In a groundbreaking move that underscores the accelerating convergence of artificial intelligence and national security, Los Alamos National Laboratory (LANL) has deployed advanced AI models on its Venado supercomputer, marking a significant leap in computational capabilities for classified research. The initiative, detailed in a recent announcement, involves integrating OpenAI’s latest o-series reasoning models into Venado’s architecture, which transitioned to a classified network earlier this year. This setup allows researchers to harness AI for accelerating simulations and analyses critical to national defense, from nuclear stockpile stewardship to complex physics modeling.

The Venado system, powered by NVIDIA’s Grace Hopper Superchips, represents a fusion of high-performance computing and AI acceleration. According to reports from the Department of Energy, the supercomputer can process tasks at unprecedented speeds, enabling AI-driven insights that were previously infeasible due to computational constraints. LANL officials emphasize that this deployment not only boosts efficiency but also positions the lab at the forefront of AI applications in secure environments.

Unlocking New Frontiers in AI-Driven Science

Industry observers note that Venado’s integration with OpenAI models could transform how scientists approach intractable problems. For instance, the system’s NVIDIA GH200 chips deliver performance metrics that outpace predecessors, with reports from Inside HPC & AI News highlighting its role in running reasoning models for national security science. This comes amid broader collaborations, including partnerships with NVIDIA and OpenAI, aimed at expanding AI resources for future projects.

Recent news from LA Daily Post elaborates that Venado’s success underscores the demand for AI-powered supercomputers, with NVIDIA’s Ian Buck describing it as a “frontier AI factory” that simulates the unobservable and generates scientific discoveries. The supercomputer’s energy-efficient design, achieving higher flops per second at lower costs, addresses longstanding challenges in scaling AI for high-stakes applications.

The Strategic Implications for National Security Research

Delving deeper, the deployment reflects a strategic pivot toward AI in defense sectors. Posts on X, formerly Twitter, from AI enthusiasts like those discussing NVIDIA’s advancements in accelerating large language models (LLMs) by up to 53 times, echo the excitement around such integrations. These sentiments align with Venado’s capabilities, where optimized models could slash inference times dramatically, as seen in broader industry breakthroughs.

Furthermore, coverage in ExecutiveGov points out that this partnership with OpenAI is part of a larger effort to install cutting-edge models on supercomputers for secure research. LANL’s history of innovation, including contributions to datasets for training AI as reported by Newswise, amplifies the potential impact, enabling multistage reasoning in visual and scientific tasks.

Challenges and Future Horizons in Supercomputing AI

Yet, this advancement isn’t without hurdles. Experts warn that deploying frontier AI on classified systems raises questions about data security and ethical AI use, particularly in national security contexts. News from The Hill highlights OpenAI’s commitment to scientific progress through such collaborations, but insiders stress the need for robust safeguards.

Looking ahead, LANL plans to expand its AI infrastructure, building on Venado’s foundation. As detailed in Scientific Computing World, the supercomputer’s role in accelerating national security-related science could inspire similar initiatives globally. With ongoing investments, including those in structured pruning and parallelized inference techniques discussed in recent X posts about AMD and NVIDIA optimizations, the trajectory points to even more powerful AI-supercomputer hybrids.

Broader Industry Ripples and Innovations

The ripple effects extend beyond LANL. Comparable efforts, such as those involving Meta and other national labs for molecular screening as per Newswise, illustrate a growing ecosystem where AI enhances scientific discovery. Venado’s reported 10 AI exaflops capability, noted in Data Center Dynamics, sets a benchmark for performance in AI workloads.

In essence, this launch signals a new era where supercomputers like Venado become indispensable tools for AI advancement, blending computational might with intelligent reasoning to tackle the most pressing challenges in science and security. As collaborations deepen, the potential for breakthroughs in fields from DNA research to complex simulations grows exponentially, promising a future where AI not only computes but truly innovates.

A leading Google scientist and recent Nobel laureate has highlighted “learning how to learn” as the paramount skill for future generations, given the transformative impact of Artificial Intelligence.

Demis Hassabis, CEO of Google’s DeepMind, delivered this insight from an ancient Roman theatre in Athens, emphasising that rapid technological advancements necessitate a fresh approach to education and skill acquisition. He stated that this adaptability is crucial to keep pace with AI’s reshaping of both the workplace and educational landscape.

“It’s very hard to predict the future, like 10 years from now, in normal cases. It’s even harder today, given how fast AI is changing, even week by week,” Hassabis told the audience. “The only thing you can say for certain is that huge change is coming.”

The neuroscientist and former chess prodigy said artificial general intelligence — a futuristic vision of machines that are as broadly smart as humans or at least can do many things as well as people can — could arrive within a decade. This, he said, will bring dramatic advances and a possible future of “radical abundance” despite acknowledged risks.

Hassabis emphasized the need for “meta-skills,” such as understanding how to learn and optimizing one’s approach to new subjects, alongside traditional disciplines like math, science and humanities.

“One thing we’ll know for sure is you’re going to have to continually learn … throughout your career,” he said.

The DeepMind co-founder, who established the London-based research lab in 2010 before Google acquired it four years later, shared the 2024 Nobel Prize in chemistry for developing AI systems that accurately predict protein folding — a breakthrough for medicine and drug discovery.

Greek Prime Minister Kyriakos Mitsotakis joined Hassabis at the Athens event after discussing ways to expand AI use in government services. Mitsotakis warned that the continued growth of huge tech companies could create great global financial inequality.

“Unless people actually see benefits, personal benefits, to this (AI) revolution, they will tend to become very skeptical,” he said. “And if they see … obscene wealth being created within very few companies, this is a recipe for significant social unrest.”

Mitsotakis thanked Hassabis, whose father is Greek Cypriot, for rescheduling the presentation to avoid conflicting with the European basketball championship semifinal between Greece and Turkey. Greece later lost the game 94-68.

Many artificial intelligence (AI) stocks have taken off this year, rebounding strongly from early-year weakness. Still, there has been differentiation among AI beneficiaries. For instance, companies that have inked deals with current leader OpenAI, such as Oracle (NYSE: ORCL) and Broadcom (NASDAQ: AVGO), have soared. Meanwhile, those perceived to be on the outside of OpenAI and its immediate suppliers have lagged.

Yet, while investors have bid up recent outperformers to stratospheric valuations, we’re really just in the second inning of the artificial intelligence revolution. That means certain stocks that have sold off for short-term reasons this summer could be excellent pickups to ride the AI wave, as long as they find their place in this ongoing paradigm shift. In that light, the following three look like strong buys on weakness.

Super Micro Computer

Super Micro Computer (SMCI 2.50%) has been on a roller-coaster ride over the past year, crashing after its accounting firm quit last October, only to recover strongly after its new accountant gave the thumbs-up to its books in February.

However, Supermicro’s stock sold off after its recent earnings report, which underwhelmed on both the top and bottom lines. Supermicro said that its customers were a bit slow in making architectural decisions, while tariffs and write-downs on old inventory pressured gross margins.

But there could be better things on the horizon. Supermicro still grew revenue 47% in the fiscal year ending in June and forecasted at least 50% revenue growth in fiscal 2026. Supermicro management also said it expects to increase its large-scale data center customers from four to between six and eight in fiscal 2026. That could be a good thing for customer diversification.

Meanwhile, Supermicro is just ramping up its data center building block solutions (DCBBS), wherein the company will install not just server racks but also an entire data center in turnkey fashion, greatly speeding up deployment. Those efforts should help margins grow back toward the company’s old range of between 14% and 17%, up from 11.2% in the latest fiscal year, even if those margins don’t get all the way there in 2026.

In any case, Supermicro has sold off to a forward price-to-earnings (P/E) ratio of just 16 after the sell-off. Given the exceptionally strong longer-term guide for AI infrastructure growth provided by Oracle and others recently, that still seems like a low price to pay for a leading AI hardware player growing that quickly.

Applied Materials

Like Super Micro, Applied Materials (AMAT -1.21%) sold off after its own recent earnings release. While Applied beat revenue and earnings estimates for its third quarter, which ended July 27, management forecasted a slight revenue and earnings decline in the current quarter. Management attributed the downturn to “digestion” in China, as well as “uneven” ramps in leading-edge logic.

While that may seem worrisome, the reasons given seem reasonable. Applied’s results actually held up better than some peers during the post-pandemic downturn in semiconductors, so it may make sense that there is a little air pocket today.

And while the leading-edge logic fab buildout may be uneven, the rise of artificial intelligence should bolster growth over the medium term. Oracle forecasts robust AI data center growth through 2030, and all those data centers will need lots of chips.

Applied is the most diverse semiconductor equipment supplier, so it should get a solid piece of that growing pie. Its equipment is concentrated in etch and deposition machines, which should see better-than-average growth over the next few years as chipmakers begin to implement new innovations such as gate-all-around transistors, backside power, and 3D architectures for both DRAM and logic chips, all of which are etch- and deposition-intensive.

Applied now trades at just 20 times earnings and 17 times next year’s estimates, which are below-market multiples. That seems absurdly cheap for a high-margin, cash-generating tech leader that should benefit from AI growth. Fortunately, Applied has rewarded shareholders with consistent share repurchases and a growing dividend, and that should continue going forward, even if the company has an off quarter here and there.

Intel

Finally, perhaps no tech company has been as maligned over the past few years as Intel (INTC -2.15%). After falling behind Taiwan Semiconductor Manufacturing (NYSE: TSM) in process technology and failing to anticipate the AI revolution, Intel spent the last four years on a spending spree in an attempt to catch up. That spending has added to Intel’s debt load and degraded its cash flow, while a lot of the fruits of that spending have not yet emerged.

Still, Intel recruited former board member and Cadence Design Systems (NASDAQ: CDNS) CEO Lip-Bu Tan as its new CEO, who is just a matter of months into his turnaround plan. Tan has unmatched experience and contacts within the semiconductor industry and seems like an ideal candidate to lead Intel at this stage.

Tan has made waves, cutting a massive amount of costs and restructuring the company. At a recent conference, CFO David Zinsner said Tan has already reduced management layers at the company from 11 to five. Meanwhile, Tan has also refreshed much of Intel’s leadership. In June, Tan promoted a new chief revenue officer and brought in several outside engineering leaders to lead Intel’s AI chip efforts.

At a recent industry conference, CFO David Zinsner stated that Tan would be laying out the company’s new AI roadmap soon. Then just last week, Tan named new heads of client and data center chip groups, completing his refreshment of Intel’s senior leadership. Given Tan’s wide experience as head of Cadence and his venture capital firm Walden Capital, which invests in AI start-ups, this new leadership is likely to strengthen Intel’s product portfolio.

Meanwhile, Intel’s first chip on its important 18A node will make its debut later this year, which management believes will give Intel equal or better technology than TSMC. And with the U.S. government recently taking a stake in the company and Tan having deep industry relationships, it seems likely Intel will land more external customers for its foundry, which will be another key to its success.

And yet, Intel trades just a touch above book value. But given that Tan is early in his transformation plan and the 18A node is just about to hit late this year, the stock is a great-looking risk-reward at these levels.