Transcript
A (0:04)
Welcome to the Tech Brew Ride home for Thursday, October 30, 2025. I'm Brian McCullough. Today, three big tech earnings reports give a mixed message. Wall street wants you to spend on the AI buildout, but not too much. And maybe not you. Meta Sora already has some new features, more details on OpenAI's IPO timeline, and maybe the AI agents are not yet ready for primetime. Here's what you missed today in the world of tech if you're looking for enterprise grade identity automation minus the enterprise grade baggage, aka having your users log on 500 times, Yeshid delivers advanced IAM automation without moving teams onto a legacy identity provider. Whether you use Google Workspace, Microsoft 365 or Okta, Yeshid integrates directly. No rebuilds or RIP and replaces are required. Yeshid helps IT and security teams reduce risk, not just tickets. And IT teams everywhere might have just breathed a collective sigh of relief. Every access change, review and approval is tracked and exportable, helping security teams effortlessly demonstrate compliance with SOC2, ISO or HIPAA. IT and security teams can spot risk before it becomes a finding. Learn more at yeshid.com techbrew that's Y-E-S-H I d.com techbrew I know tech earnings week always sort of sneaks up on me, but who decided that three of the biggies would all report on the same day? This time we're talking Meta, Alphabet, and Microsoft. And let's do it like this. Everybody basically did well. Meta's revenue was up 26%, Alphabet's revenue up 16%, and Microsoft's up 18%. But of those three, the street apparently didn't like Meta's report for reasons we can get into in a second. All the stocks are mixed this morning, but Meta is really taking it on the chin, down more than 10% as I write this. So what's $100 billion in market cap, give or take? Let's take it in reverse order. Microsoft's OpenAI investment actually resulted in a $3.1 billion hit to Q1 net income, for reasons I was not able to ascertain. But the key to all of these reports is cloud sector earnings and CapEx qu CNBC. Microsoft's Intelligent Cloud unit, which includes Azure, reported $30.9 billion in revenue, up 28% from a year ago and above the street Account consensus of 30.25 billion. Growth in Azure, which competes with Amazon Web Services and Google Cloud, also beat estimates, as analysts pulled by street account had anticipated 38.2% in terms of expansion. Cloud continues to be the big driver of growth at Microsoft, as the business has proven to be a major beneficiary of the AI boom. Last quarter, Microsoft disclosed the scale of its Azure cloud infrastructure business in dollars for the first time. The company said revenue in fiscal 2025 from Azure and other cloud services jumped 34% from the prior year to more than $75 billion. Investors were listening closely for Microsoft's commentary on capital expenditures as the company races to build out the infrastructure necessary to support AI demand. Microsoft's Hood said capex for the first quarter came in at $34.9 billion. She told investors in July that the company expected to spend $30 billion in capex and assets acquired through leases during the quarter. Hood said that the capex growth rate for fiscal 2026 will be above the rate in 2025. She'd previously said there would be a slowdown in growth. So again, keeping track here, capex for Microsoft came in slightly above what was expected, but not majorly. Microsoft is down only 2% as I write this this morning. As for Alphabet, they passed $100 billion in quarterly revenue for the first time in their history in Q3. Sundar Pichai said AI overviews drive meaningful query growth and AI mode has 75 million daily active users. I think the reason their stock is up 8% as I write this is because I think investors are thinking that Alphabet is threading the needle in terms of AI obviation of their search cash cow. Which is to say it's not quite obviating it yet in any meaningful way. Or at least investors believe Alphabet has a plan to address any obviation. YouTube revenue rose 15% to $10.26 billion. But then remember, here's the key graph. Google Cloud revenue beat estimates and they raised their 2025 capex forecast to between 91 and $93 billion after projecting $85 billion in July. So again a rise in capex. But investors seem to think that Alphabet is getting bang for their Cape. But now Meta again Meta Stock is down 10% this morning. Why? Well they did have to record a $15.93 billion one time income tax charge related to Trump's big beautiful bill and said they expect 2025 capex of $70 to $72 billion which is up from 66 to $72 billion they had forecast previously. And maybe that's the thing, investors don't see Meta getting bang for that capex buck anytime soon. Quoting the ft the varied reactions to their earnings and spending plans that were revealed on Wednesday afternoon quote underscores how sensitive investors are to how quickly the AI build out can deliver revenue, said Dec Mullarki, managing director of $30 billion asset manager SLC Management. Investors are worried that the rush to grab market leadership may cause an overshoot, he added. No one needs reminding that history is full of episodes of technology exuberance that eventually left the early investors battered. Microsoft, which surpassed a $4 trillion valuation this week after finalizing its restructuring pact with OpenAI, also suffered a share price drop. Its stock fell 2% despite beating profit estimates and posting a 39% jump in revenue at its key Azure cloud computing unit. It reported Capital expenditure was $35 billion in the quarter, a 74% increase year on year and $5 billion more than expected. Executives forecast spending of almost $140 billion next year. Chief Executive Satya Nadella told analysts that the software group was building Planet Scale cloud infrastructure and plans to double Microsoft's data center footprint over the next two years. Google and Microsoft, which both sell cloud computing to other businesses, had an easier time showing investors that elevated spending on chips, data centers and electricity will lead to income after a slow start in the AI race. Chief Executive Sundar Pichai said that the Gemini app, its main consumer AI product, now has 650 million monthly users, up from 450 million in July and closing in on ChatGPT's 800 million. Pichai added that growth in its cloud unit was driven by enterprise AI products, which are generating billions in quarterly revenues, and that it had an order backlog for computing services worth $155 billion. The 15% boost to core search advertising revenue also helped address concerns that ChatGPT is taking market share and and AI is cannibalizing traditional search. We believe this performance demonstrates successful AI integration across ad based platforms, said Angelo Zeno, an analyst at CFRA Research. Google's ability to maintain margins while scaling AI infrastructure demonstrates effective use of spending Mark Zuckerberg, meanwhile, defended huge spending on infrastructure for Meta's own use as the tech group viz to be the first to build artificial superintelligence. He said it was quote, the right strategy to aggressively front load building capacity. He added that any excess data center space could be repurposed to serve Meta's core advertising functions, which he said were compute starved. A 26% increase in quarterly revenue to $51.2 billion failed to mollify the market as investors fretted that Meta's huge outlay on chips and staff has yet to produce a large language model as capable as rivals. The social media company said capex could hit $72 billion by year end and that spending growth would be notably larger in 2026, implying a number far in excess of of an earlier forecast for $105 billion. Investors were disappointed by a rise in research and development costs, which accounted for 30% of revenue, the highest level in more than two years. Its operating margin narrowed 3 percentage points to 40%. Expenses are growing faster than revenue, said Gene Munster at Deepwater Asset Management. Next year it's going to be more like 18% revenue growth and 35% expense growth. Meta has indicated that its AI efforts are unlikely to generate meaningful revenue this year or in 2026. Zuckerberg promised that his new superintelligence team was focused on novel work that could be rapidly rolled out to three and a half billion users on Facebook, WhatsApp and Instagram and could make money via advertising, commerce or subscriptions. Investors worry Zuckerberg's quest to dominate advanced AI is disconnected from Meta's underlying business, despite his insisting that it can improve advertising, ranking and recommendations. Brian Weiser, an analyst at advisory firm Madison and Wall, said Google and Microsoft Quote are doing much more from a tech perspective. Meta's actual business is selling ads. There are so many more arrows in the quiver for Google and Microsoft, he added. End quote the digital world is more connected behind the scenes than you may realize. Interconnected is a video podcast series by Equinix that explores the hidden infrastructure behind our connected future. From data centers to cloud ecosystems to the platforms and people who use them, Interconnected's hosts bring tech leaders, industry experts and innovators together in candid conversations to break down and discuss the future of global connectivity. The third episode of Interconnected, for example, covers the digital infrastructure for a food secure world. They discuss how farmers are moving from 20th century operations to AI and machine learning that analyzes soil, weather and crop data to tackle 21st century risks. Plus how digital platforms are now connecting local producers to global demand through cloud Linked supply chains. Follow Interconnected on Apple, Spotify or wherever you get your podcasts. When your data goes dark, Veeam turns the lights back on. Partner with Veeam to increase your data resilience and get the right data recovery options for any kind of disruption so you can undo the unpredictable and get your data back so fast you won't even have time to miss it. With Veeam, it's all good. Keep your business running. @veeam.com Sora has added new features, including character cameos to create AI deepfakes of pets, also illustrations, clip stitching and leaderboards. Quoting the Verge, OpenAI's Sora app will now let you turn almost anything into a reusable avatar for its AI generated videos. These character cameos are one of several new features that have been added to the Sora 2 video generator, alongside clip stitching and leaderboards that showcase the app's most popular videos and cameos. Character cameos were teased last week and build on the existing feature that allows Sora users to create AI deepfakes of themselves that can be used by everyone else on the platform if permitted. Those same capabilities can now be used on other subjects such as pets, illustrations and toys, and the feature is launching with a selection of pre made characters that users can add to videos. Once created, each character comes with its own permissions, separate from your personal likeness. Keep it just for yourself, share it with mutual followers, or open it to everyone on Sora. OpenAI says in the latest Sora release notes give your character a display name and handle and tag it whenever you want it to appear in a video. OpenAI says that character cameos can be created using an original Persona that users generate in Sora, but it's unclear if the feature will accept fictional people generated using other AI tools or how photorealistic they're permitted to be if you can just upload any AI generated person. OpenAI hasn't explained how Sora would distinguish these from images of real people. The update is being launched within days of celebrity video platform cameo slapping OpenAI with a trademark infringement lawsuit over its usage of Cameo in Sora app features. Sora is also introducing video stitching capabilities that can connect multiple videos together to create longer multi scene clips and leaderboards that show the most remixed videos and the most cameoed users and characters. End quote. Cursor has launched Cursor 2.0 with its first homegrown coding model composer that it says is four times faster than similar models and a multi agent interface. Quoting Bloomberg, the overhauled service, called Cursor 2.0, marks a shift for the company, according to Andrew Millich, Cursor's head of product engineering. It's now positioning AI much more centrally in its software, visually and functionally, in the hopes that people will want to use the technology to field more and more multi step tasks that they traditionally carried out on their own, he said. Development has changed more in the last 18 months than it probably has changed in the last 18 years before that, Millich said in an interview. Cursor 2.0 will also include features like a Web browser to help users search for information such as documentation related to a certain kind of coding library, the company said. Cursor has raised $1.1 billion in funding to date and was most recently valued at about $10 billion. More than 1 million people now use the service each day, the company said, including coders at companies such as OpenAI, Spotify and Nvidia. End Quote More on OpenAI's path to an eventual IPO now that it is for profit quoting Reuters OpenAI is considering filing with securities regulators as soon as the second half of next year, some of the people said. In preliminary discussions, the company has looked at raising $60 billion at the low end and likely more, the people said. They cautioned that talks are early and plans, including the figures and timing, could change depending on how business growth and market conditions evolve, chief Financial Officer Sarah Fryer has told some associates. The company is aiming for a 2027 listing, the people said, but some advisors predict it could come even Sooner, around late 2026. The IPO preparations signal a new urgency inside the ChatGPT maker to tap public markets now that a complex restructuring is complete that reduces its reliance on micro Microsoft. An IPO would open the door to more efficient capital raising and enable larger acquisitions using public stock, helping to finance CEO Sam Altman's plans to pour trillions of dollars into AI infrastructure, according to people familiar with the company's thinking. With an annualized revenue run rate expected to reach about $20 billion by year end, losses are also mounting inside the $500 billion company, the people said. During a live stream on Tuesday, Altman addressed the possibility of going public. I think it's fair to say it is the most likely path for us given the capital needs that we'll have, he said. End quote. But in the background, Elon Musk's lawyer pledged to fight OpenAI's for profit restructuring and says California's and Delaware's attorneys general cannot sanitize OpenAI's on lawful conduct, quoting Bloomberg Although state officials didn't object to OpenAI's restructuring plan, California's attorney general said in a memorandum of understanding that the shift will have, quote, no impact on any other case, unquote, which will allow Musk's litigation to continue. QUOTE OpenAI's attempt to preempt that court and jury through an 11th hour restructuring only underscores why this case must proceed to trial, said Mark Toboroff, Musk's lead attorney in the OpenAI litigation. The CAAG preserved Mr. Musk's rights for good reason. We intend to vindicate them. End quote finally today not so fast, AI overlords according to Scale AI and CAIS's Remote Labor Index benchmark, which tests AI agents on freelance tasks, apparently the best AI could perform was just a little under 3% of the work assigned to it, earning a mere 1800 bucks. Quoting Wired the Remote Labor Index, a new benchmark developed by researchers at data annotation company Scale AI and the center for AI Safety. AIs, a nonprofit, measures the ability of frontier AI models to automate economically valuable work. The researchers gave several leading AI agents a range of simulated freelance work and found that even the best could perform less than 3% of the work, earning $1,810 out of a possible $143,991. The researchers looked at several tools and found the most capable to be Mannus from a Chinese startup of the same name, followed by Grok from Xai, Claude from Anthropic, ChatGPT from OpenAI, and Gemini from Google. I should hope this gives much more accurate impressions as to what's going on with AI capabilities, says Dan Hendricks, director of cais. He adds that while some agents have improved significantly over the past year or so, that does not mean that this will continue at the same rate. The researchers generated a range of freelance tasks through verified upwork workers. The tasks span a range of work, including graphic design, video editing, game development, and administrative chores like scraping data. They combine a description of each job with a directory of files needed to perform the work and an example of a finished project produced by a human. Hendrix says that while AI models have gotten better at coding, math and logical reasoning in recent years, they still struggle to use different tools and to perform complex tasks that involve numerous steps. They don't have long term memory storage and can't do continual learning from experiences. They can't pick up skills on the job like humans, he says. The analysis offers a counterpoint to a benchmark of economic work offered in September by OpenAI called GDPVAL, which purports to measure economically valuable work. According to GDPVAL, Frontier AI models such as GPT5 are approaching human abilities on 220 tasks across a range of office jobs. OpenAI did not provide a comment end quote by the way, since I haven't mentioned it recently, reminder that if you're an accredited investor, I operate a rolling fund, the Ride Home Fund, that invests in Pre Seed and Seed stage companies. The LPs in that fund are almost entirely fellow listeners to this podcast. So if you're interested in investing in early stage startups alongside me and fellow members of the Mutant Podcast army, check it all out@ridehomefund.com if anyone has a serious interest in maybe getting on board, I'd also be happy to give you more info myself. Hit me up@brianidehomefund.com talk to you tomorrow.
