Divya van Mahajan

The AI Race Just Got a SpaceX-Sized Twist

ai-daily-brief-podcast

The AI Race Just Got a SpaceX-Sized Twist

Overview

This episode of The AI Daily Brief (dated February 3, 2026) covers two major stories in the AI industry: the merger of xAI and SpaceX into a single vertically integrated entity, and OpenAI’s launch of the Codex desktop app for macOS. The host argues that both stories represent significant structural shifts—one in the competitive financing and vision of frontier AI labs, and one in the paradigm of how developers interact with AI coding agents. The speaker is the unnamed host of the AI Daily Brief, a daily podcast and video covering AI news.

Source video: URL not provided in submission.

Prerequisites

  • Basic familiarity with the major AI foundation model labs: OpenAI, Anthropic, Google DeepMind, xAI, and Meta AI
  • Understanding of what large language models (LLMs) and coding agents are
  • Familiarity with Elon Musk’s company portfolio: Tesla, SpaceX (Starlink), X (formerly Twitter), and xAI (Grok)
  • Basic understanding of startup financing terms: Series E funding, EBITDA, valuation multiples, and IPOs
  • Awareness of AI coding tools: Claude Code, Cursor, GitHub Copilot, and terminal/CLI-based workflows
  • Conceptual understanding of agentic AI (agents that autonomously complete multi-step tasks)
  • Familiarity with the Kardashev scale (a measure of a civilization’s energy use/technological advancement)

Main Points

1. The Competitive AI Landscape Entering 2026

  • Anthropic had established leadership in coding and enterprise use cases, driven largely by Claude Sonnet 3.5.
  • OpenAI retained the strongest consumer brand and ChatGPT search adoption but was losing ground in enterprise and coding to Anthropic, and in general use to Google Gemini.
  • Google Gemini entered the year with strong momentum.
  • Meta was in a rebuilding year, with its Ray-Ban smart glasses as a bright spot.
  • xAI/Grok was the biggest question mark: could it leapfrog competitors via its Colossus compute clusters, and would Musk pursue synergistic consolidation across his companies?

2. xAI and SpaceX Merge: The Deal Structure

  • The combined entity is valued at $1.25 trillion: SpaceX at $1 trillion, xAI at $250 billion.
  • xAI’s prior valuation from its $20 billion Series E (closed earlier in January 2026) was approximately $230 billion—meaning the deal is roughly flat in valuation terms for xAI.
  • SpaceX’s official rationale: to create “the most ambitious vertically integrated innovation engine on and off Earth,” combining AI, rockets, space-based internet, and direct-to-mobile communications.
  • Musk personally framed the deal as an “acceleration of sci-fi technologies at the intersection of AI and space.”

3. Musk’s Vision: Orbital Data Centers

  • Musk’s stated thesis: within 2–3 years, the lowest-cost method of generating AI compute will be in space.
  • Advantages cited for orbital data centers:
    • Heat venting directly into space (no active cooling or water use)
    • Higher solar energy efficiency without atmospheric interference
    • Potential to scale to hundreds of terawatts—far beyond Earth’s current ~200 gigawatt data center capacity
  • SpaceX filed an FCC application to launch 1 million AI satellites into orbit; no competitor has launched more than a single test satellite for this use case.
  • The application describes the orbital network as a “first step toward becoming a Kardashev Type II civilization” (one that can harness the full power of its star).
  • Long-term vision includes lunar manufacturing capable of deploying 500–1,000 terawatts of AI satellites into deep space annually.

4. Skeptical Reactions and Financial Realities

  • Financial weakness of xAI: Net loss of $1.46 billion in Q3 2025 on $107 million revenue; $7.8 billion spent in the first three quarters of 2025. Revenue doubled quarter-over-quarter but would need to double four more additional times to match its burn rate.
  • SpaceX financials: $15 billion in revenue and $8 billion in “profit” (EBITDA) in the prior year—but critics note that depreciation is a significant factor in a rocketry/satellite business, making EBITDA less representative.
  • The combined entity trades at roughly an 80x revenue multiple, high even by private market standards.
  • Skeptic investor Ross Gerber summarized a common concern: a pattern of Musk companies sequentially merging when financially strained (X → xAI → SpaceX → Tesla?).
  • Critics argued SpaceX does not need xAI, but xAI needs SpaceX; the deal was seen as disadvantageous to SpaceX shareholders (roughly an 80/20 split).
  • The Information argued the merger is primarily financially motivated, not technology-driven, noting that xAI’s role in actually building orbital data centers is unclear.

5. Optimist and Accelerationist Perspectives

  • Optimists framed the combination as “turning rocket fuel into unlimited revenue”—SpaceX generates energy, intelligence (AI) turns it into economic value.
  • Engineer Andrew McAuliffe wrote a detailed blog post, Economics of Orbital vs. Terrestrial Data Centers, concluding the concept is “not rational but physically possible.”
  • The Starlink precedent was cited: satellite launch costs dropped 20x after Starlink launched, making what seemed economically impossible into a viable business.
  • Host’s framing: sufficiently advanced technology routinely starts as an idea that appears absurd, and SpaceX’s infrastructure track record warrants at least serious consideration.

6. xAI’s Product Momentum: Grok and Aurora

  • On the same day as notable competitor launches (Google Genie 3, etc.), xAI launched Grok Aurora 1.0: a new state-of-the-art image and video generation model.
    • Features: 10-second video generation, 720p resolution, significantly improved audio.
  • This signals xAI is not solely playing a financial engineering game—it is also competing on product and model capability.

7. OpenAI’s Codex App: A New Interface Paradigm

  • OpenAI launched the Codex desktop app for macOS, framed not as a simple coding assistant but as a “command center for agents.”
  • The stated rationale: since the April 2025 launch of Codex CLI, developer behavior has shifted from single-task completion to orchestrating multiple agents in parallel over long-running tasks (hours, days, or weeks).
  • Key features:
    • Manage multiple agents simultaneously
    • Built-in work tree support so multiple agents can operate on the same repository without conflicts
    • Seamless context switching between parallel tasks
    • Automations: scheduled/cron-style tasks (e.g., daily issue triage, CI failure summaries, release briefs)
    • Skills: extend Codex beyond code generation to information gathering, writing, and problem solving
  • OpenAI described a three-era evolution: autocomplete era → IDE/command line era → agentic/sub-agent orchestration era.

8. Reception and Competitive Impact of Codex App

  • OpenAI leadership was enthusiastic: Sam Altman called it “a bigger step forward than I imagined”; Greg Brockman said returning to the terminal “felt like going back in time.”
  • Notable AI creator Theo called it a potential “Cursor killer”.
  • Every.to founder Dan Shipper reported his usage split shifted from 80% Claude Code / 20% Codex to 50/50 within weeks, noting Codex was “slower but smarter and more reliable” for large production apps.
  • Developer Peter Steinberger (creator of popular open-source tools) publicly stated he uses Codex exclusively over Claude Code, citing bugs with Opus for his use cases.
  • Swix and Latent Space noted the significance: serious coding apps are now shipping without an IDE, reflecting how far coding models have advanced.
  • Krasinko cautioned the features would likely be copied by all agent managers within weeks.

9. Broader Identity and Labor Implications

  • Sam Altman posted publicly that building with Codex made him “feel a little useless”—noting the AI was generating better feature ideas than he was.
  • Multiple commentators observed a profound identity crisis forming among software engineers whose professional identity has been built around writing code manually.
  • Signal framed it on a civilizational scale: AI is severing the Western identity loop equating what you do for money with what you are worth—a shift that will force white-collar workers broadly to adopt new identities.
  • Host’s summary: “The coding language is English and anyone can build things by talking.”

10. Near-Term Outlook

  • OpenAI is winning back mindshare and conversation equity in coding from Anthropic—though this may be short-lived.
  • Strong signals pointed to an imminent Claude Sonnet 5 release, which could again shift the competitive dynamic within hours of this episode’s recording.

Key Concepts

  • Orbital data center: A proposed space-based computing infrastructure where satellites house AI compute hardware, benefiting from direct solar energy and passive heat dissipation into space.
  • Kardashev Type II civilization: A theoretical civilization capable of harnessing the total energy output of its home star, used here as a long-term framing for the orbital compute vision.
  • EBITDA: Earnings Before Interest, Taxes, Depreciation, and Amortization; a non-standard profitability measure that critics argue understates costs in capital-intensive businesses like rocketry.
  • Colossus cluster: xAI’s large-scale GPU compute infrastructure, cited as a potential source of performance leapfrogging over competitors.
  • Agentic era (in software development): A phase of AI-assisted coding where developers orchestrate multiple autonomous agents handling complex, long-running tasks rather than using AI for single-prompt code completions.
  • Work trees (git worktrees): A Git feature allowing multiple working copies of a repository to be checked out simultaneously, enabling parallel agent workflows on the same codebase without conflicts.
  • Automations (Codex feature): Scheduled or trigger-based tasks within the Codex app, analogous to cron jobs, used for recurring workflows like CI monitoring or issue triage.
  • Skills (Codex feature): Extensible capabilities within Codex that allow the agent to go beyond code generation to broader computer-use tasks like research, writing, and information synthesis.
  • Vibe coding: Informal term for building software by describing intent in natural language rather than writing code directly; previously associated with non-technical users, now increasingly applicable to professional developers.
  • Vertical integration (Musk context): The strategy of combining multiple companies across a value chain (AI, rockets, satellite internet, social media) under unified ownership to create synergistic advantages.
  • Rorschach test (media framing): Used by the host to describe how coverage of Musk news tends to reflect the commentator’s prior opinion of Musk rather than objective analysis of the facts.

Summary

This episode argues that the AI competitive landscape in early 2026 is being reshaped by two major structural developments. First, the merger of SpaceX and xAI represents Elon Musk’s most ambitious consolidation move yet—officially justified by a vision of orbital data centers that could unlock compute at scales Earth-bound infrastructure cannot support, but widely interpreted by skeptics as primarily a financial maneuver to shore up xAI’s heavy losses against the cash advantages of Google and Meta. The host takes a measured view: the financials are genuinely strained and the cynicism is understandable, but SpaceX’s track record with Starlink warrants taking the technical vision seriously even if it sounds extreme. Second, OpenAI’s Codex desktop app signals a deliberate bet that the coding paradigm has shifted from single-agent assistance to orchestrating parallel agent swarms, requiring a purpose-built interface rather than a terminal or IDE fork. Early reception suggests the app is meaningfully competitive with Claude Code for the first time in a long while, though Anthropic’s Claude Sonnet 5 is expected imminently. Underlying both stories is a deeper cultural disruption: as AI agents increasingly do the actual work of building software, the professional identity of the software engineer—and by extension, the broader equation of economic value with cognitive labor—is under serious pressure.