In this episode of the AI Marketing Podcast, host Mark Fidelman sits down with David Amar, founder of Makina (a new conference dedicated to physical AI), to explore how robots and humanoids will change the future of marketing.
They discuss why robots are such powerful brand activations, when we might see in‑home humanoid housekeepers, how China is leading on hardware while the West leads on software, and why 2025–2026 feels like the "GPT moment" for physical AI. David also shares what to expect at Makina in Paris on July 7 and why marketers should get ahead of this trend now.
Guest
David Amar
- Background in computer science and neuroscience (UCL)
- Formerly worked in prosthetics
- Founder of Makina, a conference that brings together the fragmented physical AI ecosystem: humanoid builders, robot "brain"/OS providers, capital, industrial partners, and talent
Key Topics & Timestamps
1. Why Robots Are Marketing Gold
[0:00:00 – 0:02:24]
- David's background and the launch of Makina
- Why robots are "premium marketing material":
- Robots tap into deep cultural fascination (e.g., Star Wars, Star Trek)
- Simply announcing "Robot X/Y will be on site" can materially boost event attendance
- Humanoids as especially compelling because of their uncanny, human-like form
"I don't think I've ever met somebody that says this isn't interesting… It's just premium marketing material." – David [0:01:31]
2. Timeline: When Physical AI Hits Everyday Life
[0:02:24 – 0:03:25]
- David's long‑range outlook:
- Short term: impressive demos, but still lots of technical bottlenecks
- ~10–15 years: expect robots/humanoids in places we never imagined, with deep dependence on them
- Contrast with digital AI:
- We're already "slaves" to ChatGPT and cloud AI for knowledge work
- Physical dependence on robots will follow later
3. How Robots Show Up in Marketing (Beyond a Robot at a Desk)
[0:03:25 – 0:06:27]
- Robots won't replace marketers by typing at a desk—that's the realm of LLMs and digital AI
- Instead, robots will act as:
- Brand avatars and mascots (e.g., "the Amazon robot," "the Walmart robot")
- Physical activations at events, retail, and public spaces
- Product demo agents in stores, on the street, or wherever target audiences gather
- Comparison to today's street activations (e.g., sign spinners) but in a far more advanced, interactive form
- Emotional/branding angle:
- A charming C‑3PO‑style humanoid pitching products can be more captivating than a celebrity
"There's just something more charming about a C‑3PO showing the new Coca‑Cola than just a regular old Joe… even if it's George Clooney." – David [0:05:33]
4. Humanoids vs. "Robots" – What's the Difference?
[0:06:27 – 0:07:55]
- Humanoid:
- Robot with human‑like physiology and form (height, posture, movement)
- Tends to get anthropomorphic traits projected onto it
- Robot:
- Any robotic form, e.g. a single robotic arm, a robot dog, or R2‑D2‑style platforms
- Long‑term: Mark expects humanoids to become increasingly indistinguishable from humans in 20+ years
5. In‑Home Humanoids: How Close Are We Really?
[0:07:55 – 0:11:29]
- West vs. Asia split:
- West: stronger on software and AI models
- Asia (especially China): stronger on hardware and shipping units at scale
- Today you can already order multiple Chinese robot models online and have them delivered within a month
- Current leading players mentioned:
- 1X – focused on household/housekeeping tasks
- Sanctuary (Sunday Robotics) and others delivering early trial units
- Reality check on timelines:
- No one truly knows, but David's informed estimate:
- 5–7 years to order functional in‑home humanoids online
- Dependent on breakthroughs in:
- Fine manipulation of small objects
- Robust computer vision
- Autonomous navigation in unmapped environments
- Many "impressive" demos are partly marketing:
- Used to raise capital, build momentum, and buy time while teams fight through technical bottlenecks
6. Data, Compute, and How These Robots Actually Learn
[0:10:42 – 0:13:16]
- Today's deployed robots are often trial models used primarily to:
- Collect huge amounts of real‑world data
- Train the next generation of more capable robots
- Data and compute needs:
- Humanoids need even more data than LLMs:
- Touch, force feedback, vision, balance, navigation, etc.
- Massive compute, similar or greater than what's used for digital AI
- Where the compute lives:
- Training: in large data centers, often the same infrastructure used for AI
- On‑device inference:
- Onboard boards like NVIDIA Jetson inside the robot's "chest"
- Local models run on-device, optionally connected via Wi‑Fi for streaming data and updates
- Most robots in the wild are still tightly constrained and far from general-purpose autonomy
7. The Makina Physical AI Event in Paris
[0:13:59 – 0:19:11]
- Date: July 7
- Location: Station F, 13th district of Paris (central), the world's largest startup campus
- Format:
- One‑day dedicated physical AI conference
- Paired with the RAISE Summit (July 8–9), a broader AI conference
- Makina's mission:
- Fix the current ecosystem problem where events are:
- "Tech geeks talking to tech geeks"
- "Commercial to commercial" with limited cross‑pollination
- Bring together the full vertical stack of physical AI:
- Humanoid builders
- Robot brain / OS providers
- Investors and capital
- Industrial partners and adopters
- Talent and researchers
- Expected scale & hardware:
- Targeting 1,500+ attendees
- David's goal: 18–20 robots on site, split between stage demos and exhibitor robots
- Notable participants mentioned:
- Boston Dynamics (CEO Amanda)
- 1X (CEO)
- Google DeepMind Robotics leadership
- Other leading US, European, and Asian robotics companies
"We really are trying to regroup the very fragmented ecosystem that is physical AI… the vertical stack of physical AI in terms of ecosystems." – David [0:13:59]
8. Why Marketers Should Care (Now, Not Later)
[0:16:43 – 0:18:07] & throughout
- Humanoids and robots as future marketing must‑haves:
- Likely every major brand will have a robot avatar/mascot within ~10 years
- Use cases:
- Product demos & in‑store experiences
- Public activations and stunts
- Content creation, fail/reaction videos, and social media hooks
- Strategic advantage:
- Marketers who understand physical AI early will:
- Shape the first killer use cases
- Align brand positioning with new capabilities
- Avoid being late adopters in a fast‑moving bull run
- David's framing:
- Physical AI is in a "bull run", possibly a GPT‑moment equivalent for the physical world
- Huge capital flows, many new robotics startups, and intense industrial interest
9. Robotics Reality Check: Hype vs. Capabilities
[0:19:11 – 0:21:59]
- GTC/NVIDIA event anecdote:
- Few humanoids present, many robots were wired or constrained
- Mostly robot dog style units; limited truly autonomous humanoids
- Chinese hardware advantage:
- Companies like Unitree and others can ship robot dogs today, often ahead of US peers on commercialization
- Software and usefulness still lag:
- Hardware works, but what you can actually make them do is still narrow
- Many current units are about data collection more than true deployment
- Industrial partnerships:
- Examples: hexagon robotics & Mercedes, Figure & BMW, Boston Dynamics & Hyundai
- Present robots can perform very specific, tightly defined tasks with minimal uncertainty
- Outside of that, most can still just walk, dance, wave, and pose for marketing
"We're safe, and the Terminator is not coming next year." – David [0:21:59]
