AI Revenue Explodes, Dario’s Memo, McDonald's CEO’s Baby Burger Bite

Summary of AI Revenue Explodes, Dario’s Memo, McDonald's CEO’s Baby Burger Bite

by Alex Kantrowitz

1h 0m•March 6, 2026
Listen to Episode

Overview of Big Technology Podcast — "AI Revenue Explodes, Dario’s Memo, McDonald's CEO’s Baby Burger Bite"

This episode (Friday edition) with host Alex Kantrowitz and guest Ranjan Roy reviews the latest rapid-growth signals in the AI business, the Anthropic–Pentagon dispute (including a leaked memo from Anthropic CEO Dario Amode), OpenAI product and finance news, infrastructure shifts (including Apple's surprising role), and a light cultural moment: McDonald’s CEO’s tiny bite of the new “Big Arch” burger and the ensuing viral reaction.

AI revenue snapshot and what it means

  • Reported annualized revenue (ARR) estimates:
    • OpenAI: ≈ $25 billion ARR (up from $21.4B at end of 2025 — ~17% increase).
    • Anthropic: ≈ $19 billion ARR (nearly 3× from end of 2024; up ~36% vs. two weeks prior).
  • Context:
    • Both companies went from almost zero in 2022 to ~$40B+ combined ARR in 2026 — a hockey-stick trajectory.
    • Hosts warn against naively extrapolating short-term bursts to full-year/year-over-year forecasts (method and base period matter a lot).

OpenAI projections, costs, and IPO implications

  • Reported revenue targets (from The Information’s reporting):
    • 2026: $30B; 2027: $62B; 2028: $113B; 2029: $184B; 2030: $284B.
  • Cash burn estimates:
    • Burned ~$8B in the last year; projections cited ~$25B burn in the current year and ~$57B next year (major compute/capacity investments).
  • Key discussion points:
    • Those revenue/valuation projections imply enormous capacity build-outs and sustained exponential adoption — hosts consider such long-term extrapolations unlikely to be >50% probability.
    • Public markets (S‑1s) will reveal economics, margins, and query-level costs; banks, legal firms, and regulatory scrutiny will be central to any IPO.
    • Jensen Huang / NVIDIA signaled their large investments (cited $30B into OpenAI, $10B into Anthropic) may be one-time allocations — public exit dynamics could reduce future private mega‑bets.

Competition, agentic commerce, and infrastructure

  • Agentic commerce & shopping:
    • OpenAI reportedly scaling back some agentic shopping initiatives; Meta and others are pushing into agentic commerce and browser/commerce integrations.
    • Amazon remains a major, distinct competitor because of logistics and ad revenue (ads are a major profit pool).
  • Outages & fragility:
    • Example: Amazon experienced a multi‑hour outage, underscoring fragility of major web destinations.
  • Apple as unexpected infrastructure:
    • Apple devices (M5 Max, Mac minis) running large models locally (e.g., LLaMA 70B) position Apple as an on‑device AI infrastructure provider.
    • On-device models change the narrative: if more workloads shift locally, cloud compute economics for OpenAI/Anthropic could be affected.

New model: OpenAI GPT‑5.4

  • Highlights:
    • Marketed as a step toward autonomous agents: improvements in reasoning, coding, and professional workflows (spreadsheets, docs, presentations).
    • First OpenAI model with native computer‑use capabilities — can write code to operate apps, issue keyboard/mouse commands from screenshots.
  • Reception:
    • Mixed reactions: some view it as meaningful progress (tool use, automation), others see it as iterative or already-available functionality in other integrations (e.g., Google/Excel add‑ins, Claude Code).

Anthropic vs. the Pentagon — supply chain risk and a leaked memo

  • Pentagon action:
    • U.S. Department of Defense (DoD) formally notified Anthropic that it is a supply‑chain risk for DoD projects — meaning Anthropic products cannot be used on those projects (but not an across‑the‑board ban).
  • Anthropic response:
    • Dario Amode: “We do not believe this action is legally sound and we see no choice but to challenge it in court.” (paraphrased)
    • Anthropic remains in talks with the Pentagon despite the designation.
  • Leaked Dario Slack memo:
    • Strongly worded, accusing OpenAI/Sam Altman of working with the DoD and contrasting Anthropic’s stance on regulation and “red lines.”
    • Claimed political/donation dynamics and criticized collusion for “safety theater.”
    • The leak has cultural implications internally (may make Anthropic more guarded) and externally (public relations/legal fallout).

Privacy, training data, and user guidance

  • Important practical note from the show:
    • Many chatbot services opt users in to using their conversations for model training by default.
    • To avoid having sensitive data used for training, users should proactively toggle the “data opt‑out” setting in each service (hosts recommend doing this, while acknowledging it’s not a 100% guarantee).
  • Broader concern:
    • The hosts discuss the increased surveillance and targeting risk in a “post‑AI” world, where analyzing petabytes of data at scale becomes more feasible.

Usage metrics: Claude vs ChatGPT

  • Anthropic claims:
    • Daily signups reportedly quadrupled since the start of the year; Anthropic says >1 million people sign up per day (claimed).
  • Third‑party app metrics (Aptopia):
    • Claude downloads spiked (e.g., up 220% on a certain day).
    • ChatGPT remains far larger: ~900 million weekly active users vs. Claude representing less than ~4% of daily mobile chatbot users as of February.
  • User/host opinion:
    • Ranjan cautioned Anthropic may be overstretching toward consumer positioning; their strongest story may be enterprise and API adoption.

McDonald’s “Big Arch” CEO bite — viral cultural moment (Burgergate)

  • What happened:
    • McDonald’s CEO Chris Kempczinski posted a video taking a very small, tentative bite of the new “Big Arch” burger; the micro‑bite went viral and drew wide mockery.
  • Cultural aftermath:
    • Other fast‑food CEOs (Burger King, Wendy’s) posted response videos, escalating the online fun.
    • Hosts view it as light, 2010s‑style internet culture resurfacing and a marketing win (increased attention and likely sales).
  • Takeaway:
    • Small, human PR moments can be hugely amplifying — even awkward CEO performances can drive virality and brand attention.

Key takeaways and recommended actions

  • For readers/listeners:
    • Be skeptical of headline ARR extrapolations — ask how annualization was calculated and whether recent bursts reflect sustained adoption.
    • Watch S‑1 filings and public disclosures for unit economics, cost per query, and stated burn rates — that will clarify sustainability.
    • Toggle off model‑training/data collection if you use chatbots for sensitive information.
    • Monitor competition (Google, Meta, Amazon, Apple) — on‑device AI (Apple) and platform/service integration could materially change cloud demand.
    • For brand/marketing teams: small CEO missteps can be diamond mines for viral attention — prepare social media playbooks that can pivot rapidly.
  • For investors and operators:
    • Expect heavy capital spending and potentially large reported losses as companies build capacity; evaluate credibility of growth assumptions and competitive moat.
    • Consider the geopolitical/regulatory risks — DoD/defense designations and government contracts have political dimensions and operational consequences.

Notable quotes (paraphrased)

  • Dario Amode: Anthropic will “challenge” the DoD designation in court and argued the DoD’s surveillance authorities take on new meaning in a post‑AI world.
  • Hosts’ perspective: The AI industry is growing extremely fast, but whether it’s sustainably exponential (vs. a rapid boom) remains open and will be tested by costs, competition, and public‑market scrutiny.

— End of summary —