THE APEX TIMES
Speaking at SC26 in Chicago, NVIDIA Vice President Ian Buck previewed Nemotron for Science, a locally runnable AI agents stack designed to plan and test hypotheses on researchers’ own compute. He also described new Vera CPU expansions and the Rubin HPC GPU, plus quantum co-design plans with IBM and Fermilab and claimed breakthroughs aimed at cutting quantum error-correction compute time.
NVIDIA used its opening remarks at SC26 in Chicago to frame the next phase of high-performance computing as an extension of simulation, augmented by AI agents and quantum progress. Ian Buck, NVIDIA’s vice president of hyperscale and high-performance computing, said the “scientific stack” is moving toward systems that can plan experiments, run code, and verify results within the environments where researchers already work.
Buck kicked off the company’s SC26 week with announcements centered on agents, software and hardware intended for scientific workloads. He said agents have changed how scientists interface with supercomputers, instruments, and devices, positioning them as a practical layer between models and real experiments. He added that while AI, quantum, and agents are gaining ground, simulation remains the core foundation of scientific computing.
One of NVIDIA’s key product announcements was Nemotron for Science. Buck said it combines the Nemotron 3.5 Super VL open model with a new open chemistry dataset and post-training recipes for life sciences and chemistry. The point, he said, is to let researchers run the system locally rather than depending on a remote service, including on a DGX Spark, a workstation, or their own supercomputer.
Buck also emphasized a data and control message aimed at enterprise and research institutions: “You own your data. You own the model.” To illustrate how agents might work in practice, he cited an example at Argonne National Laboratory, saying an agent built with NVIDIA’s toolkit took a single prompt, planned and tested hypotheses, selected which simulations to run, and identified two new binding sites.
On the HPC side, NVIDIA introduced additions to its Vera platform. Buck described an expanded Vera CPU platform and said DKRZ, the German Climate Computing Center, has selected Vera CPUs for its HLRE-5 supercomputer, which is expected to arrive in early 2028. He also said Vera systems are already planned at Texas Advanced Computing Center, indicating a broader deployment push beyond any single customer site.
The company’s other hardware focus was Rubin HPC, a new GPU in the Vera Rubin platform built for science, engineering, analytics, and AI. Buck said Rubin supports precisions from FP64 to FP4, includes sparse-data acceleration, and connects using seventh-generation NVIDIA NVLink. NVIDIA also described the GPU as having more than 50,000 CUDA cores and about three times as many as Hopper, with an expected 2028 timeline for availability.
For researchers looking toward quantum computing, Buck used SC26 to connect near-term software and systems work to a longer-term goal of faster, more feasible error correction. He described plans announced at the event by NVIDIA, IBM, and Fermilab for a quantum computing co-design platform that targets both error-correction research and application development.
Buck said the pairing would involve NVIDIA DGX B200 systems with IBM quantum computers, including systems powered by IBM Quantum Nighthawk. He also discussed a claimed path to reduce decoding latency from roughly 300 microseconds to 10, and introduced a CUDA-Q decoding algorithm named TCBP-JOBS, saying it cuts the amount of computation used for error-correction research from about 12 weeks to about 45 minutes. Separately, he described work being added to CUDA-Q related to a technique called magic state cultivation, which he said can reduce the operations needed for one logical step by a factor of ten on a simulation performed on Jülich’s JUPITER supercomputer.
Buck’s remarks also included an early deployment example in photonics, saying NVIDIA Photonics co-packaged optics switches shown at Lambda were described as 15 times more reliable on bit errors and link flaps. NVIDIA did not provide additional rollout timelines or performance measurements beyond those comparative reliability claims during the address.
Several details remain unclear from the publicly shared remarks. NVIDIA did not specify pricing, support terms, or availability dates for Nemotron for Science beyond the stated ability to run locally on certain systems. For Rubin HPC and the HLRE-5 schedule, Buck provided expected timeframes but not formal contractual milestones or benchmark results, and the quantum latency and compute-time claims were not accompanied by experimental methodology in the talk.
The tech investor argues that regulatory or compliance pressure may force NVIDIA’s latest GPUs to include buyer verification checks that amount to a biometric-style “kill switch,” aiming to deter resale or diversion to restricted end users.
A recent market note argues that Alphabet’s platform could offer better five-year stock upside than Amazon, pointing to how AI is being embedded into cloud products and distribution. The piece stops short of providing new, primary data on near-term earnings.
Ahead of its late-Tuesday results, Nvidia is set to update investors on performance for its fiscal third quarter, a key checkpoint for demand trends in AI-linked data center spending.
In results for its third quarter of fiscal 2027, NVIDIA said revenue rose more than twofold year over year, with data center revenue accounting for nearly all of the company’s growth.
Broad-market indexes slipped on Tuesday, with Nvidia’s upcoming earnings acting as a key weight on sentiment, according to Yahoo Finance.
Nvidia’s chief executive, Jensen Huang, indicated he would be willing to come back if the company’s board asked. A JPMorgan analyst said outside pressure would likely need to escalate into a proxy fight to change the situation.
In a new Google blog post, Alphabet points to five developer projects that use its Gemini 4 Carbon model for tasks ranging from large-scale physics simulation to turning messy public records into interactive 3D applications.
A market update cited a continued dip in the broader index heading into Nvidia’s next earnings release, underscoring how closely traders are watching the AI-chip leader’s quarterly results.
A coalition including Teamsters staged a rally in New York City supporting electric-bike delivery drivers they say were terminated after Amazon canceled a delivery contract, drawing renewed attention to labor practices in last-mile logistics.
Bonta says his office is weighing whether the law and facts justify action in an ongoing scandal involving Jensen Huang, as the regulator examines potential wrongdoing related to the artificial-intelligence chipmaker’s business.
Comments
Community guidelines. Comments may be removed for violations.
Loading comments...