a16zHow Kong Was Born: APIs, Hustle, and the Future of AI Infrastructure
CHAPTERS
- 0:00 – 2:18
From Milan to SF: a 90-day, $600 “make it or break it” bet
Augusto (Aghi) Marietti recounts arriving in the U.S. on a tourist visa with almost no money and a hard 90-day deadline to raise funding. The stakes were simple: raise fast or return to Italy broke.
- •Entered the U.S. on a tourist visa with ~$600 left
- •Set a 90-day fundraising window as an existential deadline
- •Early living situation underscored how close to failure they were
- •Establishes the “seven years of struggle” theme that follows
- 2:18 – 3:02
Hustle playbook: stealing the attendee list and cold-emailing ~400 investors overnight
At Stanford Entrepreneurship Week, Aghi takes a high-initiative (and ethically gray) approach to getting meetings. The tactic yields a small number of replies, but enough to generate real investor conversations.
- •Attended Stanford Entrepreneurship Week to meet founders and VCs
- •Took the registration email list and sent ~400 cold emails overnight
- •Response funnel: hundreds ignore, ~30 reply, ~10 show interest, ~5–6 meetings
- •Demonstrates early growth mindset: brute-force outreach to create luck
- 3:02 – 3:42
Closing the first checks: $51K from YouTube angels—signed on an air mattress
Aghi describes landing critical early angel funding from members of YouTube’s founding team. The money is small by Silicon Valley standards, but life-changing given their situation.
- •Met Kevin Donahue (YouTube founding team) through cold outreach momentum
- •First check written literally on an air mattress
- •Structured checks to avoid ‘too many sixes’ (superstition meets cap table)
- •$51K total became the bridge between survival and debt
- 3:42 – 5:21
Negotiating at Travis Kalanick’s ‘Jam Pad’: bathroom tactics and a desperate term sheet
The founders negotiate a tough convertible note in an unconventional setting—Travis Kalanick’s house—while Aghi is also crashing there in exchange for cooking. The story highlights naivete, urgency, and mentorship in high-stakes moments.
- •Stayed at Kalanick’s house; cooked carbonara as “rent”
- •Convertible note negotiation: investors push for extreme discount terms
- •Travis intervenes to prevent investors from walking and to force a counteroffer
- •Outcome: slightly improved terms (e.g., 42% vs 50% discount, higher cap)
- 5:21 – 8:31
Living on $1,000/month for three people: Airbnb mattresses, Starbucks offices, tuna pasta
With limited legal/work options and no U.S. system credentials, the team stretches funds to survive in San Francisco. Aghi details the day-to-day tactics—housing hacks, food math, and constant scrappiness—while trying to build momentum.
- •Returned to Italy, then re-entered on a B1 visa to stay longer
- •No SSN/credit: couldn’t operate like a normal U.S. startup
- •Three co-founders lived on ~$1,000/month total for over a year
- •Worked out of Starbucks; ate ultra-cheap staples (rice/beans/tuna/pasta)
- •Aghi handled blog/website/user calls/investor outreach while others coded
- 8:31 – 10:43
First major product shift: from drag-and-drop app builder to an API marketplace
Realizing the original “build apps via APIs” concept was too early for the market, the team pivots. A trip to Hawaii becomes the reset moment that reframes Mashape around an API marketplace thesis and leads to early traction and press.
- •Original idea: HTML5-era drag-and-drop app composition using APIs
- •Market not ready; missing ecosystem pieces to make it viable
- •Pivot decision formed during a ‘walk on the beach’ reset in Honolulu
- •Relaunch as an API marketplace connecting producers and consumers
- •Early coverage (TechCrunch-era outlets) helps validate the new direction
- 10:43 – 13:49
The seed round stack: NEA + Index, plus Bezos and Eric Schmidt (via creative networking)
Mashape’s seed round comes together with major firms and high-profile angels, enabled by unconventional paths to introductions. The round is large for the time and gives credibility, but also raises expectations.
- •CRV seed check used as early “momentum” signal; NEA leads seed
- •Index participates as it builds its U.S. presence
- •Jeff Bezos invested via the founders hiring his family office lawyer, then asking for an intro
- •Eric Schmidt invested through co-working proximity and referrals from a neighboring startup
- •Seed total ~ $1.5M—considered huge in 2010/2011
- 13:49 – 14:30
Getting legal: O-1 visas, letters of recommendation, and Sam Altman’s help
With real funding, the next existential challenge is immigration status. Aghi and Marco secure O-1 visas through letters—highlighted by Sam Altman writing a recommendation after meeting on an NEA retreat bus.
- •Visa status became the gating item to operate and hire in the U.S.
- •Sam Altman (then Loopt CEO) writes a key recommendation letter
- •Multiple letters assembled to qualify for O-1 visas
- •Once legal, they can rent space, hire, and build a real team
- 14:30 – 16:46
Series A realities: raising $6.5M and trying to prove a marketplace can monetize
Mashape raises a Series A but struggles to meet the economic profile expected for marketplaces. Even with traction, revenue and unit economics lag, setting up the next major pivot.
- •Series A process takes months; CRV leads with Index co-leading
- •Investors validate the theme (APIs as an ‘assembly line’) more than the execution
- •Round size: ~$6.5M (below their initial $10M ambition)
- •Post-raise expansion: bigger team and office, but business performance disappoints
- 16:46 – 18:39
Why the API marketplace stalled: liquidity, power-law supply, exclusivity, and quality problems
Aghi dissects the structural reasons the marketplace model didn’t become the “Airbnb/Uber of APIs.” Even as GMV grows, take-rate and cloud costs limit margins, making the model hard to scale sustainably.
- •Marketplaces require long-tail, low-power participants—APIs skew to a few winners (power law)
- •Lack of exclusivity: users can find APIs directly without the marketplace
- •Quality/trust issue: marketplace gets blamed for unreliable API supply
- •Economics challenged: ~10% net on $1.5M gross; margins eroded by AWS costs
- •Leads to recognition that the real asset is the underlying API gateway engine
- 18:39 – 19:56
The pivot that created Kong: open-sourcing the API gateway (and betting the company)
Facing runway pressure, the team identifies its internal gateway as the true product: routing, auth, logging, caching, and more for massive API traffic. They open-source Kong in 2015, and adoption takes off quickly.
- •Built a sophisticated gateway powering ~20,000 marketplace APIs
- •Core capabilities: billing, rate limiting, routing, caching, authN/authZ, logging
- •Decision: release the engine to the world as Kong API Gateway (open source)
- •Timing: April 2015 launch; immediate community traction
- •Marks the transition from marketplace operator to infrastructure platform
- 19:56 – 22:09
Bridge round to survive, then Series B with a16z: ‘weeks from dead’ during diligence
Even with Kong’s early momentum, the company needs an insider bridge to keep operating, then raises a crucial Series B. Martin describes the diligence signals—especially organic user love—that proved Kong was a real phenomenon.
- •Insider bridge (~$2M) extended runway after they were ‘out of gas’
- •Despite open-source breakout, fundraising was hard due to company age and prior pivots
- •Diligence validation came from repeated unsolicited user adoption stories
- •Aghi ‘spammed’ usage/customer signals to reinforce momentum
- •Deal closed after key meetings (including dinner with Marc Andreessen)
- 22:09 – 23:36
From breakout to scale: ARR milestones, category leadership, and the ‘car bet’
After the Series B, Kong transitions from near-death to rapid growth, eventually crossing major revenue thresholds. The chapter also captures the lighter cultural moments that mark true momentum—like the investor “car” wager.
- •Company grows from sub-$1M ARR to meaningful scale (eventually $100M+ mentioned)
- •Martin’s “hit X and I’ll buy you a car” promise becomes a story of outsized execution
- •Kong’s rapid growth follows years of starvation—showcasing non-linear outcomes
- •Competitive dynamics: earlier crowded field; Kong breaks away as independent leader
- •Market tailwinds include consolidation (Apigee/MuleSoft acquisitions)
- 23:36 – 34:29
AI, agents, and the next internet: why ‘machine-consumed’ connectivity is all APIs
Aghi argues AI will reshape how the internet is consumed: agents will interact programmatically rather than via UI. That shift increases the strategic importance of API and AI connectivity infrastructure—especially governance and authentication.
- •Agents consume services via programmatic interfaces, not webpages/UX flows
- •Protocols like MCP help standardize agent-to-tool interactions
- •AI needs ‘boring’ infrastructure: auth, key provisioning/rotation, monitoring, billing
- •Thesis: API traffic and AI (token) traffic converge into a unified connectivity layer
- •Gateway pattern repeats: enterprises will govern many LLMs like they govern microservices
- 34:29 – 37:29
Founder lessons from seven years of struggle: pick durable trends, stay lean, don’t quit
In closing, Aghi shares the mindset that kept him going through repeated near-death moments. He emphasizes choosing long-lived trends, keeping burn low, and maintaining conviction when timelines inevitably stretch.
- •Motivation anchored in refusing to return home as a failure
- •Big outcomes often require far longer timelines than founders expect
- •Choose trends that last 10–20 years so you have time to iterate and compound
- •Keep burn rate low early; survival creates optionality
- •Core directive: don’t die—persistence through setbacks is decisive
