What Happened

Ollama, the open-source tool that lets developers run AI models locally, has raised a $65 million Series B led by Theory Ventures. Founder and CEO Jeff Morgan confirmed the round to TechCrunch, bringing the company's total funding to $88 million following a $15M Series A led by Benchmark's Peter Fenton.

The company reports 8.9 million monthly developers, 176,000 GitHub stars, and presence in 85% of Fortune 500 companies — achieved with just 14 employees. Morgan declined to disclose revenue or valuation. Ollama also operates a neocloud with subscription tiers from free to $100/month, pricing by GPU time rather than token limits.

The founders, Morgan and Michael Chiang, previously built Docker Desktop after Docker acquired their startup Kitematic. Ollama applies the same containerization philosophy to AI models: abstracting away hardware configuration so developers can run open-weight models in minutes.

Why It Matters

This round closes a two-week stretch of major funding in open-source AI infrastructure. Together AI raised $800M at an $8.3B valuation. RunPod raised $100M for its AI developer cloud. Venice AI raised $65M at a $1B valuation for private, uncensored AI. Ollama's raise confirms that investors are betting heavily on the open-weight model ecosystem as a structural shift, not a niche.

The strategic differentiator is Ollama's dual model: free local execution plus paid cloud inference. Morgan says the proving point came around January 2026, when open models demonstrated real agentic capabilities — particularly coding tasks. That shifted the conversation from "open models are good enough for experimentation" to "open models can do production work," which directly drives demand for Ollama's neocloud.

Fenton argues the open-vs-closed debate is misframed: enterprises will use both, but every company with high inference costs has what he calls a "vital existential project" to move workloads to open-weight models. Ollama is positioning itself as the on-ramp for that transition.

The commercialization hasn't been frictionless. Roughly a year ago, community posts criticized Ollama for "enshittification" — the perceived degradation of free tooling as the company prioritized its cloud business. Morgan and Fenton both insist the free desktop product remains unchanged, but the tension reflects a broader challenge for open-source AI companies navigating monetization.

Who Is Affected

AI application-layer startups facing rising inference costs on closed models now have another funded neocloud option, with Ollama's GPU-time pricing model offering a different cost structure than token-based providers.

Enterprise IT teams — given the 85% Fortune 500 penetration figure — are likely already running Ollama somewhere in their environment, possibly outside official procurement channels. This funding means a formal vendor relationship is now available.

Open-source developers who depend on Ollama's free local runner should monitor whether the 14-person team can maintain the core project while expanding cloud operations.

Strategic Implications

For AI startup founders: Ollama's neocloud now has enterprise-grade backing and a pricing model worth benchmarking against Together AI, RunPod, and Venice AI. The GPU-time pricing structure may be advantageous for workloads with high token volume but predictable compute patterns. The broader signal: open-weight model infrastructure is attracting serious capital, which means pricing pressure on closed-model APIs will intensify.

For developers/operators building with AI APIs: The local runner remains free and unchanged per Morgan and Fenton. The strategic question is whether to standardize on Ollama's neocloud for hosted inference or maintain portability across multiple providers. Given that Ollama supports model discovery and deployment across both local and cloud, it's becoming a workflow layer — not just a runtime.

For non-technical business owners: Ollama's Fortune 500 penetration suggests open-weight models are already running inside your organization. This funding round means you'll soon have a vendor relationship option for what may currently be shadow IT usage. Worth asking your engineering teams what they're running and where.

What to Watch Next

Monitor whether Ollama expands its 14-person team significantly with this funding — that will indicate whether they're scaling the neocloud infrastructure or investing in enterprise sales. Also watch for pricing changes in the neocloud tiers and any shifts in the open-source project's release cadence, which would signal whether commercial priorities are pulling engineering attention.

Frequently Asked Questions

Q: What is Ollama and what does it do?

A: Ollama is an open-source tool that lets developers run open-weight AI models on their local machines, getting them running in minutes without complex hardware configuration. It also offers a neocloud service for running larger models that can't fit on local hardware, with subscriptions from free to $100/month.

Q: How does Ollama's neocloud pricing work?

A: Ollama charges based on GPU time rather than token limits, which differentiates it from most closed-model API providers. This means costs scale with compute usage rather than text volume, which may be more cost-effective for certain workload patterns.

Q: Is Ollama's local tool still free?

A: Yes. According to Morgan and board member Peter Fenton, the free desktop product remains unchanged. The neocloud is an additional paid service for models too large to run locally.