Kong

Kong started in 2009 as Mashape, two Italian schoolmates building a no-code tool out of a rented garage with IKEA furniture. When their API marketplace grew into an unworkable monolith, Augusto Marietti and Marco Palladino built an API gateway to break it apart, realized every other company had the same problem, and in 2017 open-sourced it as Kong.

That gateway became the most-deployed one in the world, and Kong has spent the years since bolting a full platform on top of it: service mesh, developer portals, monetization, and now an AI Gateway. The pitch for 2026 is blunt. There is no AI without APIs, every agent and model produces traffic, and someone has to manage it. Kong wants that someone to be Kong.

Kong sells the connectivity layer that sits between applications and everything they talk to. Its platform, Kong Konnect, lets platform teams design, publish, secure, observe, and monetize APIs across any cloud, with the open-source Kong Gateway doing the actual traffic handling at the edge.

The newer bet is treating AI traffic like API traffic. The AI Gateway puts a single control point in front of LLMs, MCP servers, and agents so teams can enforce auth, sanitize PII, cap token spend, and cache responses. Kong's own numbers give a sense of the scale it plays at: it says the platform moves more than a trillion API and AI requests a day.

Modern architectures fragment fast. Teams end up with a pile of microservices, multiple clouds, and legacy API management tooling that is slow, expensive, and hard to govern. Kong's job is to consolidate that into one gateway and one control plane, cutting latency and compute while giving security and platform teams a single place to enforce policy.

The AI era makes the sprawl worse. A single agent can fire thousands of calls a second across different models, and most companies have no way to govern the token spend, block prompt abuse, or even see what is happening. Kong reframes those as the same connectivity problem it already solves for APIs, which is the whole reason it can move in on AI infrastructure without starting from zero.

The core buyer is the platform engineering team standing up an API or Kubernetes platform for the rest of the company, plus the developers who live in it day to day. Kong leans hard into open source to win that crowd, then sells the enterprise controls once the gateway is load-bearing.

Above them sit the executives Kong markets to directly: leaders trying to migrate off legacy API management, consolidate platform cost, or get a handle on AI governance and spend. Kong's own solution pages carve the audience into platform teams, AI teams, developers, and executives, which is a fair map of who actually signs versus who actually uses it.

Kong's lineup fans out from the original open-source gateway. Konnect is the managed platform that ties it together; the AI Gateway and Event Gateway extend the same model to LLM and streaming traffic; and a run of acquisitions filled the gaps, with Kuma for service mesh, Insomnia for API design, and OpenMeter for usage-based billing.

The through-line is that everything is a gateway of some kind, sitting in the path of traffic and adding governance, security, and now monetization on top.

Kong runs the classic open-core playbook. The Kong Gateway is free and open source to drive adoption, and the money sits in Konnect, the managed platform, priced per gateway per month with a metered layer on top.

There is a 30-day free trial with no card required, a Plus plan billed monthly for smaller teams, and an Enterprise plan on custom annual pricing that unlocks the things large orgs actually need: SSO, audit logs, unlimited gateways, and dedicated support. The 2025 OpenMeter acquisition points at where this is going, letting Kong meter and bill AI token consumption the same way it bills API traffic.

Kong competes with the incumbents of API management and the cloud providers' own gateways, and increasingly with the crop of AI gateway startups chasing the same LLM traffic. Its case rests on being the most-adopted open-source gateway, so its distribution and community are hard to match, and on being genuinely fast and cloud-agnostic rather than tied to one provider.

The strategic tell is how early Kong reframed itself as the AI connectivity company. It is trying to own the AI gateway category before it fully forms, using the API gateway install base as the wedge.

Kong puts real weight on its customer roster: United Airlines, Nasdaq, HSBC, Goldman Sachs, PayPal, BMW, Sony Interactive Entertainment, Wayfair, ING Bank, and AI lab Mistral all appear on its site. It claims more than 100 Global 2000 customers and over 900 small business and enterprise accounts.

The case studies lead with hard numbers rather than vibes. One payment provider reports 20x faster time to market and a 75 percent cut in API management total cost of ownership; one bank cut app delivery from seven months to three and operating costs by 57 percent; a media company cites a 70 percent performance gain. Treat the figures as customer-reported, but the specificity is the point.

Developers consistently praise Kong for scaling cleanly across big microservices setups, keeping latency low, and supporting modern protocols like gRPC and GraphQL, with a deep plugin ecosystem doing a lot of the heavy lifting. The open-source mindshare is real: tens of thousands of GitHub stars and a community Kong pegs at over 182,000 members.

The recurring gripes are just as consistent. Setup has a steep learning curve, docs can be thin for beginners and for the developer portal, several of the genuinely useful features are gated behind paid tiers, and pricing plus support responsiveness draw complaints. It is powerful, but not the tool you casually pick up on a Friday afternoon.

Kong raised a $175 million Series E in November 2024 at a $2 billion valuation, a mix of primary and secondary money led by Tiger Global and co-led by Balderton, bringing total capital raised to roughly $345 million. Longtime backers Andreessen Horowitz, Index Ventures, CRV, Sapphire Ventures, and Notable Capital all came along, with Teachers' Venture Growth and 137 Ventures new to the cap table.

The raise landed as Kong crossed $100 million in ARR and repositioned around AI connectivity. The valuation is the bar it now has to clear, and the bet is that governing AI traffic becomes as unavoidable as governing API traffic already is.

Kong's strength is unusually durable: it owns the most-adopted open-source API gateway, a real community, and a roster of banks and Fortune 500 names that do not switch infrastructure lightly. That install base is its wedge into AI.

The open question is whether AI connectivity becomes the standalone market Kong is betting on, or whether the model providers and cloud platforms absorb governance themselves. Kong is moving early and pricing the future into a $2 billion valuation. If token traffic needs a neutral control plane the way API traffic did, Kong is positioned to be it. If it does not, the growth math gets harder.

Kong is remote-first with people in 25-plus countries and offices from San Francisco and Chicago to Milan, Bangalore, Tokyo, London, and Singapore. It still runs on the founders' origin myth of grit and hunger, and frames its values around being owners, explorers, real, and unstoppable.

Engineering skews toward hard distributed-systems work: billing platforms where correctness is not optional, multi-tenant identity infrastructure, and the gateway internals themselves. The perks lean into the remote model, with stock options for most employees, a four-week paid sabbatical every five years, flexible time off, a home-office stipend, and company-wide unplug days.

Kong discloses pay on many of its roles, and the ranges are wide because they span levels from mid to staff. US engineering roles run from roughly $100k to $260k base, with Canadian roles posted separately in CAD around $98k to $246k.

Sales pay skews high on the top end, up to roughly $360k, reflecting on-target earnings rather than base, and Kong is explicit that commercial roles come with uncapped commission plus equity. Most roles include stock options, so base is only part of the story.

The registered entity is Kong Inc., operating from 44 Montgomery Street in San Francisco, with a privacy policy that carves out explicit California, EEA, and UK disclosures for GDPR and CCPA. Data submitted through the hosted Konnect cloud is governed by the separate customer agreement rather than the site privacy policy.

Compliance shows up as a product feature, not just a checkbox. Kong builds SOC 2, FedRAMP, and GDPR support directly into its identity and gateway roadmap, which fits a customer base of banks and Fortune 500 buyers that will not deploy without it.

Kong's recent headlines track its shift from API management to AI connectivity: the open-source AI Gateway launch in early 2024, the $175 million Series E that November, and the 2025 OpenMeter acquisition that gave it usage-based billing for AI and API traffic. Read together, they are one story about a gateway company trying to become the toll booth for machine traffic.