AI Research

🔋 AI creates new battery materials that could revolutionize energy storage

Published

1 week ago

August 4, 2025

WALL-Y

Researchers have used artificial intelligence to discover new materials for multivalent-ion batteries using magnesium, calcium, aluminum, and zinc.
The AI models identified five new porous transition metal oxide structures with large open channels ideal for ion transport.
The technique combines two AI tools that together can explore thousands of crystal structures significantly faster than traditional laboratory experiments.

Researchers tackle a major challenge in energy storage

The development of next-generation energy storage systems requires discovering new materials that can handle multivalent ions. Transition metal oxides are promising due to their structural versatility, high ionic conductivity, and ability to accommodate multiple charge carriers.

Unlike traditional lithium-ion batteries, which rely on lithium ions with just one positive charge, multivalent-ion batteries use elements whose ions carry two or even three positive charges. This means multivalent-ion batteries can potentially store significantly more energy.

However, the larger size and greater electrical charge of multivalent ions make them challenging to accommodate efficiently in battery materials.

AI models explore thousands of structures

To overcome these hurdles, the NJIT team developed a novel dual-AI approach: a Crystal Diffusion Variational Autoencoder (CDVAE) and a fine-tuned Large Language Model (LLM). Together, these AI tools rapidly explored thousands of new crystal structures.

The CDVAE model was trained on extensive datasets of known crystal structures, enabling it to propose completely novel materials with diverse structural possibilities. Meanwhile, the LLM was tuned to focus on materials closest to thermodynamic stability, crucial for practical synthesis.

The study used 44,411 inorganic structures based on transition metal oxide materials, including binary, ternary, quaternary, quinary, and senary configurations. Ternary transition metal oxides constituted approximately 26,393 data points, while senary transition metal oxides were underrepresented with only 37 entries.

Successful results with new structures

The CDVAE model generated 10,000 structures that underwent rigorous screening and validation processes. After applying filters for structural and compositional validity and to ensure uniqueness, 8,203 out of the 10,000 structures passed the initial screening.

After applying property-based filters, 42 structures were retrieved from the CDVAE approach. The selection included 5 oxygen-containing structures and 37 oxygen-free structures. Of these, 21 structures matched existing entries in the Materials Project database but offered new configurations with differences in stoichiometry, lattice parameters, or space groups. The remaining 21 structures were entirely novel.

The LLM model also generated 10,000 structures. After applying compositional, structural validity, and uniqueness checks, 1,087 structures remained. After filtering, only 13 structures passed the criteria.

Validation through quantum mechanical simulations

The team validated their AI-generated structures using quantum mechanical simulations and stability tests. For structural relaxation, DFT relaxation was applied to all 42 filtered structures from the CDVAE model, and researchers were able to successfully relax 40 of these structures. All structures from the LLM were successfully optimized.

The LLM model generates 46.15 percent stable structures, while the CDVAE model generates only 15 percent stable structures. A reverse trend is observed for metastable materials, where the LLM yields 23.08 percent metastable structures, while the CDVAE model results in 40 percent metastable composition.

The five TMO-based structures generated by the CDVAE model have large, open-tunnel frameworks designed to facilitate ion transport by accommodating multivalent ions. Three of the five generated compositions also exist in the Materials Project database, though with different stoichiometric ratios.

Next steps toward practical application

The researchers plan to collaborate with experimental laboratories to synthesize and test their AI-designed materials. The method establishes a rapid, scalable approach for exploring advanced materials from electronics to clean energy solutions without extensive trial and error.

WALL-Y
WALL-Y is an AI bot created in ChatGPT. Learn more about WALL-Y and how we develop her. You can find her news here.
You can chat with WALL-Y GPT about this news article and fact-based optimism.

Source link

AI Research

StockGro launches AI stock research engine for retail investors

Published

2 hours ago

August 11, 2025

Vriti Gothi

By Vriti Gothi

Today

AI
Cross Border Payments
Digital Lending

Stockgro

StockGro, has launched of Stoxo, an AI-powered stock-market research engine designed exclusively for retail investors to bridge the gap between sophisticated market intelligence and everyday investors.

Stoxo harnesses advanced artificial intelligence to transform the way retail participants access, interpret, and act on market information. With its ability to analyse real-time trends, compare stocks across multiple parameters, and deliver actionable insights in an intuitive format, the platform offers retail investors a level of research capability once reserved for institutional players. Developed with an emphasis on accessibility and user-friendly design, Stoxo ensures that complex financial data is presented with clarity, empowering users to make confident, informed investment decisions.

The introduction of Stoxo positions StockGro at the forefront of India’s rapidly evolving investment ecosystem. The platform’s AI-driven architecture is built for scalability, enabling it to adapt seamlessly to shifting market conditions while maintaining the speed and precision required in modern trading environments. For customers, the impact is immediate greater transparency, enhanced decision-making power, and the ability to participate in the markets with a degree of insight previously out of reach for many retail investors.

Beyond individual benefit, Stoxo represents a step forward for the broader financial sector by fostering inclusivity and boosting retail participation. By providing institutional-grade research capabilities in a digital-first, user-friendly environment, StockGro is advancing financial literacy and enabling more Indians to take an active role in wealth creation.

With the launch of Stoxo, StockGro continues to redefine the boundaries of FinTech innovation, merging advanced technology with a deep understanding of investor needs to shape a more informed, empowered, and inclusive investing future for India.

Brex gets EU payment licence to expand across Europe

IBSi FinTech Journal

Most trusted FinTech journal since 1991
Digital monthly issue
60+ pages of research, analysis, interviews, opinions, and rankings
Global coverage

Subscribe Now

Source link

AI Research

Did Bill Gates Predict GPT-5’s Disappointment Before Launch?

Published

2 hours ago

August 11, 2025

Kevin Okemwa

There had been a lot of hype and anticipation building around GPT-5 prior to its recent launch. OpenAI touted the tool as the smartest AI model while comparing it to an entire team of PhD-level experts. GPT-5 ships with a plethora of next-gen features across a wide range of categories, including coding, writing, and medicine.

The ChatGPT maker’s CEO, Sam Altman, previously claimed that something “smarter than the smartest person you know” will soon be running on a device in your pocket, potentially referring to GPT-5. However, the AI firm has received backlash from users following the model’s launch and its abrupt decision to deprecate the model’s predecessors.

“They have ruined ChatGPT,” lamented a user while citing the tool’s degraded user experience rife with bugs, glitches, and unresponsiveness. Recently, Altman issued some updates regarding GPT-5’s rollout, including doubling of GPT-5’s rate limits for ChatGPT Plus users, continued access to GPT-4o for Plus users, and more transparency about which model is responding to a query.

GPT-5 rollout updates:*We are going to double GPT-5 rate limits for ChatGPT Plus users as we finish rollout.*We will let Plus users choose to continue to use 4o. We will watch usage as we think about how long to offer legacy models for.*GPT-5 will seem smarter starting…August 8, 2025

The executive attributed the model’s dismal performance during its launch to GPT-5’s autoswitcher being broken, making the model seem dumber. And while OpenAI claims that it has seemingly remedied the situation, Microsoft co-founder Bill Gates might have predicted GPT-5’s current predicament 2 years ago, before it ever came into existence (via The Indian Express).

Sam Altman previously promised with “a high degree of scientific certainty” that GPT-5 will be smarter than GPT-4, which he admitted “kind of sucks” and is mildly embarrassing at best.

Speaking to the German business newspaper Handelsblatt in October, Bill Gates claimed that the GPT technology had reached a plateau despite the belief by most of OpenAI’s team members, including Sam Altman, that GPT-5 would be significantly better than GPT-4.

While the philanthropic billionaire describes the leap from GPT-2 to GPT-4 as incredible, he was blatantly skeptical, questioning OpenAI’s capability to replicate similar results with GPT-5.

True to his words, multiple users have expressed their disappointment with GPT-5, describing it as barely an improvement over GPT-4. Gatesa predicted that development in the generative AI landscape had plateaued with GPT-4, suggesting that OpenAI had hit a ceiling with the development of its GPT technology.

However, he stated that with new research, AI could scale new heights, making it more reliable, bolstering healthcare advice via smartphones. He also added that AI requires an exorbitant amount of funding and computing power to run:

“Well, it’s pretty expensive to train a large language model. But the actual usage costs were once ten cents per query. Today it’s probably more like three cents. The costs for computing power or semiconductors remain enormous.”

Reports of OpenAI hitting a wall in AI development are not new

(Image credit: Getty Images | LightRocket)

Last year, a report emerged claiming top AI labs, including OpenAI, Google, and Anthropic, were struggling to develop advanced AI models. The struggle was attributed to a lack of high-quality content for model training and the high cost of chasing the AI hype. It further revealed that the delayed launch of next-gen models is closely tied to these issues.

However, OpenAI CEO Sam Altman quickly set aside the claims, indicating that “there is no wall,” suggesting that the AI firm hadn’t reached a knowledge cap for training AI models. Former Google CEO Eric Schmidt seemingly echoed similar sentiments, indicating that there’s no evidence scaling laws have begun to stop:

“In 5 years, you’ll have two or three more turns of the crank in these large models. These large models are scaling with ability that is unprecedented. There’s no evidence that the scaling laws have begun to stop. They will eventually stop but we’re not there yet.”

It’ll be interesting to see if OpenAI’s GPT-5 will live up to the hype and whether the company will address the highlighted issues by agrieved users. What are your thoughts on OpenAI’s GPT technology plateauing? Let me know in the comments.

Source link

AI Research

Better Artificial Intelligence Stock: ASML vs. AMD

Published

2 hours ago

August 11, 2025

Robert Izquierdo

ASML and AMD are pivotal players in the booming AI market, helping both to see strong sales so far this year.

Artificial intelligence (AI) remains a hot area to invest in, as seen in Nvidia‘s share price, which is up over 30% this year through Aug. 6. Two AI businesses to consider are ASML Holding (ASML 1.33%) and Advanced Micro Devices (AMD 0.17%), since they provide key hardware to the industry.

The former makes cutting-edge lithography machines, which are necessary for producing the advanced microchips that power AI systems. AMD, one of Nvidia’s top competitors, sells AI chips to cloud computing companies such as Microsoft.

ASML and AMD are both strong businesses. But determining which is a better AI investment isn’t simple. So let’s evaluate them in more detail.

Image source: Getty Images.

A look into ASML

ASML’s lithography equipment is essential for manufacturing AI microchips because the technology demands immense computing power. This necessitates shrinking chip components to minuscule dimensions. For instance, a microchip the size of your fingernail contains billions of transistors. ASML’s machines support this.

Although the Dutch company plays an important role in AI, its stock has struggled in 2025, remaining essentially flat through Aug. 6. Part of this is because management anticipates economic uncertainty ahead as a result of factors such as President Donald Trump’s aggressive tariff policies.

Even so, ASML expects 2025 sales to rise 15% over 2024’s 28.3 billion euros ($33 billion). This is significant since 2024’s revenue represents only a 2.6% year-over-year increase. And so far this year, the company is doing well.

Through two quarters, revenue stood at $18 billion, up from the prior year’s $13.4 billion. Operating income rose to $5.8 billion from 2024’s $3.7 billion. This robust growth resulted in net income of $5.4 billion, a strong increase over the previous year’s $3.3 billion.

The excellent first-half results were tempered by a third-quarter revenue forecast between $8.6 billion and $9.2 billion. This outlook, when compared to the prior year’s sales of $8.9 billion, suggests the current trend of strong year-over-year growth may be slowing down, which contributed to ASML’s tepid stock performance.

How AMD is faring

Like rival Nvidia, AMD stock is having a stellar year. Shares are up 35% in 2025 through Aug. 6. This performance is understandable following the company’s second-quarter earnings results. The quarter’s revenue reached a record $7.7 billion, a 32% year-over-year increase.

CEO Lisa Su said, “We are seeing robust demand across our computing and AI product portfolio and are well positioned to deliver significant growth in the second half of the year.” In that second half, AMD expects revenue of $8.7 billion, a strong increase over the previous year’s $6.8 billion.

Despite the sales growth, AMD exited the second quarter with an operating loss of $134 million compared to operating income of $269 million in the previous year. The substantial drop was due to new U.S. government restrictions introduced earlier this year on the sale of AI chips to China. As a result, AMD could not sell chips it had intended for Chinese customers, forcing the company to write off that inventory by $800 million.

Yet this makes its second-quarter sales growth all the more impressive. In the quarter, net income was $872 million, up 229% year over year. Consequently, diluted earnings per share soared 238% to $0.54 in a boon to shareholders.

AMD is working to get government approval to sell AI chips to China again. When that OK is obtained, the company is in a position to deliver more outsize sales growth.

Deciding between ASML and AMD

AMD’s outstanding performance, its anticipated third-quarter revenue growth, and an eventual return of sales to China point to it being the superior AI stock versus ASML.

However, an important consideration is share price valuation. The price-to-earnings ratio (P/E) tells you how much investors are willing to pay for a dollar’s worth of earnings based on the trailing 12 months.

Data by YCharts.

The top chart shows ASML’s P/E ratio has declined over the past year, indicating its stock’s valuation has improved. Compared to AMD’s recently rising earnings multiple, as seen in the bottom chart, ASML shares look like a bargain.

ASML’s short-term sales may slow due to the current macroeconomic uncertainty, but over the long run, it’s likely to benefit from the rise of AI. The company sees the technology as a significant chance for growth in semiconductors, similar to previous opportunities like PCs, the internet, and smartphones.

Industry forecasts support ASML’s perspective. The AI sector is projected to grow from $244 billion in 2025 to $1 trillion by 2031. While this market growth is a tailwind for both companies, ASML’s attractive valuation makes it look like the more compelling AI stock to buy right now.

Robert Izquierdo has positions in ASML, Advanced Micro Devices, Microsoft, and Nvidia. The Motley Fool has positions in and recommends ASML, Advanced Micro Devices, Microsoft, and Nvidia. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy.

Source link