PARIS — January 15, 2026 — French deep-tech chip company VSORA has named semiconductor industry veteran Sandra Rivera, the former CEO of Altera, as Chair of its Board of Directors, marking a major leadership addition as the startup prepares to ramp production of its flagship AI inference processor, Jotunn8. VSORA says Rivera brings more than three decades of executive experience across semiconductors, data centers, AI, networking, and cloud infrastructure—experience the company believes will be crucial as it moves from breakthrough engineering to real-world deployment at scale. Rivera spent more than 20 years at Intel in increasingly senior roles and most recently led Altera, the FPGA business Intel spun out in partnership with Silver Lake.

Why this appointment matters right now

The timing is not accidental. Across the tech world, AI is quickly shifting from “training” (building huge models) to “inference” (running those trained models for billions of everyday tasks). Inference tends to be continuous and always-on, so it puts relentless pressure on power, cooling, and total cost of ownership. That’s exactly the problem VSORA is aiming to solve with a purpose-built inference chip designed for next-generation data centers, cloud infrastructure, and edge deployments. Put simply: training might be a sprint, but inference is the marathon. And the marathon is expensive—especially when electricity and cooling become the limiting factors. VSORA positions Jotunn8 as an “ultra-high-performance” inference chip that focuses on both speed and energy efficiency—two requirements that data center operators increasingly treat as non-negotiable. On its website and in prior product updates, the company emphasizes that Jotunn8 is engineered to reduce bottlenecks between compute and memory—an issue often described as the “memory wall”—so hardware spends less time waiting for data and wasting power. Industry observers have also pointed to a broader wave of AI infrastructure investment (including power-hungry data centers and long-term capacity deals) as organizations prepare for a future where inference demand keeps climbing. What VSORA announced In its January 15 release, VSORA confirmed Rivera “today assumed the role” of Chair of the Board of Directors. The company described her as a proven CEO and C-suite leader, highlighting her long track record in corporate transformation, realignment, and scaling multi-billion-dollar businesses. Khaled Maalej, VSORA’s founder and CEO, framed the appointment as a strategic step as the company accelerates toward production of Jotunn8. In the announcement, Maalej said Rivera’s background aligns closely with VSORA’s mission and that her leadership will support the organization as it builds high-performance global teams and pushes technical innovation in a competitive AI hardware landscape. Rivera, for her part, said she is excited by VSORA’s capabilities and understands firsthand how challenging it is to build a new chip architecture that is both power-efficient and meaningfully differentiated. She also emphasized the advantage small, focused teams can have—speed and agility—especially in fast-moving AI markets. Rivera’s background: why VSORA is betting on her VSORA’s choice is notable because Rivera’s career touches many of the exact pressure points that emerging chip companies face: building trust, simplifying product roadmaps, and executing through industry cycles. Leadership at Altera during a major transition VSORA highlighted Rivera’s most recent role as CEO of Altera, where she led the successful spinout of Intel’s FPGA business in partnership with Silver Lake. The announcement credits her with streamlining roadmaps, empowering technical leadership, and rebuilding customer and distributor trust—three areas that can make or break a semiconductor business, especially during ownership transitions. Two decades of senior leadership at Intel Before Altera, Rivera spent years in top Intel roles, including: Executive Vice President and General Manager of Intel’s Data Center and AI Group (overseeing Xeon CPUs, GPUs, FPGAs, and AI accelerators, plus the company’s enterprise-wide AI strategy) Executive Vice President and Chief People Officer Senior Vice President and General Manager of Intel’s Network Platforms Group, where Intel grew its networking business and became a market leader in network logic This blend—technical portfolio leadership plus organizational leadership—is particularly relevant for a company scaling from “chip design success” into “global operations success,” where hiring, partner ecosystems, and execution discipline matter as much as architecture. Board and academic ties VSORA also noted Rivera serves on the Board of Directors of Equinix, a major data center infrastructure company, and sits on an advisory board at UC Berkeley’s College of Engineering. Educationally, she holds an electrical engineering degree from Penn State and has earned executive leadership and sustainability certificates from Stanford’s Graduate School of Business. Those connections matter because AI inference hardware is not just a chip question anymore—it’s a systems question involving data centers, power, cooling, supply chains, and deployment realities. What Rivera says she will focus on first Rivera outlined a near-term priority list that reads like a scaling playbook: Strengthen foundational infrastructure (the internal operating structure that keeps development and execution aligned) Review and refine the product roadmap (what ships when—and why) Evolve product strategy to differentiate and scale (clear positioning, measurable goals, consistent execution) This suggests her role won’t be symbolic. It implies active governance and strategy oversight as VSORA tries to translate technical promise into repeatable customer outcomes. VSORA and Jotunn8: the product at the center of the story VSORA was founded in 2015 and describes itself as a French deep-tech company building next-generation AI inference silicon. The company says its flagship data center chip, Jotunn8, is expected to be available in early 2026, aiming for “breakthrough performance and energy efficiency” for large-scale inference. While many AI accelerators are widely associated with training workloads, VSORA positions Jotunn8 as purpose-built for inference—where speed-per-watt and cost-per-query often decide winners. In earlier updates, VSORA stressed architectural choices meant to reduce wasted compute cycles and power draw caused by memory bottlenecks. Trade publications have also discussed VSORA’s roadmap in practical deployment terms—development boards, reference designs, and server plans—reflecting the reality that successful inference adoption depends on an ecosystem, not just silicon. The bigger picture: why inference efficiency is becoming a battleground If you zoom out, the announcement speaks to three big trends reshaping AI hardware: 1) Inference is becoming the dominant workload As models move into everyday products—search, assistants, translation, customer support, security analytics, industrial automation—inference runs constantly, not occasionally. McKinsey has projected that inference will surpass training and become the dominant AI data-center workload by 2030, which would reshape how hyperscalers plan sites, networks, and power. 2) Power and cooling constraints are now strategic constraints AI data centers aren’t only limited by how many GPUs or accelerators you can buy. They’re limited by how much energy you can deliver to a site, how effectively you can cool it, and how reliably you can operate it 24/7. That’s one reason why major tech players have been signing long-term infrastructure and energy arrangements and pursuing diversified compute strategies. 3) “Sovereign compute” and regional capability are rising priorities VSORA’s release explicitly frames its work at the intersection of data center modernization, AI efficiency, and “sovereign compute infrastructure,” signaling a focus on regional control and resilience in critical AI supply chains—especially relevant in Europe. In that context, adding a globally recognized semiconductor executive as board chair is a credibility and execution signal to partners, customers, and investors. What happens next VSORA says it operates globally with offices in Japan, Korea, Singapore, and Taiwan—an important detail, because chip success depends on international collaboration across design, validation, packaging, manufacturing, systems integration, and customer enablement. With Rivera as Chair, the company is effectively telling the market: we’re moving from “great chip story” to “disciplined scale-up.” For VSORA, the near-term milestones that will likely matter most include: Demonstrating repeatable inference performance and efficiency in real customer workloads Building robust software and systems support around Jotunn8 Executing a reliable path to volume production and deliveries Proving cost competitiveness at scale (cost per token/query, cost per watt, and operational simplicity) If those pieces come together, VSORA could become a more visible player in the global race to make AI inference cheaper, faster, and far less power-hungry—exactly where the market is heading. External reference (company information): VSORA’s official website. #SandraRivera #VSORA #AIInference #Semiconductors #SlimScan #GrowthStocks #CANSLIM