Our $445M Series D

Original Article ↗ Hacker News Discussion ↗

A hardware–software startup building fully integrated, open source server racks for on‑prem “cloud-like” infrastructure has raised a $445M Series D at multi‑billion‑dollar scale, impressing many with its profitability and capital strategy. Commenters debate its choice to raise equity rather than debt, the implications of repeated large funding rounds for dilution, and how it plans to address GPU‑heavy AI workloads. The company’s unusually intensive, writing‑heavy hiring process and lack of transparent pricing draw mixed reactions, but overall sentiment is strongly positive about its technical ambition and potential to offer an alternative to hyperscale cloud providers.

Overall sentiment

  • Many commenters are enthusiastic, calling the company inspiring, well-run, and unusually thoughtful, especially for a hardware startup.
  • Others remain skeptical about aspects like hiring, pricing, lack of GPUs, and long‑term differentiation, but still treat the business as serious and credible.

Business performance, funding, and risk

  • Thread notes rapid fundraising cadence (Series B, C, D) and a large backlog of orders.
  • Company is reported as profitable and paying income tax, which several see as remarkable for a startup.
  • Debate: why raise equity instead of using loans or trade finance for inventory.
    • Pro‑equity side: debt can kill a volatile hardware startup if demand collapses; equity spreads risk, especially given memories of prior hardware crashes.
    • Pro‑debt side: if capital is mostly to fulfill existing orders, short‑term financing might be cheaper than selling more equity.
  • One linked source gives a ~$6B valuation; commenters note dilution vs. growth trade‑offs and discuss how new shares affect existing equity.

Product, positioning, and cloud vs on‑prem

  • Multiple explanations converge: it’s a fully integrated rack (compute, storage, networking, power, cooling) plus firmware and cloud‑style control plane, shipped to customer premises or colo.
  • Emphasis on hardware–software co‑design, elimination of traditional BIOS/UEFI layers, and open‑source software.
  • Customers get an on‑prem “cloud-like” experience with APIs/console akin to hyperscalers, but under their control.
  • Several commenters say they are actively looking to exit AWS/Azure/GCP due to cost, lock‑in, and perceived quality decline; Oxide is seen as a candidate vendor.
  • Others question why to choose this over established vendors (Dell/HPE/mainframes) and whether support track record and pricing are competitive.

GPUs, AI, and compute mix

  • Some say “no GPU support” is a hard sell in the current AI‑heavy landscape.
  • Others counter that general CPU compute remains critical for orchestration, agents, and non‑ML workloads, and that the size of this funding round shows GPU‑less racks can still sell.
  • Linked FAQs and podcast clips suggest GPU support is on the roadmap, especially given AMD’s investment, but specifics remain high‑level.

Hiring process and candidate experience

  • The hiring pipeline features extensive written materials and up to ~9 hours of interviews split into multiple blocks.
  • Supporters argue:
    • Written materials are the primary filter, so relatively few people reach interviews.
    • A high fraction of interviewees receive offers.
    • Long, conversational interviews (with pre‑shared written materials) avoid repetitive “elevatorpitching” and select for strong communication and cultural fit.
  • Critics argue:
    • The up‑front writing is a large unpaid time investment, especially for applicants applying widely or with other obligations.
    • Nine hours of interviews per company scales poorly in a tough job market.
    • Some report slow or missing responses despite promises to reply to every application, which feels disrespectful when applications are so involved.
  • There’s an extended meta‑discussion: some prefer rigorous, time‑intensive processes for “dream jobs”; others see them as exclusionary and indicative of poor respect for candidates’ time.

Pricing, target market, and accessibility

  • There is no public pricing page; prospective customers must contact sales.
  • Comments indicate racks start around roughly the seven‑figure range.
  • Some argue that at this scale, bespoke quoting is normal and comparable to other enterprise vendors.
  • Others say lack of transparent pricing prevents engineers from even knowing if it’s worth initiating contact or experimentation.

Homelab / smaller-scale offerings

  • Many wish for a small “homelab” or SOHO version (“mini Oxide”), even knowing it’s unlikely.
  • Several point out that much of the software stack (e.g., control plane) is open source and can be run on commodity or second‑hand hardware, though this loses the hardware–software co‑design benefits and is non‑trivial to set up.

Culture, communication, and DEI

  • Commenters praise the blog, podcasts, and technical communication as unusually clear and engaging.
  • The equal‑salary policy and transparent principles are viewed by several as “refreshingly principled.”
  • Some discussion arises around inclusive language (e.g., “you guys”) and cultural expectations:
    • One side welcomes explicit inclusivity and sees it as aligned with the company’s stated values.
    • Another side feels such norms could make the environment unwelcoming for people whose speech habits differ, perceiving potential over‑correction or rigidity.
  • Overall, culture is seen as a strong filter: attractive to some precisely because it’s opinionated; off‑putting to others for the same reason.