In partnership with

The AI Infra Market Is a Dependency Graph. Now You Can See It.

The AI compute market is growing faster than anyone can track. If you're new to the industry, the hard part is figuring out who the players are and how they fit together. If you've ever stood up GPU capacity, the hard part is that a deployment can stall on transformer lead times, a liquid-cooling retrofit, an optics shortage or a grid interconnect queue that's years deep. Both problems have the same cause. The market is split across layers, companies and headlines that change every week.

NeoNews is Ignite's answer to both, and our next step toward becoming the source of truth for AI infrastructure. At its core is an interactive map of more than 1,000 companies across eight layers of the AI infrastructure stack, from power all the way up to LLMs. Each company is placed by what it actually does. The map also shows how companies relate to each other: who supplies whom, who partners, who competes and who is financing whom.

If you're new to the space, the map is the fastest way to get your bearings. You don't have to piece the market together from press releases and podcasts, because you get one structured view of the whole landscape. You can see the major players in each layer and why power and cooling matter as much as chips. You can also see how a model company depends on suppliers it may never mention by name.

If you build and run this infrastructure, the relationships are the useful part. The stack works like a dependency graph, not a list of vendors. Take a neocloud deciding where to expand. It isn't just comparing GPU pricing. It needs to know which colo providers have powered capacity coming online, which cooling vendors can handle high-density racks and which suppliers are already committed to someone else. When you see the layers as connected, you can ask better questions. Where is the concentration risk? Which single supplier sits under half the market? Where is there a gap nobody has filled yet?

On top of the map is a daily intelligence layer. Users get a five-minute briefing on the developments that matter most, which they can filter by layer and priority. Custom watchlists track competitors, suppliers, funding rounds, capacity announcements and partnerships. For newcomers, that's an easy daily habit for learning the market. Operators can follow the layers they depend on without reading every press release in the industry. Either way, hours of research become something you can take in over a cup of coffee.

Then there's the question every technical reader asks: where does the data come from? NeoNews is built on Ignite's own database. It comes from years of running events with the operators, builders, suppliers and investors on the map and working with them directly. That relationship data is combined with public industry data that is aggregated continuously. The result is a map built by people who know the companies, not a scrape of logos. It also makes NeoNews a proof point. Mapping, organizing and monitoring this market in real time takes knowing it well.

Billions of dollars are flowing into AI compute, and as that number grows, the stack only gets denser and more interdependent. Whether you're learning the landscape or building inside it, you need a clear picture of how the pieces fit together and what's changing underneath them. NeoNews is built to be that picture.

Explore the map.

Bay Area Startups Collectively Secured $9.7B+ in October MTD

Startup funding surged this week to $9.7B, almost doubling the total for October. Ten megadeals for $3.9B anchored the week, including AI infra companies Tenstorrent ($1.5B), Flapping Airplanes ($1B) and Peak Energy ($108M). Megadeal rounds also went into life sciences (Warp Sciences, $215M) , manufacturing and robotics (Arda Industries, $160M).

Exits: It was a slow week for M&A, with only two priced acquisitions - Salesforce's $5.6B acquisition of Braze, and Kapital's acquisition of Woz. On the IPO front, despite other companies pausing their offerings, Anthropic is pushing ahead and is now expected to go out to investors in the second week of November. The company is still looking for a $2T valuation, despite some investor pushback following recent disclosures that revenue growth is slowing.

Startups: Looking for investors, partners, possible competitors? Best deal in the valley is the Pulse of the Valley weekday newsletter and alerts – get a daily summary of just the activity that is most important to you: prospects for investors, partners and customers. Founders get the alerts and newsletter 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 trends, companies and investors impacting the startup ecosystem.

Early Stage:

• Warp Sciences closed a $215M Series A, a life sciences infrastructure company that integrates AI and advanced robotics to automate small-molecule discovery chemistry.

• Dream Labs closed a $90M Seed, an AI company building the brain for everyday robots.

• Maxinsights closed a $30M Seed, they are building the datalayer for physical AI.

• Preference Model closed a $16M Seed, they develop reinforcement learning environments for training AI systems to perform machine learning research.

Growth Stage:

• Tenstorrent closed a $1.5B Series E, an AI compute company that builds RISC-V-based AI processors and systems for developers, enterprises, and sovereign infrastructure worldwide.

• Flapping Airplanes closed a $1000M Series A, a foundational AI research lab devoted to solving the data efficiency problem.

• Peak Energy closed a $108M Series B, delivering low-cost, giga-scale energy storage purpose-built to meet today's grid challenges.

• Blacksmith closed a $66M Series C, builds infrastructure and software for validating code.

WEKA: More Tokens From Your GPUs

Every turn of an agentic session sends the full conversation history back to the model, and GPU memory fills up fast. Once the KV cache is evicted, GPUs spend their time recomputing context they already processed, which raises time to first token and cost per token. Inference now pays the bills, so the memory wall has become a margin problem.

For NeoClouds, the priority is getting more served tokens out of the GPUs already racked and powered, not buying more of them. That is where WEKA fits.

WEKA is the AI data and memory infrastructure company behind NeuralMesh, a software platform that combines high-performance storage with extended GPU memory. Its Augmented Memory Grid turns pooled NVMe into a shared KV cache tier. In an eight-hour test on Nebius AI Cloud running NVIDIA HGX B300, it served 2.4x the requests with a 93% cache hit rate, versus 40% for GPU memory alone. NeuralMesh 6, announced in July, adds native multi-tenancy, unified file and object storage on NVMe, and always-on data reduction. The third-generation WEKApod systems deliver 1.1 exabytes of effective capacity in a single rack.

Who they serve

AI clouds, frontier model builders, government agencies and enterprises, including Nebius, Together AI, TensorWave, Firmus and Andromeda. WEKA says it is trusted by 30% of the Fortune 50.

At NeoCloud Summit, CEO Liran Zvibel said inference rewards whoever is "effective, nimble and cost effective." Chief AI Officer Val Bercovici pushed the room toward "cost per task, cost per outcome." WEKA is building its platform around that scorecard.

Explore WEKA to learn how NeuralMesh can raise token throughput on the GPUs you already own.

// NEXT UP

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

Your SOC 2, handled end to end with AI agents.

Enterprise buyers won’t put your product near their customer data without a SOC 2 report. Twenty-one state privacy laws now hold them accountable for the vendors they share data with, so their diligence lands on you.

Sprinto gets you audit ready in 14 days, across three working sessions. AI agents connect your stack, collect the evidence auditors ask for, and close the gaps as they appear. You approve, they execute.

Sprinto also answers the security questionnaires your buyers send, using the same evidence base, so security reviews stop holding up deals.

No compliance hire. No consultant. Your auditor signs off.

Learn more and get $1,000 off.