8090 Solutions Raises $135M to Build Software With AI Agents

8090 Solutions has raised $135 million in a round led by Salesforce. The Redwood City startup, founded in 2024 with Chamath Palihapitiya as co-founder and CEO, builds enterprise software using coordinated teams of AI agents working under human-led oversight. The thesis is blunt: custom enterprise software should be built faster and cheaper by orchestrated agents than by traditional dev shops, and Salesforce is putting strategic money behind it.

• $135M led by Salesforce, a strategic investor whose own products would compete with or consume this model.
• The approach is coordinated agents under human oversight, not a single autonomous coder, closer to an AI-run software factory.
• It targets custom enterprise software, the expensive, slow-to-build internal apps large companies commission constantly.
• Chamath's involvement guarantees attention; the round's real signal is Salesforce validating agent-built software.

What does 8090 actually build?

8090 sells outcomes, not a chatbot. Enterprises constantly need bespoke internal software, workflow tools, integrations, dashboards, line-of-business apps, and traditionally they pay consultancies or in-house teams to build it slowly and expensively. 8090's model puts a coordinated set of AI agents on that work, each specialized (planning, writing code, testing, reviewing), with a human lead directing the effort and accountable for the result. The bet is that this pipeline delivers working software dramatically faster than a conventional team, at a fraction of the cost, without the quality collapse pure autonomy tends to produce.

RelatedPoetic Raises $50M to Build Software That Learns

Why did Salesforce lead the round?

Strategic investors put money where their roadmap points. Salesforce has staked its future on Agentforce, its own agentic platform, so backing 8090 is both a hedge and a signal: the company that popularized enterprise SaaS believes agent-built custom software is a real market. It also gives Salesforce a window into how coordinated agents perform on production workloads, intelligence worth more than the check. For 8090, a Salesforce-led round buys credibility with exactly the enterprise buyers it needs, and potential distribution into the largest CRM ecosystem on earth.

Is "human-led oversight" the key detail?

Yes, and it is what separates this from the autonomous-coding hype. Fully autonomous agents remain unreliable on real enterprise software: they hallucinate requirements, miss edge cases, and produce plausible code that fails in production. 8090's framing, agents do the volume, a human owns the judgment and the sign-off, is the pragmatic middle path that enterprises can actually buy, because someone remains accountable. It is a bet that the near-term winner is not "no humans" but "far fewer humans, each far more leveraged."

What it means for the market

The signal for investors is that capital is flooding the layer where AI does labor, not just the models. This round sits alongside a July that saw Together AI raise $800M and a broad tilt toward AI infrastructure and tooling. For the incumbent consultancies and custom-dev shops, 8090 is a direct threat to the billable-hours model; for Salesforce (NYSE: CRM), it is a call option on the agentic-software thesis it is already selling. Watch whether 8090 publishes real delivery metrics, cost per app, time to ship, defect rates, because agent-built software lives or dies on whether the output survives production, not on the pitch. Track the round and the wave it belongs to on our Funding Tracker and the ranked Biggest AI Funding Rounds.

RelatedLinqAlpha Raises $22M to Put AI Agents on Market Research

How is this different from AI coding tools?

Copilot, Cursor, and Claude Code assist a human developer who stays in the driver's seat writing software. 8090 inverts the ratio: the agents do the bulk of the building and the human supervises, closer to running a software agency than using an autocomplete. That distinction matters commercially, because it changes what 8090 sells. It is not a per-seat developer tool competing on productivity; it is an outcome, a finished application, competing with consultancies on cost and speed. The risk profile differs too. A coding assistant's mistakes get caught by the developer using it; an agent-built app's mistakes have to be caught by review and testing, which is exactly why the human-oversight layer is load-bearing rather than decorative.

Our take

Strip away the Chamath star power and this is a serious bet on the most defensible version of agentic software: agents for throughput, humans for judgment. That framing is why Salesforce, of all backers, led, it validates the exact thesis Salesforce is selling with Agentforce, and it gives 8090 a credibility and distribution edge most AI startups would kill for. The risk is the oldest one in enterprise software: shipping a demo is easy, shipping software that survives real production, with security, edge cases, and maintenance, is hard, and that is precisely where autonomous coding still breaks. 8090's answer is to keep a human on the hook. If the delivery numbers hold up, this is a template for how custom enterprise software gets built for the next decade. If they do not, it is another well-funded reminder that judgment does not automate cheaply.

Original analysis by GenZTech. Figures current as of July 2026. Source: Tech Startups.