Ever wonder why your AI assistant can’t just buy that flight ticket and charge it smoothly — without handing over your card details every damn time?
Skyfire’s CEO, Amir Sarhangi, has the answer: we’re missing the agentic commerce infrastructure. In a recent CB Insights interview, he lays it out plain — AI agents need their own layer of digital identity, trust, and payments to become real economic players on the web. And it’s sitting pretty on top of the $30 trillion global e-commerce beast, plus trillions in B2B.
Here’s the thing. Sarhangi doesn’t mince words on the scale.
At Skyfire, we are at the intersection of digital identity, agent trust, and payments in what we call the agentic commerce market, the infrastructure that allows AI agents to act as trusted economic actors on the internet.
Massive, right? But let’s pump the brakes — is this vision grounded, or just CEO bravado?
Is Agentic Commerce Really a Hundreds-of-Billions Opportunity?
Sarhangi pegs it there, no question. AI’s already nibbling at tasks — think chatbots booking rides or optimizing supply chains. Scale that to autonomous agents handling your entire shopping cart? E-commerce morphs overnight.
Current digital commerce? $30 trillion and climbing, per their estimates. B2B payments? Another multi-trillion shadow. Agents won’t replace it all — they’ll layer on, automating the friction. Skyfire positions itself as the Visa for machines: identity verification (no more ‘prove you’re not a bot’ for bots), trust protocols (agent reputations), and smoothly micropayments.
But. History whispers caution. Remember the blockchain hype for payments? Trillions promised, mostly vaporware outside crypto niches. Agentic commerce feels similar — exciting, yet unproven. My take: it’ll hit, but slower. Regulatory moats around identity and money movement could snag it for years. (Think KYC for AIs — who signs off on that?)
Skyfire’s edge? They’re building now, when incumbents like Stripe or Plaid are still human-focused. Early movers win rails wars.
Numbers back the buzz. Gartner predicts 30% of enterprise apps will be agent-driven by 2028. McKinsey sees AI automating 45% of work activities. Mash that with e-comm growth at 15% CAGR — yeah, hundreds of billions isn’t wild.
Why Do AI Agents Need Machine-Only Payment Rails?
Humans click ‘buy now’ with passwords and 3D Secure. Agents? They’ll swarm servers, negotiating deals in milliseconds, splitting payments across wallets.
Existing rails choke on that. Visa’s geared for your iPhone tap, not a fleet of Grok-like shoppers. Friction kills autonomy — agents need zero-trust identity (ironic, huh?), programmable trust scores, and atomic transactions that settle instantly.
Sarhangi nails it: “As AI begins to shop, access services, and complete tasks on behalf of users and businesses, the internet needs a trust and transaction layer designed for machines, not humans.”
Spot on. Without this, agents stay leashed — fun demos, no prime time.
Look, I’ve seen fintech pitches like this before. Remember embedded finance? Banks freaked, then adapted. Same here. Big Tech (Anthropic, OpenAI) will push agents; they’ll need Skyfire-like plumbing or build their own. Prediction: by 2030, 15-20% of online transactions flow through agent rails. That’s my unique call — not in the interview, but drawn from agent adoption curves mirroring mobile payments’ rise in Asia.
Critique time. Skyfire’s market sizing feels PR-polished — ‘hundreds of billions over time’ is vague enough to dazzle VCs without specifics. Where’s the TAM breakdown? Still, dynamics favor them. Competition? Sparse. Maybe Zeta or Unit for B2B, but no pure agent play.
Skyfire’s Play: Trust as the Killer App
Not just payments. Digital identity first — agents prove sovereignty without user creds. Then trust: slashing fraud in a world where deepfakes rule.
Market fit? Perfect for fintechs building agent copilots (hello, Ramp, Brex). Enterprises too — automate procurement, no human oversight.
Risks? Volatility. If agents flop (hallucinations, anyone?), market craters. Or regulators step in — EU’s AI Act could mandate agent audits, hiking costs.
But bullish dynamics win. VC poured $50B into AI last year; fintech slice grows. Skyfire, fresh off funding (assume seed vibes), rides the wave.
Wider ripple? E-comm platforms pivot — Shopify agents? Amazon’s Rufus scales to transact. Winners: infra like Skyfire. Losers: legacy gateways stuck in human mode.
And the PR spin? Sarhangi’s interview screams ‘invest in us’ — fair game, but readers, probe the demos. Show me agents paying live.
The Roadblocks No One’s Talking About
Short para. Talent wars — AI + fintech unicorns scarce.
Then scalability. Billions of agents? Infra buckles without quantum-ready crypto or whatever.
My insight: parallel to 90s internet plumbing. Netscape didn’t own browsers forever; standards emerged. Agentic commerce? Open protocols or bust.
Skyfire smart if they push standards body involvement.
🧬 Related Insights
- Read more: Iran’s $2 Million Bitcoin Demand: The Strait of Hormuz Toll That Could Shake Oil Markets
- Read more: Stablecoins Just Dethroned the US Banking Backbone—And Nobody’s Talking About It
Frequently Asked Questions
What is agentic commerce?
Agentic commerce is the infrastructure letting AI agents handle buying, services, and tasks online with their own identity, trust, and payments — built for machines on top of human e-commerce.
Skyfire competitors in AI payments?
Few direct rivals yet — players like Stripe experiment, but Skyfire focuses purely on agent trust and identity. Watch for Big Tech moves from Google or xAI.
When will AI agents dominate shopping?
Not tomorrow — 3-5 years for pilots, 10 for mass adoption, per market dynamics. Regulatory hurdles slow it, but $30T e-comm upside accelerates.
