The pitch from Stefaan Vervaet sounds straightforward enough: What if you could move massive AI training datasets between cloud providers—or off the cloud entirely—without rewriting code, renegotiating contracts, or bleeding money on egress fees?
That premise just landed Akave, his Austin-based storage startup, $6.65 million in seed funding announced for March 2, 2026. The round, which includes checks from Protocol Labs, Avalanche Foundation, and a roster of crypto-focused investors like Blockchange and No Limit Holdings, brings the company's total disclosed capital past $10 million following a $3.45 million pre-seed last November.
At its core, Akave is building S3-compatible decentralized storage—infrastructure that mimics Amazon Web Services' ubiquitous object storage standard but lives outside the hyperscaler ecosystem. The company's newly launched Akave Cloud charges $14.99 per terabyte per month with no egress fees and no per-request nickel-and-diming, a cost structure that, if it holds at scale, could appeal to machine learning teams tired of surprise bills when they shuttle data across environments.
Whether that model can muscle into an enterprise storage market still dominated by AWS, Google Cloud, and Microsoft Azure is another question entirely.
The Egress Fee Problem (and the Blockchain Solution?)
Vervaet, who previously ran network growth for Filecoin at Protocol Labs, frames Akave around a specific pain point: organizations that want control over where their data lives and how it moves. "AI training pipelines accumulate costs fast when you're moving large datasets between environments," he said in a recent conversation, though he declined to share revenue figures or customer counts beyond a handful of early design partners.
The startup's pricing undercuts AWS S3's standard tier significantly, particularly once data retrieval enters the picture. For context, competitors like Wasabi and Impossible Cloud advertise even lower per-terabyte rates—$6.99 and $7.99 respectively—also without egress tolls. Akave isn't trying to win on price alone. Instead, the company is leaning into programmability and what Vervaet calls "on-chain auditability," courtesy of a dedicated Avalanche L1 blockchain that handles access controls and immutable audit logs.
The architecture is deliberately compute-agnostic. Teams can plug Akave into AWS, Google Cloud, on-premises data centers, or the emerging crop of "neoclouds" without touching application code, at least in theory. Object versioning, locking, and bring-your-own-key encryption sit at the protocol level. The company claims "11 nines" durability through erasure coding—a figure that, if accurate, would match or exceed most hyperscaler guarantees.
An optional Filecoin-backed archiving tier, rolled out late last year, adds verifiable long-term storage for compliance-heavy workloads. And Akave has quietly integrated with Apache Iceberg and earned a spot in Snowflake's external storage documentation alongside other S3-compatible providers, a nod that suggests at least some enterprise traction beyond crypto-native use cases.
Early Traction, Familiar Customers
The customer list so far tilts toward data-intensive niches. Intuizi uses Akave for consumer intelligence datasets. LaserSETI stores astronomical observations on the platform. 375ai, focused on edge-collected AI training data, is another early adopter, as is Skymapper, a telescope data outfit.
None of these names will move the needle for CIOs at Fortune 500 companies—yet. But they do signal the kinds of workloads Akave is targeting: high-volume, high-churn environments where storage costs and data mobility matter more than legacy vendor relationships.
The seed funding will primarily go toward sales and marketing, a necessary step if Akave wants to graduate from design partnerships into broader enterprise adoption. The company reports a team of roughly 15, with about 80 percent in engineering roles—a lean operation by Silicon Valley standards, though perhaps appropriate for a startup still proving product-market fit.
Middleware Ambitions in a Decentralized Stack
Akave has woven itself into a broader decentralized infrastructure ecosystem—partnering with Akash Network for GPU compute, Baselight for queryable data marketplaces, and tapping the Filecoin network for archival storage. The bet, essentially, is that it can serve as middleware between AI workloads and a more modular, less centralized storage layer.
Whether that vision scales beyond crypto-adjacent customers is the lingering uncertainty. Blockchain-based infrastructure has promised to disrupt cloud incumbents before, often with more fanfare than follow-through. And enterprises, particularly those handling sensitive data, tend to move cautiously when it comes to experimental storage architectures—no matter how attractive the pricing.
Still, the investor roster suggests real appetite for alternatives to what Vervaet calls the "hyperscaler oligopoly." Protocol Labs and Avalanche Foundation both have strategic reasons to back Akave: the former because of Filecoin integrations, the latter because Akave runs on its blockchain. But funds like Blockchange and Lightshift are placing broader bets on decentralized infrastructure gaining enterprise legitimacy.
For now, Akave is positioning itself as the pragmatic middle ground—S3-compatible enough to feel familiar, decentralized enough to offer optionality, and priced aggressively enough to get meetings with cost-conscious engineering teams. Whether that's enough to carve out meaningful market share in a storage landscape still overwhelmingly controlled by three companies remains to be seen.
But in a market where data gravity and egress fees keep customers locked in, even incremental gains in portability might be worth $10 million and change to find out.
