Tech Brew Ride Home - Silicon Valley Circling The Wagons Around Anthropic?

Episode Date: March 6, 2026

Dario might need some message discipline as Anthropic is officially designated a risk by the US government. GPT-5.4 is here. Oracle is considering laying off a ton of people and Softbank is considering taking on a ton of debt, both for the same reason. An early warning system for AI job destruction. And, of course, The Weekend Longreads Suggestions. Anthropic says it will challenge Defense Department's supply chain risk designation in court (Engadget) Anthropic CEO apologizes for lashing out at Trump as he gears up for court battle with Pentagon (NYPost) OpenAI’s new GPT-5.4 model is a big step toward autonomous agents (The Verge) Oracle Plans Thousands of Job Cuts in Face of AI Cash Crunch (Bloomberg) SoftBank Seeks Record Loan of Up to $40 Billion for OpenAI Stake (Bloomberg) Anthropic launches AI job destruction detector (Axios) Labor market impacts of AI: A new measure and early evidence (Anthropic) Weekend Longreads Suggestions: SpaceX: the final frontier of IPOs (FT) Who needs data centers in space when they can float offshore? (TechCrunch) Learn more about your ad choices. Visit megaphone.fm/adchoices

It's peak pollination season, and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speed. That's why I chose GoogleFi Wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans start at $35 a month. Now, that's a deal that doesn't stay. Explore Google Fi Wireless plans today. Plus taxes and government fees.

Google Fi Wireless is not subject to data traffic deprioritization during times of high network usage. Welcome to the Tech Brew Ride Home for Friday, March 6, 2026. I'm Brian McCullough today. Dario might need some message discipline as Anthropic is officially designated a risk by the U.S. Government. GPT. 5.4 is here. Oracle is considering laying off a ton of people, and SoftBank is considering taking on a ton of debt, both for the same reason. An early warning system for AI job destruction and, of course, the weekend long read suggestions. Here's what you miss today in the world of tech.

The U.S. DoD says it has officially informed Anthropic leadership, the company, and its products are deemed a supply chain risk effective immediately. Dario Amadai says Anthropic plans to fight the DOD's risk designation in court, claiming the DOD's letter to it has a, quote, narrow scope, and apologizes for a recent leaked memo, quoting Engadgett. Amo Dye explained that the designation has a narrow scope because it only exists to protect the government. That is why the general public and even defense department contractors can still use Anthropics Claude Chatbot and its AI technologies. Microsoft told CNBC that it will

continue using Claude after its lawyers had concluded that it can keep on working with Anthropic on non-defense-related projects. More on that in a second. The CEO has also admitted that his company had productive conversations with the department over the past few days. He said that they were looking at ways to serve the Pentagon that adheres to its two exceptions, namely that its technology not be used for mass surveillance and the development of fully autonomous weapons, and looking at ways to, quote, ensure a smooth transition if that's not possible. That confirms reports that Anthropic is back in talks with the agency in an effort to reach a new deal. In addition, Amadai apologized for a leaked internal memo, wherein he reportedly said that OpenAI's messaging about its own deal

with the department is just straight up lies, end quote. Actually, he said more than that, quoting the New York Post. Anthropic CEO Dario Amadai has apologized for bashing the Trump administration in an explosive letter to staffers as he gears up for a court battle over the Pentagon's blacklisting of his AI firm. The execs said he's sorry for the, quote, tone of a 1600-word internal missive that accused the Department of War of targeting Anthropic for not giving dictator-style praise to Trump. I also want to apologize directly for a post-internal to the company that was leaked. to the press yesterday. Amadai wrote in a note published on his company's website Thursday, Anthropic did not leak this post nor direct anyone else to do so. It is not in our interest to escalate this situation. He went on to note his inflammatory comments came hours after Trump

blasted Anthropic staff as, quote, left-wing nut jobs, and Secretary of War Pete Heggseth said on X he would designate Anthropic a supply chain risk. It was a difficult day for the company, and I apologize for the tone of the post, Amadai wrote. It does not reflect my careful or considered views, end quote. In his memo to staffers earlier this week, Amadai said Anthropic was being punished because he didn't, quote, donate to Trump while, quote, Open AI slash Greg have donated a lot, referring to OpenAI President Greg Brockman. The information reported Amo Dye, who donated to Democratic former Vice President Kamala Harris's failed presidential campaign, blasted Open AI and the Pentagon for allegedly smearing his company's name. I want to be very clear on

the messaging that is coming from OpenAI and the mendacious nature of it, he wrote in the note, end quote. Indeed, also in that memo, Amo Dye said Open AI's DoD deal is, quote, safety theater, and the DOD dislikes Anthropic in part because it hadn't, again, given, quote, dictator-style praise to Trump, so that's why he's walking that back. Meanwhile, Microsoft says it will keep Anthropics AI tools embedded in its client products after its lawyers concluded the DOD's designation is only for defense projects, quoting CNBC. Microsoft is the first major company to say it will keep working with Anthropic after the Pentagon's actions. Some defense technology companies have told employees to stop using Anthropics' cloud models and migrate

to alternatives. Our lawyers have studied the designation and have concluded that Anthropic products, including Claude, can remain available to our customers other than the Department of War through platforms such as M365, GitHub, and Microsoft's AI Foundry, that we can continue to work with Anthropic on non-defense-related projects, a Microsoft spokesperson told CNBC in an email. Microsoft supplies its technology to a variety of U.S. government agencies. The Microsoft 365 productivity software is widely used inside the Department of War. In September, Microsoft said it was integrating Anthropics generative artificial intelligence models into the Microsoft 365 co-pilot add-on for Microsoft 365 subscriptions,

alongside models from OpenAI. Many software engineers have a. adopted Anthropics clawed models for drafting source code, and they are available in the GitHub copilot AI software development service, along with OpenAI's competing codex models. In November, Microsoft said Anthropic has committed to spending $30 billion on Microsoft's Azure cloud services, while Microsoft agreed to invest up to $5 billion in Anthropic, end quote. As expected yesterday, OpenAI launched GPT 5.4, saying it is its, quote, most capable and efficient frontier model for professional work, and it's first with native computer use capabilities. Quoting the Verge. It's also Open AI's first model with native computer use capabilities, meaning it can

operate a computer on your behalf and complete tasks across different applications. The new model is a step toward the agentic future that AI companies are aiming to build, where a network of AI-powered agents operates in the background to complete complex jobs online and within software. OpenAI introduced chat GPT agent amid a flurry of other agentic tools, that emerged last year, which can take control of your computer to perform tasks such as searching for and buying ingredients for a meal. While OpenAI is bringing GPD 5.4 to its API and its AI-powered coding tool, codex, it's rolling out its reasoning model, GPD 5.4 thinking to chat GPT. OpenAI says GPD 5.4 can write code to operate computers, as well as issue keyboard and mouse commands in response

to screenshots. GPT 5.4 also shows improvements while using web browsers as well as its ability to call upon tools and APIs more accurately and efficiently to help it complete tasks. The model is better at fielding questions that require it to gather information for multiple sources to, as OpenAI says the model, quote, can more persistently search across multiple rounds to identify the most relevant sources, particularly for needle and the haystack questions and synthesize them into a clear, well-reasoned answer. Open AI claims GPD 5.4 is its most factual model yet with individual claims 33% less likely to be false compared to GPD 5.2.

Inside, ChatGPT, 5.4 thinking will provide an outline of its work for more complex queries while also allowing users to tweak or change their request during its response. This makes it easier to guide the model toward the exact outcome you want without starting over or requiring multiple additional turns, opening I says. This feature is now available in the ChatGPT web app and on Android, but OpenAI says it's coming soon to the iOS app. GPD 5.4 is rolling out now across ChatGPT, Codex and the API with the GPT 5.4 thinking model coming to Plus, team, and pro users. There's also a 5.4 pro model for maximum performance on complex tasks rolling out in the API as well as for ChatGPT Enterprise and EDU users, end quote.

Sources tell Bloomberg that Oracle is planning to cut thousands of jobs as soon as March, among its moves to handle a cash crunch. for a massive AI data center expansion effort that we've been discussing at length. Quote, the job reductions will affect divisions across the company and may be implemented as soon as this month, according to people familiar with the matter who asked not to be named discussing the still private plans. Some of the cuts will be aimed at job categories that the company expects it will need less of due to AI, two of the people said. Wall Street projects the expenditures by the cloud unit for data centers to push Oracle's cash flow negative over the coming years before the spending begins to pay off in 2030, according to

data compiled by Bloomberg. Last month, Oracle said it would raise as much as $50 billion this year through a combination of debt and equity sales. The reductions being planned are expected to be wider-reaching than the company's typical rolling job cuts, according to the people. This week, Oracle announced internally that it would be reviewing many of the open job listings in its cloud division effectively slowing down or freezing the hiring process, according to people with knowledge of the move. Oracle declined to comment. The company had about 162,000 employees globally as of the end of May 2025. Planning for the workforce reductions is still active and could change, the people said. Oracle's initial moves as an AI cloud provider drew favor from investors who boosted the stock 61%

in 2024 and 20% last year. However, as the costs increased, the market has soured on the company with the shares following 54% from their September 2025 high through Wednesday's close, end quote. Study and play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds. Get the unreal college deal, everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 premium and a year of Xbox GamePass Ultimate with a custom color Xbox wireless controller. Learn more at Windows.com slash student offer. While supplies last ends June 30th, terms at AKA.m.m.m.m. Ready to soundtrack your summer? With Red Bull Summer All Day Play, you choose a playlist that fits

your summer vibe the best. Are you a festival fanatic, a deep end DJ, a road dog, or a trail mixer. Just add a song to your chosen playlist and put your summer on track. Red Bull Summer All Day Play. Red Bull gives you wings. Visit redbull.com slash bright summer ahead to learn more. See you this summer. Ambition comes in all shapes and sizes. At First Citizens Bank, we roll with your goals

because we're built for what you're building. Fit for your ambition for Citizens Bank. Whatever your thing, it could be anything. Canva helps you make that thing a thing. Canva is a simple online tool thing. It's a way to design with our magic AI tool things. You can social media your thing, generate images or videos of your thing, make decks for presentations to show your thing.

Whatever needs to be done for your thing, Canva can make it an even better and bigger thing. Canva, the thing that makes anything a thing. Bloomberg sources also say that SoftBank. is seeking a bridge loan of up to $40 billion, its largest ever borrowing denominated solely in dollars to help finance its investments in OpenAI. Quote, the bridge loan would have a tenor of about 12 months, according to some of the people who asked not to be identified discussing private matters. Four lenders, including J.P. Morgan Chase, will be underwriting the facility, the people said.

The potential size of the loan underscores SoftBank founder Masayoshi Son's aggressive bid to try and position his company as the linchpin in the global AI boom. The $30 billion bet on Open AI comes on top of more than $30 billion the company has already injected into the startup, which now forms the centerpiece of Sons' ambitions, a gamble reminiscent of his early investments in BiteDance or Alibaba, but at a far higher price. The Japanese company, which held about 11% in Open AI at the end of December, has unloaded assets, including at stake in NVIDIA to bankroll its growing bet on Open AI. Still, the scale of SoftBanks bet, as well as persistent concerns about a bubble, given the lack of a truly mainstream use case for AI services, has spooked market observers.

This week, S&P lowered SoftBanks credit outlook, citing the danger that its investments in Open AI may hurt the Japanese company's liquidity and the credit quality of its assets, end quote. Anthropic says Claude's daily signups have grown 4x since the start of 2026 and more than a million users now sign up for Claude every single day. According to Aptopia, Claude downloads jumped more than 220 percent week on week. So again, Anthropic has really become the main character of 2026 so far for so many reasons. But what I want to tell you about now is that Anthropic has debuted a sort of early warning system for potential AI-driven destruction of white-collar jobs and says it shows

limited evidence of AI-led job loss so far. Quoting Axios, while the new index from Claude's maker shows limited evidence that AI has affected joblessness so far, the efforts enter a larger debate among economists over how a possible AI labor doom and gloom scenario should be tracked in the first place. Anthropics' new measure takes into account an occupation-specific tasks, an estimate of which of those tasks can be performed by large language models, and which tasks are actually being done by AI today. Jobs are more exposed when their core tasks could be automated by AI, and Anthropics' anonymized data shows those tasks are already being automated in the real world. According to that, computer programmers around 75% task coverage,

customer service reps, data entry keyers, and medical record specialists rank among the most exposed to AI disruption, Anthropics says. Its economists estimate roughly 30% of occupations don't clear the minimum threshold to register as exposed in their index. Those are fields you might expect to be the least susceptible to AI disruption, given how human intensive they are, like cooks, lifeguards, dishwashers, and the like. Workers in the most exposed occupations have not become unemployed at meaningfully higher rates, though, than workers in jobs considered AI proof, according to the data. The average change in the gap since the release of ChatGBTBT is small and insignificant,

suggesting that the unemployment rate of the more exposed group has increased slightly, but the effect is indistinguishable from zero. Researchers write, Anthropic does find, quote, suggestive evidence that hiring of younger workers, particularly ages 22 to 25, has slowed in exposed occupations, a sign that certain entry-level workers are so far among the most affected by the uptake of AI, end quote. I'm also linking to the Anthropic study on this as well, because if you scroll down to figure two in that post, there is a handy graphic that shows you

which industries they feel are the most at risk. So see if yours is in there, I guess. In the long reads this week, the FT takes a look at SpaceX's potential IPO later this year, which reportedly aims to raise up to $50 billion at a $1.75 trillion valuation, more than seven times higher than it's around $200 billion valuation from October of 2024. And also, like one of the biggest IPOs of all time ever. Quote, with these numbers and with the SpaceX and XAI combination happening almost concurrently with the expected IPO, investors must make comic book heroic assumptions about sustained hypergrowth to justify a $1.75 trillion market cap.

The media have been guided to expect 2026 revenues of $22 to $24 billion, and given the likely mid-year timing, investors will be pointed to, 2027 projections, presumably with even punchier year-over-year growth. Even taking those forecasts at face value, we are aware of no serious company ever having gone public at anything close to the enterprise value to sales multiple apparently being contemplated here. Facebook, for example, went public on 11 times EV forward year sales. This would be a moonshot valuation inspired by a healthy dose of moonshine. Success hinges on the premise that Musk's personal brand, the lack of terrestrial

competitors and unbounded retail enthusiasm can propel the offering off the launch pad and into deep space, end quote. And then from TechCrunch, who needs data centers in space as Elon wants SpaceX to do when you can just put them in the ocean, maybe? Quote, one startup thinks that the ocean is a better place for AI data centers. Offshore wind developer Iquito is planning to submerge a 100-kilwatt demonstration data center off the coast of Norway this year. The small unit will live in the submerged pods of a floating offshore wind turbine. If all goes well, the company hopes to build a larger version to deploy off the coast of the UK in 2028. That model will sport a 15 megawatt to 18 megawatt turbine that will feed a 10 megawatt to 12 megawatt data center. The move offshore could solve a few

challenges. Proximity to power is an obvious one since the source will sit overhead. Winds offshore are more consistent than onshore, and a modest battery could bridge any laws. Submerged data centers could eliminate concerns from Not in My Backyard groups who oppose data centers near their properties due to noise and or pollution concerns. And lastly, by floating in cold seawater, cooling the service could be a simpler proposition. Cooling is one particularly vexing issue for orbital data centers since they need to employ different techniques in the vacuum of space. But for all of the challenges, offshore data centers could solve, they do introduce a few more. The ocean is a harsh environment. While submerged servers

wouldn't be battered by waves, they also wouldn't be completely stationary, so they'd need to be fully battened down. Sea water is also corrosive, so any equipment, including the container and power and data connections, will need to be hardened against it, end quote. No weekend bonus content for you this weekend. Talk to you on Monday. California's number one entertainment destination for today's superstars. Catch the Jonas Brothers return to the Yamava Theater stage on April 30th, the powerful vocals of Demi Lovato on May 17th, and the signature Southern Country Rock of Eric Church on July 19th.

Tickets on sale now at Yamavah Theater.com, only at Yamava Resort and Casino, celebrating its 40th anniversary. UN? Must be 21 to enter.