On October 26, 2026, iPronics announced that its co-founder and chief technology officer, Dr. Daniel Pérez-López, is its new chief executive officer. Board member and investor Geoffrey Tate becomes chairman of the board. The announcement, carried by GlobeNewswire from Valencia, Spain, and Santa Clara, California, is short on product news and long on intent: the company says it is moving into production, and it is putting the person who invented its core technology in charge of that move.
That framing is the story. Leadership changes at venture-backed hardware companies are routine; what is unusual here is how explicitly iPronics ties this one to a commercial inflection. The release says the transition follows the company’s recent $125 million Series B round, which includes strategic participation from NVIDIA, and describes iPronics as “delivering high-density, rapidly reconfigurable OCS solutions, based on photonics, to its customers including tier-1 hyperscalers and AI infrastructure providers.” It is a company telling the market that the technology argument is settled and the execution argument has begun — and that the founder is the right person to make the execution argument.
The move, in full
Pérez-López co-founded iPronics in 2019 as a spin-off from the Universitat Politècnica de València, where his doctoral research produced what the company describes as world-record chip performance and the first experimental validation of general-purpose programmable photonics processors. His background is in reconfigurable photonic integrated circuits — chips whose optical routing can be reprogrammed in software rather than fixed at fabrication. Before founding iPronics he worked on quantum photonics at Xanadu Quantum Technologies in Canada. His honors, as listed in the announcement, include the IEEE Photonics Society Graduate Student Fellowship, the Extraordinary PhD Thesis Award, inclusion in ElectroOptics’ inaugural Photonics100 list, the Royal Academy of Engineering Medal, and a European Research Council grant.
Tate, the incoming chairman, brings a different kind of résumé. He spent ten years at AMD running several business units, most recently as senior vice president of processors. As founding CEO of Rambus, he joined at the Series A, took the company public seven years later, and led it as a public company for seven more years, reaching a multi-billion-dollar valuation. He later helped turn around and sell Nanosolar, and co-founded Flex Logix, a provider of embedded FPGA and AI technology that Analog Devices acquired in 2024. He is now an investor, a strategic advisor to Light Street Capital, and a board member of Ayar Labs, the co-packaged optics company.
The pairing is deliberate, and the investor quotes in the release say so in terms. Manish Muthal, senior managing director at Maverick Silicon, which co-led the Series B, said the firm backed the round “based on our conviction in Daniel’s exceptional technical achievements and leadership” and called Tate’s operating experience the complement. Amelia Armour, partner at Amadeus Capital Partners, an early investor, said the firm had “seen Daniel grow in experience building a world-class team for photonics technology development” and added: “We are excited to have him lead the company as we move into production.” Armour also thanked outgoing CEO Christian Dupont “for leading the company for the last two years, culminating in our successful Series B.”
That last detail matters for reading the transition correctly. Dupont led iPronics through its commercialization push and its largest fundraise; the handover is presented as a planned succession at a natural breakpoint, not a crisis departure. The release gives no reason for Dupont’s exit beyond the thanks, and no new role for him is announced. Readers should not invent one.
What iPronics actually sells
To judge whether a founder-CEO is the right call for “commercial scale,” it helps to be precise about what is being scaled. Optical circuit switching, or OCS, is a different answer to datacenter networking from the packet switching that dominates today. In a conventional datacenter fabric, traffic arriving as light is converted to electrical signals at every switch, routed through a silicon switch chip, and converted back to light for the next hop — each conversion costing power and adding latency.
OCS routes light directly. A photonic switch establishes a physical optical path between two endpoints — a circuit, in the old telephony sense — and traffic flows through it without optical-electrical-optical conversion at the switch. The trade-off is well understood: circuit switches reconfigure more slowly than packet switches, so they suit traffic patterns that are stable over seconds or minutes rather than microseconds. AI training workloads, with their long-lived, high-volume flows between GPU groups, are the canonical fit. The pitch is that for scale-up networks — the interconnect that lets thousands of GPUs across multiple racks behave like one computer — OCS removes the conversion tax exactly where it is heaviest.
iPronics’ product is iPronics ONE, described by the company as a rack-ready OCS platform built on silicon photonics. It provides programmable optical connectivity within and across racks for what the company calls dynamic AI infrastructure, with integrated control, telemetry, and APIs. The stated benefits are real-time reconfiguration to improve GPU utilization, lower energy consumption, and scaling without costly network overhauls. Those are company claims about a product’s purpose, not independently measured results, and the release attaches no performance figures to them — no port counts, no switching times, no power-per-port numbers, no named customer deployments. That absence is itself information, and it sets the boundary of what this announcement can support.
The operational detail the release does give is geographic. To support what it calls its global commercial rollout, iPronics is scaling its advanced development center in Valencia, expanding its supply chain and operations capability in Taiwan, and building out its executive and customer-facing teams in Santa Clara. That three-point footprint — research in Spain, manufacturing operations in Taiwan, commercial presence in Silicon Valley — is the classic shape of a photonics company moving from prototype to volume, and it is the shape the new leadership is being asked to fill in.
The $125 million context
The Series B that precedes this transition was announced September 2, 2026: $125 million, co-led by Maverick Silicon and Light Street Capital, with participation from NVIDIA and a long list of new and existing investors including Triatomic Capital, Bosch Ventures, Catalight Capital, the European Innovation Council Fund, The Tate Family Trust, Fine Structure Ventures, Amadeus Capital Partners, Build Collective, and Criteria Venture Tech. The round brought iPronics’ total funding to $177 million.
NVIDIA’s participation is the detail that carries the most weight, and it should be read carefully. A strategic investment from the dominant supplier of AI accelerators is a meaningful signal about where the interconnect bottleneck sits in the industry’s thinking. But a participation is not a partnership announcement, a purchase order, or a design win — and the release does not say NVIDIA is a customer, so no one should report it as one.
The head-to-head: two $125 million bets on light
iPronics is not the only company to have raised exactly $125 million for OCS in the last two months, and the coincidence is instructive. On October 14, 2026 — twelve days before the iPronics CEO announcement — Oxford-based Salience Labs announced its own oversubscribed $125 million Series B, led by Bullhound Capital, with participation from AMD Ventures and Intel Capital, plus SBI Group, Hudson River Trading, and the British Business Bank. Salience builds an all-optical, fully integrated silicon-photonics switch architecture for AI datacenter scale-up networking, and its stated use of funds is nearly a mirror image of iPronics’: accelerate OCS deployment, scale volume manufacturing, and expand its United States presence in Silicon Valley.
The symmetry extends to the investor map, and the investor map tells its own story. iPronics’ round carries NVIDIA. Salience’s carries AMD Ventures and Intel Capital. The two leading OCS pure-plays have, in effect, sorted themselves under the banners of the three largest merchant silicon vendors, each of which has an interest in how AI clusters are wired. That is a useful frame for following the sector, but it is not evidence of a standards war or a procurement split: chip vendors invest broadly, and a venture check is not an exclusive endorsement.
Where the two companies’ public positions differ is in product specificity. Salience has shown hardware at a named milestone: a 32-port all-optical switch debuted in March 2026, and a 64-port version unveiled in September at ECOC 2026, which the company calls an industry first for silicon photonics OCS, with 128- and 256-port configurations in development. iPronics, by contrast, talks about its platform in terms of density and reconfigurability without publishing comparable port counts or switching metrics in its recent releases. A rack-ready platform with control software and APIs is a different proposition from a switch chip, and the two may not be directly comparable on port count. But for readers tracking who is actually shipping, Salience currently offers more verifiable milestones, and iPronics’ next announcements will be judged on whether they close that gap with specifics.
The “tens of billions” claim, handled with care
The October 26 release contains one market-size sentence, and it deserves scrutiny because it will be quoted. “OCS is projected to grow to tens of billions by 2030 as AI data center connectivity shifts from copper to optical.” The release attributes no source for this projection. It is a company pitch figure, presented in a company announcement, and it should be reported as such — not as an established market forecast.
There is a third-party number in circulation for comparison. Coverage of Salience’s round cited Cignal AI projecting the optical circuit switching market above $8 billion by decade’s end. The gap between “above $8 billion” from an analyst firm and “tens of billions” from a vendor release is not a contradiction to be resolved by picking a winner; the two figures may define the market differently, count different product categories, or simply reflect different optimism. What a reader should take away is the range and the uncertainty: OCS is a market that credible observers size in the billions by 2030, with wide disagreement about how many billions. Any article that repeats the “tens of billions” figure without the attribution is doing the company’s marketing for it.
The underlying directional claim — that AI datacenter connectivity is shifting from copper to optical — is on firmer ground. Copper cabling has hard physical limits on reach and power at the data rates AI clusters now demand, and the industry’s migration to optical interconnects at ever-shorter distances is well documented across transceiver roadmaps. What is genuinely contested is not whether optics displace copper, but which optical architecture wins the switching layer: OCS, co-packaged optics, or continued evolution of electrical packet switching at higher lane rates. That contest is the reason two companies just raised $250 million between them.
What this announcement is not
Precision about what did not happen on October 26 is as important as precision about what did. This announcement contains no new product, no new customer name, no deployment figure, no revenue number, and no performance specification. It does not say iPronics has shipped in volume, and it does not say any hyperscaler has committed to a deployment. The phrase “delivering… to its customers including tier-1 hyperscalers” is a statement about the company’s customer base in general terms; it names nobody and quantifies nothing, and it cannot be converted into a claim that a specific hyperscaler has deployed iPronics ONE.
Nor does a founder becoming CEO guarantee a technical outcome. The skills that produce a world-record chip are not identical to the skills that run a volume manufacturing operation and a hyperscaler sales cycle. The counter-argument, which iPronics’ investors are plainly making, is that the person who understands the technology’s limits most intimately is best placed to make the trade-offs production forces. Both arguments are plausible; the evidence will be in the next year of announcements.
Tate’s seat on Ayar Labs’ board is worth noting as a map of the ecosystem, not a conflict: Ayar Labs works on co-packaged optics, a complementary approach to the switching layer. Whether that cross-view becomes an advantage for iPronics depends on how the market sorts out co-packaged optics versus OCS, a question still open.
What to watch
The signals that will tell whether this leadership change was the right call are concrete, and they are all missing from the current announcement — which is precisely why they are the right things to watch. First, named deployments: a hyperscaler or AI infrastructure provider willing to be named as a customer, with a scale attached. Second, product specifics: port counts, reconfiguration times, and power figures for iPronics ONE that can be compared with Salience’s published milestones. Third, manufacturing evidence: the Taiwan operations expansion producing volume shipments rather than pilot lots. Fourth, the next fundraise or revenue disclosure that tests whether the $177 million total funding is carrying the company to self-sustaining commercial scale.
None of that is a criticism of the appointment. A founder who spent seven years turning programmable photonics from a doctoral thesis into a venture-backed company with NVIDIA on its cap table has earned the chance to run the production phase. The honest summary is narrower than the release’s rhetoric and more interesting than a routine personnel note: iPronics has told the market it believes its technology is ready and its problem is now execution — and handed execution to the person who built the technology. The next announcement that matters will be the one that shows a customer agreeing.
Sources and further reading
iPronics, “iPronics Names Co-Founder & Photonics Pioneer Dr. Daniel Pérez-López as CEO to Drive Commercial-Scale for AI Optical Circuit Switching,” GlobeNewswire, October 26, 2026. All appointment details, investor quotes, biographical information, geographic expansion plans, and product descriptions in this article are from this release. iPronics, “iPronics Raises $125 Million to Scale Programmable Optical Networking for AI Data Centers,” September 2, 2026 (company announcement; carried by GlobeNewswire and HPCwire): Series B size, co-leads Maverick Silicon and Light Street Capital, NVIDIA participation, full investor list, and the $177 million total funding figure. Salience Labs, “Salience Labs Secures $125 Million in Series B Funding to Scale Photonic Networking Chips for AI Datacenter Infrastructure,” Business Wire, October 14, 2026, with coverage in Data Center Dynamics: Salience’s round size, lead investor Bullhound Capital, AMD Ventures and Intel Capital participation, 32-port and 64-port switch milestones, and 128- and 256-port development plans. The “tens of billions by 2030” figure is iPronics’ own projection as stated in its October 26 release, reported here as a company claim; the contrasting “above $8 billion by decade’s end” figure is attributed to Cignal AI as cited in coverage of Salience’s round. No customer names, deployment figures, revenue figures, or iPronics ONE performance specifications appear in the cited releases, and none are asserted here.