(Demetrius Freeman/Getty Images)

CTRL-F

Here’s all the AI stuff in Trump’s tax bill

Over $2 billion is allocated to deploying AI for nuclear weapons programs, autonomous underwater drones, and sovereign AI models.

Jon Keegan

7/2/25 12:00PM

Yesterday the Senate passed President Trump’s massive tax bill, and today it heads to the House, where it faces an uncertain future.

The text of the bill is loooong — it spans 870 pages as a PDF and is full of allocations of federal dollars for Trump’s priorities, including over $2 billion for AI programs for defense, energy, and homeland security.

Top AI insiders might hold a lot of sway in this administration, but tech companies like Microsoft, Meta, Palantir, and OpenAI did suffer a major loss yesterday when a key provision to halt all state-level AI regulation for a decade was removed from the bill.

Things could change in the House, but let’s take a look at all the AI-related things that made it through the Senate.

Defense

💣 Resources for scaling low-cost weapons into production

$124,000,000 for improvements to Test Resource Management Center artificial intelligence capabilities
$250,000,000 for the expansion of Cyber Command artificial intelligence lines of effort

💻 Improving efficiency and cybersecurity

$200,000,000 for the deployment of automation and artificial intelligence to accelerate the audits of the financial statements of the Department of Defense

☢️ Enhancement of resources for nuclear forces

$115,000,000 for accelerating nuclear national security missions through artificial intelligence

⛴️ Enhancement of Department of Defense resources for shipbuilding

$188,360,000 for the development and testing of maritime robotic autonomous systems and enabling technologies
$174,000,000 for the development of a Test Resource Management Center robotic autonomous systems proving ground
$200,000,000 for the development, procurement, and integration of mass-producible autonomous underwater munitions
$500,000,000 to prevent delays in delivery of attritable autonomous military capabilities
$75,000,000 to contract the services of, acquire, or procure autonomous maritime systems

Energy

🤖 Transformational artificial AI models ($150 million)

“American science cloud”: a system of US government, academic, and private sector programs and infrastructures utilizing cloud computing technologies to facilitate and support scientific research, data sharing, and computational analysis across various disciplines while ensuring compliance with applicable legal, regulatory, and privacy standards
Mobilize National Laboratories to partner with industry sectors within the United States to curate the scientific data of the Department of Energy across the National Laboratory complex so that the data is structured, cleaned, and preprocessed in a way that makes it suitable for use in artificial intelligence and machine learning models
Initiate seed efforts for self-improving artificial intelligence models for science and engineering

Homeland Security

🚔 Border security, technology, and screening

$168,000,000 for procurement and integration of new nonintrusive inspection equipment and associated civil works, including artificial intelligence, machine learning, and other innovative technologies, as well as other mission support, to combat the entry or exit of illicit narcotics at ports of entry and along the southwest, northern, and maritime borders

Rural hospitals

🏥 Rural health transformation program ($50 billion)

Providing training and technical assistance for the development and adoption of technology-enabled solutions that improve care delivery in rural hospitals, including remote monitoring, robotics, artificial intelligence, and other advanced technologies

Jon Keegan10/17/25

Google’s Gemini 3.0 reportedly due to be released in December

Google is aiming to release the latest version of its flagship AI model, Gemini 3.0, in December, according to a report from Sources.news.

The updated model is expected to make significant gains that should boost it to the top of the leaderboards, according to the report.

The Gemini app also spent some time at the top of the iOS App Store leaderboards, propelled by Google’s Nano Banana image generation model, which proved popular with users looking to turn themselves into action figures. Gemini briefly knocked ChatGPT from the top spot, which is now occupied by OpenAI’s other hot app, Sora.

Recently, there have been signs of ChatGPT downloads slowing, which could provide an opening for Gemini to gain market share. Adding some premium Gemini features to the free tier is a plan under discussion within Google, per Sources.news.

Sources.news also reports that a “small, secretive team” inside Google is working to integrate Gemini into Apple’s operating systems.

Google readies Gemini 3, Perplexity hits its goal

Sources.news also reports that a “small, secretive team” inside Google is working to integrate Gemini into Apple’s operating systems.

Jon Keegan10/17/25

Meta strikes $30 billion deal with Blue Owl to finance Hyperion data center

Meta’s Hyperion mega data center site in Richland Parish, Louisiana, is currently under construction. The city-sized development will be the home to one of the largest data centers in the world, housing around 2 million pricey GPUs, and will scale up to an eventual 5.5 gigawatts.

So, how is Meta planning to pay for this expensive project?

Bloomberg reports that Meta has signed a deal with asset management company Blue Owl Capital to finance $30 billion to pay for the project, marking what could be the largest private capital deal ever.

According to the report, Blue Owl and Meta would co-own the site, with Meta retaining a 20% stake in the project. PIMCO is also part of the financing for the deal, as the anchor lender.

Raising the massive capital to fund all of these huge AI data center projects is pushing companies to use unusual financing arrangements. The Information reported that xAI made such a deal with Valor Equity Partners worth $20 billion to rent the GPUs needed for its Colossus 2 data center.

Blue Owl Seals Largest Private Capital Deal for Meta’s AI Growth

According to the report, Blue Owl and Meta would co-own the site, with Meta retaining a 20% stake in the project. PIMCO is also part of the financing for the deal, as the anchor lender.

Rani Molla10/17/25

EssilorLuxottica surges to record high after saying Ray-Ban Meta glasses helped boost revenue growth

European eyewear company EssilorLuxottica said during its earnings call yesterday that its Ray-Ban Meta glasses helped boost its revenue growth, something that’s sent the ADR up to a record high.

“Clearly, there is a lift coming from Ray-Ban Meta wearables as a product category,” the company’s CFO, Stefano Grassi, said on the call Thursday. “The contribution from Ray-Ban Meta in wearables, as I mentioned before, is in excess of 4 percentage points overall for the group.”

EssilorLuxottica’s revenue was up 11.7% in the third quarter compared with a year ago.

Meta has a nearly 3% stake in the eyewear company, which it has partnered with on the smart glasses. Meta CEO Mark Zuckerberg has also claimed that its Ray-Ban Metas are a hit, saying that the “sales trajectory that we’ve seen is similar to some of the most popular consumer electronics of all time.” We looked at the numbers and aren’t so sure.

Tom Jones

ARR you watching?

10/17/25

The revenue race between Anthropic and OpenAI is getting more heated

Hey Claude, when do you think you’ll start making more money than ChatGPT?

44%

Matt Phillips10/16/25

JPMorgan economists estimate that the basket of stocks they use as a rough gauge of AI’s market impact is now worth about 44% of the S&P 500’s total market cap, up from 26% in 2022.

Using a basket of 30 AI stocks picked by the bank’s equity analysts as a barometer of AI, the economists find that American households have seen their aggregate wealth go up by about $5 trillion over the last year as a result of AI, they reported in a note published Thursday.

They also estimate the surge in stock market wealth could raise annualized US consumer spending by some $180 billion, due to wealth effects.

JPM acknowledges some uncertainty around this estimate, noting that the spending impact could be lower “if the wealth gains are accruing disproportionately to upper income households with lower [marginal propensity to spend].”