Everyone who’s anyone in the crazy artificial intelligence world attended the RAISE Summit in Paris this week. Bigger, noisier and more crowded than last year’s installment, this year’s summit featured a level of buzz I haven’t seen since the dot-com days.

And that buzz? It’s all about the money.

What I found at RAISE was a furious crescendo of investment. Billions of Euros and dollars flooding into data centers as neoclouds rush their offerings to market. Software vendors furiously pivoting from whatever business model worked in the old days of 2025 to some newly found AI religion. And the enterprises struggling to keep up, delicately balancing investment in innovation against traditional concerns about security and compliance risk.

The risk most evident at RAISE was the risk of FOMO – fear of missing out – as participants, exhibitors and speakers all proclaimed the NEXT BIG THING.

At least, until the AI bubble bursts. But in the meantime, here are my hotspots on the show floor:

Feeding the neocloud beast

Neoclouds – data centers offering AI-specific cloud capabilities – have always been at the center of RAISE (see my coverage of last year’s summit for some neocloud highlights).

This year, the neoclouds are running full steam ahead, doing whatever they can to get their offerings into the market as quickly as possible. To meet this explosive demand, several vendors at RAISE provided offerings that focus on supporting this massive investment in the neocloud marketplace.

One of the most significant bottlenecks slowing down neocloud growth is the constraint on graphics processing units or GPUs. To address this limitation, Clockwork Systems Inc. offers a software-driven AI fabric that focuses on maximizing GPU utilization, primarily for neoclouds.

Clockwork’s innovations leverage familiar capabilities in new ways to optimize neocloud’s infrastructure: high-availability fault tolerance that takes its lessons from redundant array of identical disks or RAID technology, as well as leveraging observability technologies to maintain precise clock synchronization (essential for the proper operation of any cloud) and technology reminiscent of software-defined wide-area networking or SD-WAN for resolving network congestion issues within cloud infrastructure.

Perhaps the most blatant example of the rush to neocloud investment, however, is SambaNova Inc. Fresh off a billion-dollar raise at an $11 billion valuation, SambaNova offers prebuilt racks of servers leveraging its next-generation chip technology that powers both AI training and inferencing. Neoclouds can simply drop these racks into their data centers and they’re off to the races.

The pivots: Playing Wayne Gretsky

Several vendors at RAISE have decided their best bet was to skate to where the proverbial puck is going, pivoting from whatever they were doing before to an AI-centric offering.

Hammerspace Inc. is a data virtualization vendor that recruited Linux kernel maintainers to incorporate the company’s metadata-handling capabilities directly into the kernel. As a result, Hammerspace’s data virtualization provides a unified metadata environment outside the data path for exceptionally high performance.

Data virtualization has been around for years, but now Hammerspace is finding that its technology is well-suited for meeting the voracious needs of large language models or LLMs, as it enables companies to build curated data sets from siloed data sources.

Another significant pivot is Comet ML Inc., which spent several years as an MLOps vendor. It has reinvented itself as an end-to-end agent development platform that offers agentic evaluations both during test as well as production (see my recent article on the importance of evals).

Comet’s best differentiated capability: coding agent cost intelligence. Comet can tell engineering managers where developers are spending money on AI-generated code and can even reconfigure Claude Code to reduce coding costs up to 40%.

Backblaze Inc. has also found new life with a pivot away from S3-compatible object storage (ho hum) to exabyte-scale storage for AI-centric data sets. Whereas the company historically closed relatively small deals with its customers, neoclouds’ insatiable appetite for storage has changed the game, as Backblaze’s recent eight-figure deal with neocloud CoreWeave Inc. illustrates.

My final pivot is Rafay Systems Inc. I’ve been covering Rafay’s Kubernetes management offering for years, as my 2019 coverage of KubeCon shows. Today, the company offers a complete AI software stack for neoclouds and enterprises.

Kubernetes is still in the mix, but today AI-native infrastructure takes it for granted. On top of this underlying layer, Rafay now offers graphics processing unit management, an AI marketplace ready for neoclouds to white-label, and a “token factory,” which is what Rafay calls its ability to package and abstract AI tokens across models via a consistent application programming interface.

In essence, Rafay enables neoclouds (and some large enterprises) to offer an AWS-like experience for AI capabilities and services out of the box.

You didn’t know you needed these

Two vendors at RAISE stood out for providing solutions to problems that many people don’t have – or at least don’t recognize that they have.

SuSea Inc., doing business as You.com, offers web search APIs for AI, along with an underlying data layer and web index that reflects the majority of the public web.

LLMs need to be able to search the web. Models both large and small must also be able to search public and private data sets depending on their respective use cases. You.com provides an API layer that enables organizations to standardize and control their models’ access to search.

The key to understanding You.com’s value proposition is to realize that neither Google LLC nor Microsoft Corp. wants to provide open API access to their search results, as both vendors are pushing their own AI offerings. You.com provides a neutral third-party search capability that frees customers from having to work with the dominant players in the search market.

SkipLabs Inc. also offers a product that nobody realizes they need. SkipLabs offers a development harness for AI-based coding that generates a standardized OpenAPI contract that formalizes and attests to back-end dependencies.

In other words, SkipLabs provides a declarative approach to building stateful agentic applications. It works analogous to a compiler, enabling developers to create contracted back-end services for their agents that are well-defined and testable.

And the one you know you need

My final standout vendor from RAISE is Traversal Inc. Traversal offers AI-driven automated site reliability engineering or SRE for enterprises.

Traversal combines deep expertise in distributed system resilience and causal AI to build dynamic production world models of its customers’ production environments. These world models leverage Traversal’s generative knowledge graph technology that its platform creates on the fly based upon engineers’ prompts.

Essentially, Traversal provides a vibe coding-like experience for ops teams, automating root cause analysis and resolution of production issues.

Traversal’s production world model is so powerful, in fact, that customers are finding new uses for it based upon the model’s emergent properties. In essence, the model is powerful enough to generalize from the troubleshooting solutions it finds to other similar problems in the organization.

The Intellyx take

I recognize the smell of the RAISE Summit from the Dot Com days – it’s the smell of crazy money.

AI is going to be so huge, the money people say, that we need to pour billions of dollars into it so that we have a seat at the table. So in comes the money, and all the vendors scramble to get themselves a piece of the pie.

My comparisons to the dot-com boom, however, are intentional. There is certainly an AI bubble, and the only question is when it will pop.

In the meantime, however, it’s time to innovate. Feed the neoclouds. Pivot to AI. Solve problems nobody knows they have. It’s all fair game in today’s feverish AI world.

Jason Bloomberg is founder and managing director of Intellyx, which advises business leaders and technology vendors on their digital transformation strategies. He wrote this article for SiliconANGLE.

Photo: RAISE Summit