Pomegra
Phaze, built by the Parsec founders, raised an $8.75M seed led by a16z in 2026 to give employees and AI agents shared access to powerful remote computers.
Key Takeaways
- Phaze raised an $8.75M seed round led by a16z on November 4, 2026.
- Indie, NextView, Makers Fund and Lerer Hippeau also invested; valuation is undisclosed.
- The founders previously built Parsec, which Unity bought for $320M in 2021.
Lead
Phaze, a remote-computing startup started by the team behind Parsec, announced an $8.75 million seed round on November 4, 2026. Andreessen Horowitz (a16z) led the round. Indie, NextView, Makers Fund and Lerer Hippeau also took part. The company did not disclose a valuation.
Phaze sells high-performance remote desktops that companies can share across staff and across AI agents, rather than buying a dedicated workstation for every person. CEO Benjy Boxer and co-founder Chris Dickson ran Parsec, the remote desktop tool for graphics-heavy work that Unity acquired for $320M in 2021.
What Does Phaze Actually Sell?
Phaze sells remote access to powerful machines for demanding professional workloads, now with agents treated as a second class of user. People connect to a remote workstation from any device. Agents connect through a local Model Context Protocol (MCP) server and operate the machine through the desktop or the command line.
The company also supports shared sessions, where a person and an agent work in the same remote session at once and hand control back and forth. Agents can run in the background without touching the user's own computer. Enterprise authentication and centralized access controls sit on top, with user-level permissions limiting what an agent can reach.
One early use case is an IT support bot. It takes requests in Slack, logs into the employee's machine, tries to fix the problem and escalates to a human when it cannot.
Why Would a16z Back This?
a16z backed Phaze because it sees expensive hardware sitting idle as a cost problem that agents make worse. General Partner Martin Casado framed it in the funding announcement: "High-end compute is too expensive to sit idle on individual desks. Phaze is turning those machines into shared infrastructure for people and AI agents."
The pitch rests on a familiar inefficiency. A workstation assigned to one person is busy for part of the day and idle for the rest. Pooling those machines raises utilization, and agents add a new source of demand that does not follow office hours.
For a seed investor the team matters as much as the thesis. Boxer and Dickson already shipped low-latency remote desktop software at scale and sold it for a figure well above this round. That record carries more weight than the market-size claims in most seed decks.
How Does This Round Compare With the Competition?
The round is small next to the money chasing agent infrastructure, and the competition is crowded. Baseten acquired Blaxel, which builds sandboxes for AI agents. Perplexity has built an agent platform around Mac minis. Apple's Mac mini sold out in the US during spring 2026, which points to agent demand pushing buyers toward dedicated hardware.
Most agent sandboxes are headless containers built for code execution. Phaze's bet is that many valuable tasks still require a full graphical computer, with real applications and a human who can step in. That is a narrower claim than "computers for agents," and a more defensible one, because it builds on latency work Parsec spent years refining.
The risks are plain. Customers are unnamed in the announcement, and the company has not published revenue or usage figures. Large cloud providers and virtual desktop vendors already sell remote workstations, and they could add agent features. Phaze also depends on enterprises trusting agents with logged-in desktops, a security question that most buyers have not settled.
What Comes Next for Phaze?
Phaze said it will spend the money on engineering hires, product development and customer growth. The first test is whether the shared-session model, where humans and agents work side by side, becomes a reason to buy rather than a demo feature. Named enterprise customers would be the clearest evidence.
A follow-on round will depend on whether usage by agents grows faster than usage by people. If agents drive most of the demand, Phaze looks like agent infrastructure. If people do, it looks like a well-funded successor to Parsec with an AI feature set.
Outlook
Phaze enters a market where agents need real computers and companies want to stop paying for idle ones. It has a proven founding team, a lead investor with a clear thesis, and $8.75M of runway. It lacks disclosed customers or a valuation, and larger rivals can copy the pitch. The next 12 months should show whether agent demand for full desktops is a lasting category or a niche.
