Add ‘post-AI’ encryption to your to-do list
THE SO WHAT
If OpenAI’s 400+ math results are even partially applicable to cryptanalysis, you can’t wait for a formal “AI-broken” moment to rethink encryption. CISOs should start inventorying where long-lived secrets sit today and which of those need AI-resilient schemes over the next 3–5 years, not 10–15.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechCyberattack Hits India’s Largest Bank, Exposes Customer Records
A breach at India’s largest bank with records dumped to the dark web is a reminder that core financial infrastructure remains a high-yield target as digital penetration rises. Any institution with retail-scale data in emerging markets should revisit third-party, legacy-core, and identity controls this quarter—assume your attack surface is larger than your last audit showed.
Deep & Emerging TechExclusive: Photon Bridge raises €10m to build laser devices that connect AI chips
Optical interconnect startups like Photon Bridge are going after the real choke point in AI scaling—moving data between chips without melting your power budget. If you’re planning multi-GPU or custom accelerator clusters, start tracking photonics roadmaps alongside GPU SKUs, because bandwidth and energy, not raw FLOPs, will set your ceiling.
Deep & Emerging Tech'How do you know what's real?': US utilities are making AI data centers prove they exist before handing over power connections — as 50% of planned buildouts are currently being delayed, stalled, or canceled
Half of planned AI data center buildouts being delayed or canceled because utilities doubt the customers are real means power is now a due-diligence problem, not just an engineering one. If you’re planning large-scale AI infra, lock in credibility with utilities early—contracts, financing, and visible progress—before you assume timelines are real.
Deep & Emerging TechThis silent all-flash NAS with 10GbE and dual USB4 is cheaper if you buy it bare-bone and add your own SSDs
Fanless, local-AI-capable NAS with 10GbE and USB4 at bare-bones pricing is another nudge toward owning more of your inference and data stack. For teams with privacy or latency constraints, it’s time to re-run the math on a small on-prem tier alongside cloud.