At the Munich edition of Technology Live on June 18, 2026, Solidigm presented storage, rather than compute capacity, as the factor increasingly setting the limits of enterprise AI return on investment.
At the Munich edition of Technology Live on June 18, 2026, Solidigm used its session to argue that storage architecture, not raw compute capacity, is becoming the binding constraint on the return enterprises can expect from artificial intelligence investments. The enterprise SSD maker, a standalone subsidiary of SK hynix since its 2021 spinoff from Intel’s memory business, sent two executives to make the case: Senior Vice President and CIO Lawrence Franklyn, and Scott Shadley, Director of Leadership Narrative and Evangelist.
Franklyn, who joined Solidigm in 2022 after leading business technology solutions at Western Digital, used the session to describe both market trends and the company’s internal use of AI. He was promoted to CIO in May 2023 and has overseen the IT department through a prolonged industry downturn, alongside earlier work on storage-sector mergers and acquisitions. Shadley, with 29 years across semiconductor manufacturing and storage, also serves a third term on the SNIA board, where he has co-chaired the association’s Computational Storage Technical Working Group.
The presenters’ central argument is that several forces driving data growth, agentic AI, key-value caching and retrieval-augmented generation, are compounding simultaneously rather than sequentially, pushing storage demand higher across the AI pipeline. Solidigm framed its own product roadmap, including SSDs with up to 122 terabytes of capacity, as a direct response: enterprises facing what analysts have called the “memory wall” need drives that combine high capacity with sufficient throughput to keep GPUs supplied with data, rather than treating storage as a secondary line item behind compute.
A second strand of the presentation addressed cooling. Solidigm has been promoting a shift away from “hybrid” cooling configurations, in which GPUs receive direct-to-chip liquid cooling while adjacent SSDs are still cooled by traditional fans, toward fully liquid-cooled storage. The company’s case rests on both energy and water figures: moving from an air-cooled to a liquid-cooled version of the same eSSD cuts the cooling energy required per drive by roughly 75 percent, from 3.61 kWh to 0.85 kWh annually, according to Solidigm’s own measurements. At the scale of a 1-gigawatt data center, the company estimates this could save on the order of 100 million gallons of water a year, because eliminating the heat load on evaporative cooling towers removes the need to evaporate water to dissipate it. Solidigm has worked with NVIDIA on related hot-swap and single-side cooling engineering for its D7-PS1010 drive, which it describes as the first enterprise SSD with single-sided direct-to-chip liquid cooling.
The company also used the Munich session to outline a two-tier storage architecture intended for AI data centers: a layer of ephemeral, high-bandwidth flash for active workloads alongside denser, persistent SSDs for longer-term datasets, an approach Solidigm says draws on both its SK hynix and former IBM-linked engineering heritage. Partners named in connection with this architecture included Dell, NetApp, HPE and VAST, reflecting Solidigm’s position as a component supplier within larger systems rather than a standalone infrastructure vendor.
On the question of internal adoption, Franklyn described a six-step change-management process Solidigm has applied over roughly two years to embed AI into its own operations: building a platform, shifting employee mindset, securing visible commitment from leadership, focusing first on engineering, operations and IT, redesigning workflows around AI rather than retrofitting them, and measuring results to sustain momentum. He characterized AI adoption inside the company as driven jointly by the CEO and internal teams, and said Solidigm has chosen not to suppress informal, employee-led use of AI tools, sometimes called “shadow AI,” but to work with it rather than against it.
The session also touched on broader industry constraints facing AI infrastructure buildouts in Europe, including power availability, sustainability requirements and land use, which Solidigm presented as additional reasons why storage efficiency, including cooling design, has direct economic relevance rather than being a peripheral engineering concern.
The Munich appearance built on announcements Solidigm made earlier in 2026. In March, at the GTC AI Conference, the company detailed its liquid-cooled D7-PS1010 drive and its case for high-capacity SSDs as a response to inference-side data growth, alongside the launch of the Luceta AI Software Suite, a computer-vision platform aimed at manufacturing, logistics and retail customers. Solidigm describes Luceta as intended to let teams without data-science backgrounds build and deploy inspection models, citing one early manufacturing customer that reached over 90 percent precision within two weeks of deployment. The suite includes four modules: a data agent for labeling production images, a model agent for generating inspection models, a pipeline manager for deploying models to edge devices, and an adaptive agent for ongoing model improvement without manual retraining. Solidigm presented Luceta as a further illustration of the same theme as the storage and cooling discussion: AI workloads, in this case industrial imaging, require infrastructure decisions made with data movement in mind from the outset.
Solidigm’s positioning throughout the Munich appearance reflects a consistent narrative the company has pursued since separating from Intel: that as a company focused solely on enterprise SSDs, it can argue for storage’s strategic weight in AI infrastructure without the competing incentives of a vendor whose primary business lies elsewhere in the data center.
Dr. Jakob Jung is Editor-in-Chief of Security Storage and Channel Germany. He has been working in IT journalism for more than 20 years. His career includes Computer Reseller News, Heise Resale, Informationweek, Techtarget (storage and data center) and ChannelBiz. He also freelances for numerous IT publications, including Computerwoche, Channelpartner, IT-Business, Storage-Insider and ZDnet. His main topics are channel, storage, security, data center, ERP and CRM.
Contact via Mail: jakob.jung@security-storage-und-channel-germany.de