AI Update: Grok’s Going to Space

Welcome to the newest edition of ID Tech’s AI Update. Here’s the latest big news on the shifting landscape of AI and identity technology:

Anthropic and OpenAI both launched new frontier models – pretty much simultaneously, yesterday. Anthropic’s Opus 4.6 is its smartest model, and is impressing people across a range of domains, including in physics. And OpenAI’s coding-focused GPT-5.3-Codex is both smarter and faster than its previous flagship coding model. Perhaps more importantly, it’s the first OpenAI model to have been “instrumental in creating itself,” according to the company. “Today will go down as some kind of turning point,” said the economist Tyler Cowen. “Somewhat arbitrarily, but it is OK if journalists and historians have to present things in that manner.”

Crypto.com founder and CEO Kris Marszalek has bought the domain AI.com for $70 million. This is the largest publicly disclosed cash-only domain sale in history. Marszalek has indicated that his plan is to launch some kind of personalized AI agent, available through the domain website, during the Super Bowl. The deal was paid entirely in cryptocurrency to an undisclosed seller.

Google’s parent company, Alphabet, announced plans for capital expenditure in the range of $175 billion to $185 billion for 2026 – roughly double its 2025 capex. Executives framed the ramp as AI-driven infrastructure buildout (data centers, servers, etc.), while also warning that the company remains supply-constrained. AI helped to fuel 48 percent growth in Google’s cloud services in the most recent quarter alone. “We are seeing our AI investments and infrastructure drive revenue and growth across the board,” said CEO Sundar Pichai.

Not to be outdone, Amazon says it will spend about $200 billion on capex this year. In its latest quarterly update, the firm said Amazon Web Services revenue growth accelerated to 24 percent year-over-year (reaching $35.6 billion), helping to push AWS to an annualized revenue run rate of roughly $142 billion. The spending will be focused on data centers, chips, servers, and AI infrastructure. CEO Andy Jassy said he’s eyeing “seminal opportunities like AI, chips, robotics, and low earth orbit satellites”.

French authorities raided X’s Paris office this week, in search of evidence related to a range of concerns, including AI-generated sexual abuse imagery and sexually explicit deepfakes from the Grok AI system that operates on the X social media platform. CEO Elon Musk has been formally summoned for questioning. Musk says it’s “a political attack,” and in a statement X said the “Paris Public Prosecutor’s Office is plainly attempting to exert pressure on X’s senior management in the United States by targeting our French entity and employees, who are not the focus of this investigation.”

SpaceX has acquired xAI in a deal valuing the combined company at $1.25 trillion. These are a couple of Elon’s other companies. One of them sends rockets into space, and the other one makes AI. Part of the logic here is that Elon believes their missions are tied together, and is aiming to get an “orbital data-center system” in place that can harness abundant solar power 24/7. He talked about this at some length on a recent podcast. He seems quite serious about it. “In the long term, space-based AI is obviously the only way to scale,” he has said.

The UK government says it will work with Microsoft, academics and other experts on a “deepfake detection evaluation framework” meant to set consistent standards for assessing detection tools and to map where current detection fails. The aim is to have a framework for assessing multiple detection tools “regardless of source,” rather than a single proprietary detector. The UK recently criminalized the creation of non-consensual intimate images, and pornographic deepfakes are a serious concern, as is fraud. “Deepfakes are being weaponised by criminals to defraud the public, exploit women and girls, and undermine trust in what we see and hear,” said Liz Kendall, Labour’s Secretary of State for Science, Innovation and Technology.

The chatbot’s take: Prompt: “Below is the latest AI Update feature I’ve written for ID Tech. Normally, at the end of each AI Update, I include a response from an AI system to a particular question. But in this case I am going to include a custom HTML block, and insert whatever you give me here. Please do something creative and technically challenging, and make sure it’s appropriate for an article post. I want to be able to copy and paste it into a custom HTML block.” (Editor’s note: we had to try this twice because the first custom HTML thing wasn’t compatible with our CMS.)

Frontier Convergence: Stress-Test the Next 12 Months

A static, interactive briefing card for AI acceleration, infrastructure concentration, and policy lag.

Signal Stack

  • New frontier models from Anthropic + OpenAI Capability cycle compresses further.
  • AI.com acquired for $70M Distribution and consumer entry points harden.
  • Alphabet + Amazon capex surge Compute race deepens hyperscaler leverage.
  • French probe into deepfake harms Enforcement shifts from rhetoric to operations.
  • UK detection evaluation framework Standards-first governance begins to form.

Scenario Lens

Acceleration dominates. Frontier models improve faster than deployment institutions adapt. Operational edge goes to teams with tight eval loops, rapid model routing, and rollback discipline.
Concentration intensifies. Capex scale and supply constraints centralize power in infrastructure owners. Strategic hedge: multi-cloud failover, open-model portability, and vendor-level incident clauses.
Trust becomes the bottleneck. Deepfake abuse, fraud, and provenance uncertainty constrain adoption. Winning systems pair high capability with verifiable authenticity and externalized audit trails.
Innovation Velocity84
Governance Strain71
Synthetic Harm Risk68

Editorial note: This module is intentionally script-free for CMS compatibility; interactions are CSS-only.

–

By Alex Perala