In partnership with

The Neocloud Summit: We built a room that has never existed before

In under three years, neoclouds went from an experiment to critical AI infrastructure, delivering compute, speed, and specialization the incumbent clouds could not. The category is now entering its most consequential chapter. Workloads are shifting from training to inference, enterprise buyers are arriving in force, and the constraints have moved from chips to power, capital, and architecture. The decisions that will separate durable operators from the rest are being made right now, by a small group of people.

On October 8, at Hotel Kabuki during SF Tech Week, we are putting that group on one stage.

The NeoCloud Summit is exclusive by design. Not a tradeshow with badge-scans, It is a single main stage and a curated audience, the operators, chip architects, inference providers, enterprise buyers, and the capital financing the buildout, in one room for one day. Building on the Ignite ecosystem of AI infrastructure players, this room has never existed before.

Not a room for surface-level conversations

Every session on this agenda goes deep into the problems neoclouds and their ecosystem actually wrestle with.

The morning opens with State of the NeoCloud, mapping where demand is heading with WEKA, Silicon Data, and Remanie.ai, moderated by Supermicro. AMD follows with The Neocloud Advantage in the Agentic Era, on why a single agentic request now spans multiple models, tools, and verification steps, why memory capacity and bandwidth matter more as concurrency grows, and why coordinating CPUs, GPUs, memory, networking, and software as one system is becoming a competitive advantage.

From there it stays technical. Powering the NeoCloud puts operators on three continents on stage, Jazz Computing, Moonshot, and Panadina, with iMasons pressing on whether the real bottleneck is power, permits, or packets. Training vs. Inferencing brings GMI Cloud and Massed Compute on where compute spend is actually going. Tensor Machines and Cato Digital present The Real Cost of a Token, an open benchmark for what a token actually costs to produce. The morning closes on The NeoCloud Security Gap, the multi-tenant isolation and trust conversation most fast-scaling clouds skip, with Cisco.

The afternoon goes deep on inference

The Inference Stack traces where margin lives from silicon to serving models, with SambaNova, Digital Ocean, and Together.ai. Upscale AI keynotes on running any accelerator fleet as a single system. When Does GPU Capacity Become an AI Cloud? gets specific on build-vs-buy with Infrinia, Zyphra, and Qumulus.

Then the sharpest technical session of the day: Splitting the Brain: Disaggregated Inference with AMD and Cerebras. Prefill and decode want different things from hardware, and cramming both onto one chip wastes silicon and money. AMD and Cerebras break down running each half on hardware built for it, and what that means for anyone choosing inference infrastructure, moderated by Lightning.ai. Vultr follows on the edge advantage of a globally distributed GPU cloud.

The back half closes the loop on the business of inference, Enterprise Tokenomics with ServiceNow, Transunion, and WEKA, Inferencing + Bare Metal on what inference providers actually demand from a GPU partner with Compute Exchange and Prem.ai, and Private Capital Finance, where Morgan Stanley and Cisco talk about when compute becomes bankable.

Networking: Where the deals get done

The stage sets the agenda. The room closes it. Lunch, the expo, and the closing happy hour from 4:30 to 6:30 are engineered for the conversations that turn into partnerships, pilots, and signed deals. In a curated room, the person next to you is a potential partner, customer, or investor, not a badge in a crowd. Every operator in the sessions above is in the room the rest of the day.

This is where the next cloud gets built.
If you want a seat, now is the time.

Bay Area Startups Collectively Secured $15B+ in September MTD

More deals, smaller amounts sums up the week in SV Funding. AI infrastructure companies dominate our ten featured fundings this week. 
The projected cost of the U.S. AI infrastructure build-out has increased steadily over the last year. A new Brookings Institution paper Financing the AI Buildout projects US AI investment in data center buildings, power systems, networking infrastructure, specialized chips and other equipment will total $10.3T between 2025 and 2032, averaging 3.6% of GDP per year. That's the largest single-industry build-out ever as a share of GDP — greater than that for the canals, railroads and grid combined.

Startups: Looking for partners, keeping a watch for possible competitors? LinkSV's Pulse of the Valley weekday newsletter and alerts can keep you on top of new companies launching; database keyword search gives you prospects for partners and customers. Founders get the newsletter, and alerts for just $7/month ($50 value). Check it out and sign up here.

Follow LinkSV on LinkedIn to get early notice of new tech, stay on top of SV funding intelligence, and keep track of the companies, investors and executives impacting the startup ecosystem.

Early Stage:

• DensityAI closed a $200M Series A, stealth, building the fastest inference solution for frontier models.

• Watney Robotics closed a $80M Series A,  the frontier AI data lab, helping teams build the data and environments behind high-performing frontier and agentic AI.

• Horowitz Andreessen Academy closed a $42M Series A, a highly selective school in San Francisco for the most ambitious young builders out of high school

• PicoJool closed a $27.5M Series A, developing next-generation optical chips and modules for high-bandwidth, low-cost connectivity in hyperscale AI data centers.

• Petrarch closed a $.5M Pre-Seed, brings internal company data to frontier labs, starting with manufacturing and industrial companies.

Growth Stage:

• Snorkel AI closed a $350M Series E, the frontier AI data lab, helping teams build the data and environments behind high-performing frontier and agentic AI.

• Micro1 closed a $100M Series C, building the infrastructure for advancing intelligence.

• Firecrawl closed a $75M Series B, the web data toolkit for AI.

• BigHat Biosciences closed a $75M Series C, an AI-native biotechnology company designing and developing next-generation protein therapeutics.

• Kairos Power closed a $70M Corporate, focused on commercializing the fluoride salt-cooled, high-temperature reactor to enable deep decarbonization

Event Spotlight: AI for FSI NYC

Financial institutions have no shortage of AI pilots. The harder part is getting those systems into production while maintaining control over cost, latency, data, security, governance, and regulatory requirements.

That was the focus of our AI for FSI event in New York, hosted by AMD and Supermicro.

The day brought together leaders across financial services and AI infrastructure to look at what it actually takes to move from experimentation to production.

Why this, why now?
The conversations stretched across the stack. Market data independence and vendor lock-in. On-prem and hybrid AI factories. GPU capacity and shared clusters. Data provenance and auditability. Inference infrastructure. Token economics and how enterprises should actually measure AI ROI.

Flagstar Bank shared how modernizing its data environment helped take one CECL reporting process from three weeks to two minutes. Other discussions explored the importance of citations and point-in-time data for trustworthy AI outputs, managing finite GPU capacity across teams, and deciding when renting or owning infrastructure makes economic sense.

One idea kept resurfacing throughout the day: financial institutions need choice and accountability as they scale AI.

That means portable infrastructure, governed data, clear provenance, fit-for-purpose compute, and ROI measured against business outcomes.

Thank you to everyone who joined us in New York and contributed to the conversation.

Your Feedback Matters!

Your feedback is crucial in helping us refine our content and maintain the newsletter's value for you and your fellow readers. We welcome your suggestions on how we can improve our offering. [email protected] 

Logan Lemery
Head of Content // Team Ignite

Some teams never seem to stop moving. They're on Attio, the agentic CRM.

Every customer signal is captured in one shared context layer, always current and compounding. Agents and workflows build pipeline, chase every buying signal, and move deals forward, an always-on revenue engine running alongside your team.

With Attio, you’ll get:

• Leads automatically prioritised and routed to the right rep

• Expansion and risk signals caught the moment they land

• Follow-ups written in your voice, already there when you arrive

Teams like Parallel, Turbopuffer, and Wordsmith build on Attio. Are you one of them?

Get started now