Emerging
May 28, 20261
50%
General Compute raises $15 million to compete in AI inference market with SambaNova chips
General Compute, a new AI inference cloud platform, raised $15 million in seed funding to deploy specialized SambaNova chips in data centers. The company addresses bottlenecks in securing appropriate inference chips and deploying them efficiently, positioning itself in a competitive market where specialized architectures are increasingly seen as superior to GPUs for AI inference workloads.
Quick Facts
Who
General Compute
What
Raised $15 million seed funding
When
2026-05-28
Where
Silicon Valley
- Raised $15 million seed funding
- Launched AI inference cloud offering
- Ordered $300 million of SambaNova SN50 chips
- Pursuing colocation deals with data center providers and crypto miners
- Claimed fastest performance running MiniMax 2.7
General Compute, a new inference cloud startup that rents AI processing power, has raised $15 million in seed funding at a $60 million post-money valuation, led by FUSE VC with participation from Carya Venture Partners and Village Global Ventures. The company addresses two critical challenges in the AI ecosystem: securing the right chips for inference and deploying them in data centers efficiently.
While GPUs have dominated AI computing, industry consensus is shifting toward specialized chips designed specifically for inference—the phase where trained models respond to user queries rather than undergoing training. General Compute co-founders Finn Puklowski and Jason Goodison have turned to SambaNova, an Intel-backed chipmaker, after facing capacity constraints at established players like Groq and Cerebras. SambaNova's forthcoming chips are claimed to generate 600 to 700 tokens per second compared to approximately 250 tokens per second for GPUs, while requiring less power and air-cooling instead of water-cooling.
General Compute has ordered $300 million worth of SambaNova's SN50 chips and will be the first neocloud to deploy them. The air-cooled architecture enables installation in existing data center infrastructure without requiring new construction. The company is pursuing colocation arrangements not only with traditional data center providers but also with cryptocurrency miners seeking to repurpose their facilities as bitcoin production costs have exceeded prices.
The startup launched its cloud offering last week and claims it is already the fastest at running MiniMax 2.7, an open-source large language model. Venture investor Joe Hasselmann, who backed Groq in 2021 and recently launched the AI-focused Evercrest Capital Partners fund, sees parallels between General Compute's SambaNova partnership and CoreWeave's relationship with Nvidia. The competition reflects broader questions about which chip architecture will dominate AI's future, with inference clouds representing a bet on a fragmented market of multiple models and agents where speed and cost efficiency are paramount.
Topics
Why This Matters
General Compute's funding and SambaNova partnership signal a fundamental shift in AI infrastructure from GPU-dominated computing toward specialized inference chips. For enterprises and cloud consumers, this competition creates immediate cost and performance implications: specialized architectures promise 2.4–2.8× faster token generation and lower cooling costs than GPUs. The company's pursuit of colocation partnerships with cryptocurrency miners demonstrates how emerging AI economics are reshaping underutilized infrastructure globally. Investors and technologists should watch whether fragmented inference chip markets (SambaNova, Groq, Cerebras) can sustain competitive advantage against Nvidia's entrenched GPU ecosystem, as this outcome determines pricing, availability, and vendor lock-in across the AI supply chain.
Timeline & Sources
Jan 1, 2021
WireJoe Hasselmann invested in Groq
Jan 1, 2026
WireJoe Hasselmann launched Evercrest Capital Partners fund
Jan 1, 2026
WireSambaNova to release new inference chips
May 22, 2026
WireGeneral Compute launched cloud offering
May 28, 2026
WireGeneral Compute announced $15 million seed funding