Railway, a San Francisco-based cloud platform that has quietly amassed two million builders with out spending a greenback on advertising, introduced Thursday that it raised $100 million in a Sequence B funding spherical, as surging demand for synthetic intelligence purposes exposes the restrictions of legacy cloud infrastructure.
TQ Ventures led the spherical, with participation from FPV Ventures, Redpoint, and Uncommon Ventures. The funding values Railway as one of the vital infrastructure startups to emerge in the course of the AI increase, capitalizing on developer frustration with the complexity and value of conventional platforms like Amazon Internet Companies and Google Cloud.
"As AI fashions get higher at writing code, an increasing number of persons are asking the age-old query: the place, and the way, do I run my purposes?" mentioned Jake Cooper, Railway's 28-year-old founder and chief govt, in an unique interview with VentureBeat. "The final technology of cloud primitives have been gradual and outdated, and now with AI shifting all the things quicker, groups merely can't sustain."
The funding is a dramatic acceleration for a corporation that has charted an unconventional path by the cloud computing {industry}. Railway raised simply $24 million in complete earlier than this spherical, together with a $20 million Sequence A from Redpoint in 2022. The corporate now processes greater than 10 million deployments month-to-month and handles over one trillion requests by its edge community — metrics that rival far bigger and better-funded opponents.
Why three-minute deploy instances have grow to be unacceptable within the age of AI coding assistants
Railway's pitch rests on a easy remark: the instruments builders use to deploy and handle software program have been designed for a slower period. A normal build-and-deploy cycle utilizing Terraform, the industry-standard infrastructure instrument, takes two to 3 minutes. That delay, as soon as tolerable, has grow to be a vital bottleneck as AI coding assistants like Claude, ChatGPT, and Cursor can generate working code in seconds.
"When godly intelligence is on faucet and may remedy any drawback in three seconds, these amalgamations of programs grow to be bottlenecks," Cooper advised VentureBeat. "What was actually cool for people to deploy in 10 seconds or much less is now desk stakes for brokers."
The corporate claims its platform delivers deployments in beneath one second — quick sufficient to maintain tempo with AI-generated code. Prospects report a tenfold improve in developer velocity and as much as 65 % price financial savings in comparison with conventional cloud suppliers.
These numbers come immediately from enterprise shoppers, not inside benchmarks. Daniel Lobaton, chief know-how officer at G2X, a platform serving 100,000 federal contractors, measured deployment pace enhancements of seven instances quicker and an 87 % price discount after migrating to Railway. His infrastructure invoice dropped from $15,000 per thirty days to roughly $1,000.
"The work that used to take me per week on our earlier infrastructure, I can do in Railway in like a day," Lobaton mentioned. "If I need to spin up a brand new service and take a look at completely different architectures, it could take so lengthy on our previous setup. In Railway I can launch six providers in two minutes."
Contained in the controversial choice to desert Google Cloud and construct knowledge facilities from scratch
What distinguishes Railway from opponents like Render and Fly.io is the depth of its vertical integration. In 2024, the corporate made the bizarre choice to desert Google Cloud completely and construct its personal knowledge facilities, a transfer that echoes the well-known Alan Kay maxim: "People who find themselves actually critical about software program ought to make their very own {hardware}."
"We needed to design {hardware} in a method the place we might construct a differentiated expertise," Cooper mentioned. "Having full management over the community, compute, and storage layers lets us do actually quick construct and deploy loops, the type that permits us to maneuver at 'agentic pace' whereas staying one hundred pc the smoothest experience on the town."
The method paid dividends throughout latest widespread outages that affected main cloud suppliers — Railway remained on-line all through.
This soup-to-nuts management allows pricing that undercuts the hyperscalers by roughly 50 % and newer cloud startups by three to 4 instances. Railway costs by the second for precise compute utilization: $0.00000386 per gigabyte-second of reminiscence, $0.00000772 per vCPU-second, and $0.00000006 per gigabyte-second of storage. There aren’t any costs for idle digital machines — a stark distinction to the standard cloud mannequin the place prospects pay for provisioned capability whether or not they use it or not.
"The traditional knowledge is that the large guys have economies of scale to supply higher pricing," Cooper famous. "However after they're charging for VMs that normally sit idle within the cloud, and we've purpose-built all the things to suit rather more density on these machines, you might have a giant alternative."
How 30 staff constructed a platform producing tens of hundreds of thousands in annual income
Railway has achieved its scale with a group of simply 30 staff producing tens of hundreds of thousands in annual income — a ratio of income per worker that may be distinctive even for established software program firms. The corporate grew income 3.5 instances final yr and continues to develop at 15 % month-over-month.
Cooper emphasised that the fundraise was strategic slightly than crucial. "We're default alive; there's no motive for us to lift cash," he mentioned. "We raised as a result of we see an enormous alternative to speed up, not as a result of we wanted to outlive."
The corporate employed its first salesperson solely final yr and employs simply two options engineers. Practically all of Railway's two million customers found the platform by phrase of mouth — builders telling different builders a few instrument that truly works.
"We principally did the usual engineering factor: in the event you construct it, they’ll come," Cooper recalled. "And to a point, they got here."
From facet initiatives to Fortune 500 deployments: Railway's unlikely company enlargement
Regardless of its grassroots developer group, Railway has made vital inroads into giant organizations. The corporate claims that 31 % of Fortune 500 firms now use its platform, although deployments vary from company-wide infrastructure to particular person group initiatives.
Notable prospects embody Bilt, the loyalty program firm; Intuit's GoCo subsidiary; TripAdvisor's Cruise Critic; and MGM Resorts. Kernel, a Y Combinator-backed startup offering AI infrastructure to over 1,000 firms, runs its complete customer-facing system on Railway for $444 per thirty days.
"At my earlier firm Intelligent, which bought for $500 million, I had six full-time engineers simply managing AWS," mentioned Rafael Garcia, Kernel's chief know-how officer. "Now I’ve six engineers complete, and so they all deal with product. Railway is precisely the instrument I want I had in 2012."
For enterprise prospects, Railway presents safety certifications together with SOC 2 Kind 2 compliance and HIPAA readiness, with enterprise affiliate agreements out there upon request. The platform offers single sign-on authentication, complete audit logs, and the choice to deploy inside a buyer's current cloud surroundings by a "deliver your personal cloud" configuration.
Enterprise pricing begins at customized ranges, with particular add-ons for prolonged log retention ($200 month-to-month), HIPAA BAAs ($1,000), enterprise assist with SLOs ($2,000), and devoted digital machines ($10,000).
The startup's daring technique to tackle Amazon, Google, and a brand new technology of cloud rivals
Railway enters a crowded market that features not solely the hyperscale cloud suppliers—Amazon Internet Companies, Microsoft Azure, and Google Cloud Platform—but additionally a rising cohort of developer-focused platforms like Vercel, Render, Fly.io, and Heroku.
Cooper argues that Railway's opponents fall into two camps, neither of which has totally dedicated to the brand new infrastructure mannequin that AI calls for.
"The hyperscalers have two competing programs, and so they haven't gone all-in on the brand new mannequin as a result of their legacy income stream remains to be printing cash," he noticed. "They’ve this mammoth pool of money coming from individuals who provision a VM, use possibly 10 % of it, and nonetheless pay for the entire thing. To what finish are they really taken with going all the best way in on a brand new expertise in the event that they don't really want to?"
Towards startup opponents, Railway differentiates by protecting the total infrastructure stack. "We're not simply containers; we've received VM primitives, stateful storage, digital non-public networking, automated load balancing," Cooper mentioned. "And we wrap all of this in an absurdly easy-to-use UI, with agentic primitives so brokers can transfer 1,000 instances quicker."
The platform helps databases together with PostgreSQL, MySQL, MongoDB, and Redis; offers as much as 256 terabytes of persistent storage with over 100,000 enter/output operations per second; and allows deployment to 4 international areas spanning the USA, Europe, and Southeast Asia. Enterprise prospects can scale to 112 vCPUs and a pair of terabytes of RAM per service.
Why buyers are betting that AI will create a thousand instances extra software program than exists immediately
Railway's fundraise displays broader investor enthusiasm for firms positioned to profit from the AI coding revolution. As instruments like GitHub Copilot, Cursor, and Claude grow to be normal fixtures in developer workflows, the amount of code being written — and the infrastructure wanted to run it — is increasing dramatically.
"The quantity of software program that's going to return on-line over the subsequent 5 years is unfathomable in comparison with what existed earlier than — we're speaking a thousand instances extra software program," Cooper predicted. "All of that has to run someplace."
The corporate has already built-in immediately with AI programs, constructing what Cooper calls "loops the place Claude can hook in, name deployments, and analyze infrastructure robotically." Railway launched a Mannequin Context Protocol server in August 2025 that permits AI coding brokers to deploy purposes and handle infrastructure immediately from code editors.
"The notion of a developer is melting earlier than our eyes," Cooper mentioned. "You don't need to be an engineer to engineer issues anymore — you simply want vital pondering and the power to research issues in a programs capability."
What Railway plans to do with $100 million and nil advertising expertise
Railway plans to make use of the brand new capital to develop its international knowledge middle footprint, develop its group past 30 staff, and construct what Cooper described as a correct go-to-market operation for the primary time within the firm's five-year historical past.
"One in every of my mentors mentioned you increase cash when you’ll be able to change the trajectory of the enterprise," Cooper defined. "We've constructed all of the required substrate to scale indefinitely; what's been holding us again is just speaking about it. 2026 is the yr we play on the world stage."
The corporate's investor roster reads like a who's who of developer infrastructure. Angel buyers embody Tom Preston-Werner, co-founder of GitHub; Guillermo Rauch, chief govt of Vercel; Spencer Kimball, chief govt of Cockroach Labs; Olivier Pomel, chief govt of Datadog; and Jori Lallo, co-founder of Linear.
The timing of Railway's enlargement coincides with what many in Silicon Valley view as a basic shift in how software program will get made. Coding assistants are not experimental curiosities — they’ve grow to be important instruments that hundreds of thousands of builders depend on day by day. Every line of AI-generated code wants someplace to run, and the incumbents, by Cooper's telling, are too wedded to their current enterprise fashions to totally capitalize on the second.
Whether or not Railway can translate developer enthusiasm into sustained enterprise adoption stays an open query. The cloud infrastructure market is affected by promising startups that failed to interrupt the grip of Amazon, Microsoft, and Google. However Cooper, who beforehand labored as a software program engineer at Wolfram Alpha, Bloomberg, and Uber earlier than founding Railway in 2020, appears unfazed by the dimensions of his ambition.
"In 5 years, Railway [will be] the place the place software program will get created and advanced, interval," he mentioned. "Deploy immediately, scale infinitely, with zero friction. That's the prize value taking part in for, and there's no larger one on provide."
For an organization that constructed a $100 million enterprise by doing the other of what standard startup knowledge dictates — no advertising, no gross sales group, no enterprise hype—the actual take a look at begins now. Railway spent 5 years proving that builders would discover a higher mousetrap on their very own. The following 5 will decide whether or not the remainder of the world is able to get on board.
