up for the anticipated GBT 5.2 announcement. However, I do think it's important to have a broad-based understanding of how this technology is intersecting with the wider world, particularly in the realm of geopolitics, and today we got one of our most significant shifts ever. The headline, which by now I'm sure you've heard, is that President Donald Trump has said that the U.S. will now allow NVIDIA to sell its H-200 chips to companies in China, taking a 25% cut along the way. Now, to understand this, I think you have to go back almost a decade to get the trajectory of U.S.-China policy specifically as it relates to chips. For pretty much the entirety of the first Trump presidency as well as the Biden administration, China hawks led the policy discourse. Indeed, it was one of the few things that was fairly bipartisan. I remember back in the 2020 run-up, there was a big debate on who would be tougher on China, Joe Biden or Donald Trump. During the Biden administration, we got the CHIPS Act, which was of course in part a response to our COVID-era recognition of what many perceived as our over-reliance on Chinese-involved supply chains. And we got this big effort to reshore domestic manufacturing across a wide variety of industries, but with a particular focus on advanced chips. Throughout the Biden administration, there were not just these incentives to move the chip manufacturing industry back to the U.S., but also more restrictions on exports. Indeed, right at the end, as President Biden was getting prepared to vacate the White House, they dropped an executive order, which became known as the AI Diffusion Rule. The Diffusion Rule created three tiers of licensing requirements that basically reduced the amount of chips that we were exporting everywhere. Now, as you might imagine, the world's leader in advanced AI chips, NVIDIA, was no fan of that and started to articulate basically what their position would be during the upcoming Trump administration as well. Outside of the folks who think that, of course, their main position is, hey, we'd like to sell our product into one of the world's biggest markets. Basically, their position is that the best way to handle China is not to cut them off, but to manage them. That effectively, the U.S. wants its infrastructure to be used everywhere. Now, almost immediately as the new administration came online, the Department of Commerce announced the rescission of the AI diffusion rule, but that didn't mean that NVIDIA was in the clear. Throughout the year it has just been a total back and forth on what long US policy was going to be regarding selling chips to China Although functionally NVIDIA has simply not been selling into China even their neutered H20 chips that were designed specifically to meet export requirements When NVIDIA has given forward guidance for the last several quarters, they've basically left China off entirely. And yet, clearly in the recent weeks, things have been shifting. At the end of October, Semaphore editor-in-chief Ben Smith wrote a piece called Trump is Poised to End Washington's Decade of the China Hawks. In it, he wrote that we're seeing the, quote, end of an anti-China moment in Washington that really began with Donald Trump's election in 2016, intensified through the Biden presidency, went viral on Capitol Hill and crested on Liberation Day. The piece concluded, on the defining question of China, Trump seems to be working his way back even from a confrontation that he began and Biden escalated towards something a little more like business as usual, deferring to the old Calvin Coolidge observation that the business of America is business. Now, it was Semaphore once again who reported on Monday morning, that the Commerce Department was poised to approve the export of NVIDIA H200s for the first time. Last week, Commerce Secretary Howard Lutnick had said that the decision was in Trump's hands, but didn't give his opinion on the matter. Some of four sources, however, said that Lutnick had been supportive of the plan to loosen export restrictions to allow in the powerful chips. The decision also came in the context of a Washington visit from NVIDIA CEO Jensen Huang last week. Huang met with the president as well as congressional leaders, and at the time, the big story was that he had successfully managed to negotiate the removal of tighter restrictions that were planning on being inserted into the National Defense Authorization Act. It now appears that he's successfully negotiated for much, much more. Yesterday, the president tweeted, I have informed President Xi of China that the United States will allow NVIDIA to ship its H200 products to approved customers in China under conditions that allow for continued strong national security. President Xi responded positively. 25% will be paid to the United States of America. Now, Trump had floated the idea of the government taking a cut of NVIDIA's Chinese revenues over the summer when the issue was last up for debate. At the time, they were talking about 15%, so this 25% is, I guess, the art of the deal. Now, Trump framed the move as a policy win for the U.S. manufacturing sector, specifically taking aim on the policy of neutered versions of technology to match export controls. In that same Truth Social post, he continued, the Biden administration forced our great companies to spend billions of dollars building degraded products that nobody wanted, a terrible idea that slowed innovation and hurt the American worker. That era is over. Now, to the extent that there was a line that was drawn around NVIDIA's most advanced chips, commenting, NVIDIA's U.S. customers are already moving forward with their advanced Blackwell chips and soon Rubin, neither of which are part of the deal. Trump also said that the same approach would be applied to AMD, Intel, and other U.S. chip makers. Now, as you might imagine, the announcement triggered a huge response. Democrats immediately got up in this, basically saying it was one more example of Trump's political corruption. This attitude was summed up by Elizabeth Warren, who said, The Republican response has been a little bit more quieted, but that's a complicated story that we'll come back to in just a moment. Mostly folks tried to explore what it would mean for the U.S. with regard to competition with China in the AI race. David Shore, the head of data science at Blue Rose Research, wrote, We're on track to reduce our compute advantage over China from 33 to 1 to 1.2 to 1 because NVIDIA and Jensen Huang put millions of dollars in Trump's pocket in order to enrich themselves at the rest of our expense. Alex Staff, the co-founder of think tank IFP, wrote, massive own goal to export these AI chips to China. The H200 is six times more powerful than the H20, which was previously the most powerful chip approved for export. Our compute advantage is the main thing keeping us ahead of China in AI. Why would we throw that away? Georgetown professor Rush Doshi writes, this is a big deal, essentially a reversal of the US export control policy on advanced chips, possibly decisive in the AI race. Compute is our main advantage. China has more power, engineers, and the entire edge layer. So by giving this up, we increase the odds the world runs on Chinese AI. Others noted that the policy coming out of this White House is not exactly clear. Hudson Institute fellow Michael Sobolik wrote, the same day the admin improved sales of H200 chips to China, the DOJ says this about the exact same chips, quote, these chips are the building blocks of AI superiority and are integral to modern military applications. The country that controls these chips will control AI technology. The country that controls AI technology will control the future. We are sabotaging ourselves to make a buck. If this is the posture we're taking, we're asking for defeat. NVIDIA stock, of course, jumped on the news, racing up by about 2% in late trading. Now, while it's clear that the policy change could dramatically shift the balance of power in AI, it's not quite as simple as China getting a boost at the expense of the US. The Institute for Progress put out a report on just how much the export of H-200s would change the capabilities of China's GPU fleet. They modeled that under current policy settings, China would have just 2% of the AI compute produced next year. Now, with conservative exports of the H-200, they modeled a 12% share, with a 30% share possible under a more aggressive set of assumptions. Tim Fist, the Director of Emerging Technologies at the IFP, wrote, Huawei is not planning to make an AI chip that matches the H-200 until Q4 2027. But more importantly, China's manufacturing bottlenecks mean it'll only reach 1-2% of U.S. production in 2026. So this move is giving China a bunch of advanced AI compute it wouldn't otherwise have. However, he also recognized that there's a trade-off here, continuing, The argument for exporting H-200s is to slow China's AI chip industry. But the CCP recognizes its strategic vulnerability. It's subsidizing industry and creating demand for Huawei chips via procurement mandates and restrictions. U.S. export policy is unlikely to change this. Instead, he writes, H200 exports will counterfactually boost China's ability to both develop frontier AI models and deploy them widely, competing with USAI and cloud companies abroad. The new Chinese stack will be NVIDIA chips, Tencent, Baidu, Alibaba cloud, and DeepSea, Quen, and Kimi models. Milkroad tried to summarize the two large camps when it comes to this question. They wrote, the smart strategy camp argues that this is the best way to manage China's rise without losing money Since the H200 chip is now about 13 months old selling it to China is seen as selling them previous generation technology while the US keeps the brand new significantly faster Blackwell chips for itself The goal is to get China addicted to buying Commerce Secretary Howard Lutnick did go so far to say You want to sell the Chinese enough that their developers get addicted to the American technology stack. Going back to Milk Road, they write, This keeps them dependent on the U.S. supply chain and sends billions of dollars back to American companies like NVIDIA, which helps fuel the U.S. economy. However, the Dangerous Mistake Camp warns that this logic is flawed because the H200 is still an absolute beast of a machine. It isn't just a little better than what China currently has. It offers six times the performance of the chips China is currently allowed to buy. By selling these chips, the U.S. is essentially breaking its own safety rules and handing over hardware that is far more capable than the label of old technology suggests. This is viewed as a massive bailout for China because their own tech giant Huawei has admitted it won't be able to build a chip this good until late 2027. Right now, China is struggling to manufacture high-end chips at scale. But if they can just buy the H200 from NVIDIA, they get to skip two years of hard work and struggle. This fast-forwards their military and AI capabilities instantly. The fear is that the U.S. is trading long-term national security for short-term stock market gains, giving a rival the exact tools they need to catch up just because the chip is technically a year old. So what does China do from here? China commentator and author of Red Roulette wrote that Beijing's priorities will shift almost overnight. He argues that Beijing will read Trump as fundamentally transactional, meaning that everything is negotiable. He writes, expect a full-court press in the coming months to secure the most advanced technology available across the spectrum. Not just GPUs, but specialty materials, EDA software, and semiconductor manufacturing equipment that seemed untouchable only a day ago. Second, he argues, none of this will slow China's drive for semiconductor self-reliance, not even by a millimeter. China's long-term mission, he writes, remains unchanged. Eliminate dependence on foreign chips. Access to more advanced U.S. hardware doesn't dilute that goal. Historically, it has sped it up. Seemingly in line with that second point, the Financial Times reported this morning that the Chinese government was planning to limit access to NVIDIA's H200 chips even though they were now approved for export. The FT sources suggest that buyers of the H200 will be required to go through an approval process, submitting requests, and also in the same process explaining why they weren't able to use Huawei or another domestic manufacturer to meet their needs. Now, in addition to the Chinese government slowing down these plans, there also is legislation moving through Congress that would prevent the White House from approving the export of chips, including the H-200, to China for 30 months. Which brings us to our second policy story of the day. One of the things that we've been tracking here is the White House's push for federal policy over state policy when it comes to AI. A few weeks ago, we started to get reports that not only was the White House thinking about releasing an executive order that would set up a commission to sue states that pursued their own AI regulations, but that this was causing some serious consternation within the Republican Party. After the EO was dropped because of that pushback, it seemed the next venue for trying to get this through was to try to add an amendment to the must-pass National Defense Authorization Act that would once again block state-created AI laws. While we haven't gotten the executive order yet, but also on Monday on Truth Social, Trump tweeted this, There must be only one rulebook if we are going to continue to lead in AI. We are beating all countries at this point in the race, but that won't last long if we're going have 50 states, many of them bad actors, involved in rules and the approval process. There can be no doubt about this. AI will be destroyed in its infancy. I will be doing a one-rule executive order this week. You can't expect a company to get 50 approvals every time they want to do something. That will never work. AI czar David Sachs flushed out the concept in a longer post on X. He tried to make it a lot more political, with about half of the post raising the boogeyman of what the White House has effectively called woke AI before. Republicans continue to not be fully aligned around this. Responding to a tweet that said, I hope Florida plans to challenge Trump's new EO he's going to sign this week, Florida Governor Ron DeSantis wrote, an executive order doesn't and can't preempt state legislative action. Congress could theoretically preempt states through legislation. The problem is that Congress hasn't proposed any coherent regulatory scheme, but instead just wants to block states from doing anything for 10 years, which would be an AI amnesty. I doubt Congress has the votes to pass this because it is so unpopular with the public. Showing just how weird the bedfellows are getting, AI safety voice Max Tegmark quote tweeted DeSantis and said, I agree. It's pretty clear that what the AI lobbyists want is one rule that's de facto no rule. Let's stop this corporate welfare by treating AI companies just like we treat other powerful industries and not leaving them less regulated than sandwich shops. Now, it is way beyond the scope of this particular show to get into the merits of the state versus federal framework. Turns out that's been a conversation that we've been having for the roughly 260 years since we became a country. But what is interesting to me in this particular case is not even just the policy, but the politics. There is a growing sense that Trump is making a major bet here that is not necessarily in sync with his own party. Indeed, Axios wrote a piece called Trump Bets Party and Presidency on AI. The way that they describe the political risk is this. Trump is flooring the gas pedal at the very moment some of his most ardent MAGA backers are warning AI could destroy the working-class Americans who brought him to power. The fear is that AI and AI-powered robots will eat vital American jobs before the nation has time to prepare the U.S. workforce for sci-fi-level change. The bottom line, they write, If White House AI and Cryptozar, David Sachs and others are right that AI juices economic growth and new jobs, Republicans will likely prosper. But if they're wrong or the benefits come after a few years of pain, it could be politically catastrophic. Now, I promise you guys that this is not going to become a political podcast, even as we head into the election cycle next year. However, because of that election cycle, I do think what's going on in the political world and the policy world could have fairly dramatic impact on the technology that we're all using. And so every once in a while, I will pop over to cover that, despite the focus continuing to be on the more practical and business-oriented side of AI. For now, that is going to do it for today's AI Daily Brief. Appreciate you listening or watching. As always, until next time, peace. you