The AI Daily Brief: Artificial Intelligence News and Analysis - Is OpenAI Becoming Too Big to Fail?

Starting point is 00:00:00 This podcast is sponsored by Google. Hey folks, I'm Amar, product and design lead at Google DeepMind. We just launched a revamped vibe coding experience in AI Studio that lets you mix and match AI capabilities to turn your ideas into reality faster than ever. Just describe your app and Gemini will automatically wire up the right models and APIs for you.

Starting point is 00:00:21 And if you need a spark, hit I'm feeling lucky and we'll help you get started. Head to AI.studio slash build to create your first app. Today on the AI Daily Brief, as OpenAI announces yet another big deal this time with Amazon, are they becoming too big to fail? Before that in the headlines, Coca-Cola comes back to the well with another AI-generated Christmas ad. 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, Rovo, robots and pencils, Blitzy and KPMG. To get an ad-free version of the show, go to patreon.com slash AI Daily Brief.

Starting point is 00:01:06 and for information about sponsorship to get a link to our AI-R-OI benchmarking study, which is ROISurvey.aI.i, or to learn about jobs or anything else regarding the show, go to AIDaily Brief.com. Welcome back to the AI Daily Brief Headlines edition, all the daily AI news you need in around five minutes. There is a grand tradition of Christmas time advertising. In the UK, you have John Lewis and Sainsbury's, whose Christmas short film ads have become treasured holiday traditions, and then, of course, there is Coca-Cola, a company who, whose advertising imagery is integrally and historically linked to our modern perception of Christmas, which, by the way, is the thing I did a whole podcast on that I can dig up if anyone is interested.

Starting point is 00:01:46 In any case, in the last couple of years, Coca-Cola's Christmas advertising has also become a front in the AI culture war. In 2024, for the first time, Coca-Cola aired an AI-generated version of their iconic holiday commercial. At the time, it made impressive use of stable diffusion, but looking back a year later, it's certainly very of its time. Faces are wobbly, everything has an AI sheen, the lighting is uneven. It was impressive, but impressive, based on the time period. And it was not hard to find strong negative reaction. One person commented, The world is so over if the Christmas Coca-Cola advert is made with AI.

Starting point is 00:02:22 This year's version shows pretty significant improvements. The weather is more realistic, the lighting is better. Once again, the ad was produced by a studio called Secret Layer. Founder Jason Zada commented, Last year was a cultural milestone, and this year through animated. characters, we bring magic across the globe with the arrival of the Coca-Cola trucks. Now, this year, you can tell they sort of hedged. Rather than facing the criticism that they could

Starting point is 00:02:43 have hired humans before AI generated humans, which was a part of last year's commercial, this year it was all anthropomorphic animals. The only human shown in this ad was Santa Claus, and in order to preemptively combat some of those arguments, Coca-Cola pointed out that they hired performers to sing the music for the videos. Coca-Cola's head of Gen AI, Pertique the car, said, were committed to using AI as a human enabler where it makes sense. creative ambition, direction, and thought leadership has and always will be human-led. AI is a superpower when it comes to execution and production, making what was previously impossible possible.

Starting point is 00:03:14 Now, on the one hand, there is still a lot of negative critique. The Verge calls it a sloppy eyesore and writes, the holidays are coming to cheapen your seasonal nostalgia. But there is a bit more of a diverse perspective as well. One commenter on YouTube said, just because this ad was generated using AI doesn't make it unoriginal or slopped. This ad was concepted, ideated, scripted, and produced by a team of creative professionals using the latest tool in their box, Gen A.I.

Starting point is 00:03:37 Process would be the same with 3D animators and after effects, but Gen AI allowed the creators to ideate in concept faster and eventually create something new and memorable. Embrace this as just another way to express yourself. And for some, it's already past the point where it matters how it was created or not. Said one commenter, you can finally say it's Christmas when the Coca-Cola advert comes on.

Starting point is 00:03:58 Next up, if you were on Twitter at all over the last couple of days, you might have heard that ChatchipT had banned giving legal and health advice. The whole panic was triggered by betting market Kalshi tweeting, Just in, chat GPT will no longer provide health or legal advice. Post was referring to a policy change that went into effect on Wednesday. It included a list of prohibited uses for chatch TBT, including the provision of tailored advice that requires a license such as legal or medical advice without approved involvement by a licensed professional.

Starting point is 00:04:25 OpenAI's head of health AI quickly stepped in to add the relevant nuance with Karen Singhal writing, not true. Despite speculation, this is not a new change to our terms. Model behavior remains unchanged. ChatGPT has never been a substitute for professional advice, but it will continue to be a great resource to help people understand legal and health information. It turns out that change was simply neatening up the list of prohibitions, with OpenAI consolidating three separate documents into a single list. And while this particular fight may have been ultimately kind of a nothing burger,

Starting point is 00:04:53 expect to see a lot more debates like this that aren't just about an individual company, but about what consumers in society decides is and isn't an inappropriate use of LLMs. Now, moving over to geopolitics, during a taped interview with 60 minutes, President Trump has said that China will not get access to NVIDIA's latest Blackwell chips. He said, we will let them deal with NVIDIA, but not in terms of the most advanced, we will not let anybody have them other than the United States. Trump echoed those comments on Air Force One on Sunday commenting, we don't give that chip to other people. Now, the reason the people were paying attention to this is that Trump had said that Blackwells would be on the table heading into trade negotiations last week, but stated after the meeting that the Blackwells weren't discussed. The Wall Street Journal reports that administration figures put up a united front to deny Blackwells to the Chinese. According to the report, Secretary of State Marco Rubio, Commerce Secretary Howard Lutnik, and U.S. Trade Representative Jameson Greer, were among those who opposed a deal on latest generation chips.

Starting point is 00:05:46 Frankly, even if NVIDIA was allowed to deal with Chinese firms, there's little sign at this point that Beijing will allow the chips into the country. Speaking at a developer conference in South Korea last week, NVIDIA CEO Jensen Wang said they made it very clear that they don't want NVIDIA to be there right now. Interestingly, then, Microsoft has obtained an export license to send Blackwell chips to the United Arab Emirates. The U.S. Commerce Department reportedly approved a shipment of 60,000 chips, including the latest generation GV300 Blackwells. Microsoft President Brad Smith said they were the first company to receive a license to bring the advanced chips to the Middle East. He commented, you cannot get those export licenses unless you're able to meet the requirements that have been imposed by the U.S. government. We've earned it by satisfying very stringent cybersecurity, physical security, and other security requirements.

Starting point is 00:06:28 Now, this has been an ongoing conversation. The Gulf states are in a literal in between, both geographically and politically, between the U.S. and China. Over the last couple of years, as AI has become more and more of a flashpoint, the U.S. government has put ever-increasing pressure on the Gulf states to turn towards us and away from Beijing. Brad Smith remarked that this deal could be a linchpin for AI diplomacy in the global south. He said, we run a risk that AI diffusion will become increasingly uneven. There obviously is a race between the U.S. and China. People often focus, first and foremost on the race for advanced AI model development. But I think the AI diffusion race is probably even more important than the race on the technology frontier. And this is where the stronger

Starting point is 00:07:05 relationship between the United States and the United Arab Emirates becomes critical. Will this be a one-off for the first of many? Remains to be seen, but for now, that's going to do it for the headlines. Next up, the main episode. Meet Rovo, your AI-powered teammate. Rovo unleashes the potential of your team with AI-powered search, chat, and agents, or build your own agent with Studio. Rovo 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.

Starting point is 00:07:44 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. Robo 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. 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

Starting point is 00:08:22 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 U.S., 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 robots and pencils.com slash AI Daily Brief.

Starting point is 00:08:58 This episode is brought to you by Blitzy, the Enterprise Autonomous Software Development Platform with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand enterprise-scale code bases with millions of lines of code. Enterprise engineering leaders start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and pre-compiles code for each task. Blitzy delivers 80% plus of the development work autonomously, while providing a guide for the

Starting point is 00:09:26 final 20% of human development work required to complete the sprint. Public companies are achieving a 5x engineering velocity increase when incorporating Blitzie as their pre-IDE development tool, pairing it with their coding pilot of choice to bring an AI-native SDLC into their org. Visit blitzie.com and press get a demo to learn how Blitzy transforms your SDLC from AI-assisted to AI Native. What if AI wasn't just a buzzword, but a business imperative? On You Can with AI, we take you inside the boardrooms and strategy sessions at the the world's most forward-thinking enterprises. Hosted by me, Nathaniel Wittamore, and powered by KPMG,

Starting point is 00:10:02 this seven-part series delivers real-world insights from leaders who are scaling AI with purpose, from aligning culture and leadership to building trust, data readiness, and deploying AI agents. Whether you're a C-suite executive, strategist, or innovator, this podcast is your front-row seat to the future of enterprise AI. So go check it out at www.kpmG.org-us-slash-a-i podcasts, or search you Penn with AI on Spotify, Apple Podcast, or wherever you get your podcasts. Welcome back to the AI Daily Brief.

Starting point is 00:10:37 OpenAI has announced another mega deal, and with all of these things combined, some people are starting to ask, is OpenAI too big to fail? Today we are going to get into all of the latest in the bubble talk, starting with the latest deal with OpenAI, which is with Amazon. Amazon CEO Andy Jassy tweeted on Monday, New multi-year strategic partnership with OpenAI will provide our industry leading infrastructure for them to run and scale chat GPT inference, training, and agentic AI workloads. Allows OpenAI to leverage our unusual experience running large-scale AI infrastructure,

Starting point is 00:11:08 securely, reliably, and at scale. OpenAI will start using AWS infrastructure immediately, and we expect to have all of the capacity deployed before the end of the next year, with the ability to expand in 2027 and beyond. The specifics of the deal are that in this first incarnation, it represents a $38 billion commitment. The deal will give them access to AWS compute, quote, imprising hundreds of thousands of state-of-the-art Nvidia GPUs,

Starting point is 00:11:32 with the ability to expand to tens of millions of CPUs to rapidly scale agentic workloads. Now, most of the announcement reads as an advertisement for AWS compute products, although it is notable that it's Nvidia GPUs, not Amazon's own tranium chips that are part of the deal. Now, for Amazon, this was a reminder to the market that despite some of their cloud peers getting more press recently, AWS is the cloud giant, and they are going to have a very big footprint in this space. For Sam Altman, it feels very much like just another announcement in the never-enance. quest for more compute. In understated fashion, he tweeted, very pleased to be working with Amazon to bring a lot more Nvidia chips online for OpenAI to

Starting point is 00:12:06 keep scaling. Now, many remarked that this came hot on the heels of OpenAI's final deal with Microsoft, which allowed them to diversify their cloud providers. Amit is investing pointed out the obvious saying, looks like OpenAI is diversifying their AI workloads by partnering with every cloud. Let's go. Now, Market certainly liked it. Amazon surged over 6% after announcing the deal. And I think realistically, if you were trying to look for some big implication in terms of who is working with who, the reality is when it comes to compute and AI workloads at scale, everyone is just going to work with everyone. People tried to make it a big deal when Anthropic signed a new deal with Google, leading some to wonder if that was

Starting point is 00:12:41 a slight to Amazon. The reality, I think, is just every big AI model company is going to look for compute wherever they can get it. Now, beyond the specific implications for Amazon, the biggest part of the conversation was around the implications of OpenAI's dealmaking. The Kobayisi letter went viral with a tweet where they said, OpenAI has now one, signed a $500 billion Stargate deal, two, signed a $100 billion Nvidia deal, three, sign a $100 billion AMD deal, four, signed a $38 billion

Starting point is 00:13:06 Amazon deal, five, signed a $25 billion Intel deal, six, signed a $20 billion $2,000 TSM deal, seven, signed a $13 billion Microsoft deal, eight signed a $10 billion dollar Oracle deal, nine signed a multi-billion dollar broadcom deal, 10 launched a browser to compete with Chrome, 11, become the world's most valuable private company, 12 considered a $1 trillion

Starting point is 00:13:24 IPO by 2027. We are in the midst of a generational technological revolution. And yet for some, the question is, how could Open AI possibly live up to all these deals? On a recent podcast, investor Brad Gersner, who by the way is an investor in OpenAI, asked Sam Altman how a company that had just $13 billion in revenue could afford this $1.4 trillion in commitments. Altman's response has been dominating the conversation for the past several days. It's really worth a listen in its own right. You know, how can the company with 13 billion in revenues make 1.4 trillion of spend commitments, you know, and you've heard the criticism, Sam?

Starting point is 00:14:03 We're doing well more revenue than that. Second of all, Brad, if you want to sell your shares, I'll find you a buyer. I just enough. Like, you know, people are, I think there's a lot of people who would love to buy opening eye shares. I don't think you want to sell. Including myself. Including myself. people who talk with a lot of like breathless concern about our compute stuff or whatever that

Starting point is 00:14:25 would be thrilled to buy shares. So I think we could sell, you know, your shares or anybody else's to some of the people who are making the most noise on Twitter, whatever, about this very quickly. We do plan for revenue to grow steeply. Revenue is growing steeply. We are taking a forward bet that it's going to continue to grow and that not only will chatypity keep growing, but we will be able to become one of the important AI clouds that our consumer device business will be a significant and important thing that AI that can automate science will create huge value. So, you know, there are not many times that I want to be a public company, but one of the rare times it's appealing is when those people are writing these ridiculous, open AI is about to go out of

Starting point is 00:15:07 business and, you know, whatever, I would love to tell them they could just short the stock, and I would love to see them get burned on that. But, you know, I, we carefully plan. We understand where the technology, where the capability is going to grow, go, and how the products we can build around that and the revenue we can generate, we might screw it up. Like, this is the bet that we're making and we're taking a risk along with that. People did not like this answer. And in fact, the hyperbolic commentary around it has been pretty fascinating. I have seen so many tweets indicating that somehow this is Sam's Gentile mask slipping

Starting point is 00:15:44 and that actually he's just this evil genius. when this to me reads kind of just like a CEO who's had to answer the same question a lot, getting fed up with answering the same question, and getting annoyed that people don't see what he sees. Now, unfortunately for Sam, I think the reality is that based on the position that he and his company have very intentionally put themselves in, he is now a politician. He's a diplomat. He's a statesman. And he kind of doesn't have the privilege of being cheeky and getting away with it. This is sort of just the cost of doing business when you put the entire economy.

Starting point is 00:16:16 me on your shoulders, and effectively that's what these 1.4 or 1.5 trillion in commitments does. I think being disappointed that he reached for Cheek instead of giving a more fully articulate answer is reasonable, but it certainly doesn't seem like some crazy mask slip moment to me. V.C. Tamaz Tongu has actually tried to crunch some of the math. He and his firm built a model, with the upshot being, these implied revenue figures suggest OpenAI would need to grow from 10 billion in 2024 revenue to 577 billion by 2029, roughly the size. of Google's revenue in the same year, assuming Google grows from $350 billion in 2024 at 12% annually. If nothing else, the estimated annual spending and commitments convey an absolutely enormous level

Starting point is 00:16:57 of potential and ambition. And indeed, this is the new phase that the conversation has gotten into. Not even just whether there's an overall AI bubble, but whether Open AI specifically is becoming too big to fail. That was the name of a recent op-ed in the Wall Street Journal. Now, the idea of too big to fail comes out of the global financial crisis in 2008 and 2009. The U.S. government, of course, made a determination that there were certain G-Sibs globally systematically important banks where the failure of those institutions would have too many ripple effects that would take down too big a chunk of the overall economy, thus making it intolerable for those companies to fail, meaning that the government would and eventually

Starting point is 00:17:32 did step in to stop the failure of individual companies in order to stop larger systemic contagion. Now, there is a lot of important nuance here. Too big to fail is actually kind of a misleading term. It suggests that the main question is just one. one of size, but it's actually not about size, it's about interconnection. The problem with the financial institutions wasn't their total assets under management. It was a potential for cascading contagion, where one failure, via leverage and other forms of interconnection, would trigger a cascade of

Starting point is 00:18:02 failures that ultimately wiped huge amounts of value out of the economy. Now, whether or not OpenAI has become too big to fail with all of these deals, the idea that it might certainly has captured notice. Florida Governor Ron DeSantis tweeted that Wall Street Journal piece and said, A company that hasn't yet turned a profit is now being described as too big to fail due to it being interwoven with big tech giants. Now, of course, as quickly as those op-eds started appearing, it's far from consensus that that's actually the case. All-in-Pods Jason Calcanus writes, the actual risk is not that OpenAI collapses, but that

Starting point is 00:18:33 they become one of five players in a highly competitive market. They will face margin compression in declining market share, which reduces a 30-plus-x price-to-sales ratio down to 5x, which means the market cap slows or maybe even stays flat. Flat market cap means it's impossible to make the $1.4 trillion in commitments as quickly as planned. Rezo responded, Open AI isn't too big to fail, if anything, they're too connected to fail. And there's a fundamental difference here. Size isn't the problem. The problem is the web of dependencies. Microsoft needs open AI for their AI story. Oracle needs them for utilization. Invita needs them for their demand narrative. This isn't systematic importance at circular dependency where everyone's pretending

Starting point is 00:19:08 the emperor has closed. Now, Razo concludes, but I think the bubble narrative is still wrong. We don't have a bubble. We have real demand for compute, real demand for scaling, and developing AI technologies. Compound 248 writes, people have lost the plot. Too big to fail refers to a systemic collapse risk. Open AI, which is unlikely to fail in any case, would not be an uncontrolled failure if it someday went bankrupt, nor would it have systematic implications. It would likely happen in slow motion, with equity holders diluted into oblivion or acquired cheap by a Microsoft. People latch on to phrases or concepts like Too Big to Fail because they're scary sounding, but they don't really understand them. Too Big to Fail is not applicable here. Open AI would be

Starting point is 00:19:46 fine to fail and would not require a government bailout. Now, if you want a sense of just how widespread the AI bubble conversation is getting, a reporter recently asked President Trump about it. The reporter said, could I ask you, some experts warn about an AI bubble. Are you concerned? Trump said, what's the AI problem? Reporter says some experts say that some investors are overreacting. Trump, and this is obviously a summary, in this case from unusual whales, says, Everybody wants AI because it's the new internet, it's the new everything, it's one of the biggest things anyone's ever seen. So everyone wants it.

Starting point is 00:20:15 Yeah, I mean, the only problem is if you don't get it. Goldman Sachs CEO, David Solomon also isn't sold on the doom and gloom. Speaking at a small business conference on Monday, he said, there will be disruption, but I'm a big believer that our economy is very nimble and very flexible. When you look at the technology that has flooded over hundreds of years into our society, we adapt. We find new businesses, we find new jobs. I don't believe it will be different this time. That said, he did acknowledge that the speed at which this is happening is different this time.

Starting point is 00:20:41 Solomon said, the pace of adoption of this technology is going a little bit faster. As businesses wrestle with deploying the technology and the automation, the short-term disruption might be a little bit higher. But our economy is incredibly broad and nimble. Indeed, I think if you look at how Wall Street is behaving right now, it's actually a little bit less clear-cut than perhaps some of the Twitter posters might have you think, in ways that do show the self-correcting nature of markets, and also show the value of the AI bubble narrative, even if it is just a narrative. On the one hand, you still have plenty of analysts

Starting point is 00:21:10 who think that there's lots more room to run. Loop Capital recently predicted that Nvidia can still go 70% higher, raising their price target from 250 to 350, compared to an average target of 231 among Wall Street analysts, and implying an $8.5 trillion valuation. But on the same day, we also got some interesting counter signals around Palantir. Pallenteer posted a record quarter and boosted forecasts on Monday night, but the stock still fell. Q3 revenue came in at $1.18 billion, and earnings per share outperformed Wall Street expectations by more than 20%.

Starting point is 00:21:40 In addition, Pallantir restated their Q4 guidance to as much as 12% higher than analyst forecasts. Still, the numbers weren't enough to drive another push higher, and while the stock spiked by 7% immediately after the announcement, it quickly retraced to fall by 4% during the overnight session. Even Pallentier's CEO, Alex Karp, acknowledged that the stock is a little stretched. He remarked during a Monday interview, we're in a nosebleed zone. No one else is here. Now, Michael Burry of big short fame announced massive shorts on Palantir and Invidia. And regardless of that, the point is that all of this AI bubble discussion is at least keeping the conversation considered. Look, we are in new territory here. No startup in history has ever done anything similar to what Open AI has done,

Starting point is 00:22:20 in terms of the pace and scale of user adoption, of revenue growth, and certainly not in terms of this incredible spade of dealmaking. It is important then to stay cautious and considered, even while trying to adjust our priors around what is possible. The good news, though, for those who don't want this to be just an out-and-out bubble with big systemic implications later on, is that the more the debate rages, the less likely the bubble actually is. For now, that's going to do it for today's AI Daily Brief. Appreciate you listening or watching, as always, and until next time, peace.

The AI Daily Brief: Artificial Intelligence News and Analysis - Is OpenAI Becoming Too Big to Fail?

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.