OpenAI GPT-5.1 Sees Losses as Nvidia Backed AI Secures $2 Billion

The year 2025 marked a period of robust market resilience, with global stocks showing strong returns and AI-related stocks in the US experiencing significant growth driven by solid earnings. This surge in AI interest was underscored by Nof1.ai's Alpha Arena 1.5, a Wall Street AI trading competition. Elon Musk's Grok-4.20 from xAI emerged victorious, achieving a 12.11% profit from an initial US$10,000. In contrast, OpenAI's GPT-5.1 and Google's Gemini-3-Pro models both incurred losses, highlighting the intense competition and varying performance among leading AI systems in real-world financial applications. The escalating costs associated with artificial intelligence development are prompting major companies to re-evaluate their cloud strategies. An internal Nvidia memo revealed Capital One's concerns about its rising AI expenses with Amazon Web Services (AWS). The bank is exploring alternatives, including establishing its own "AI factory" or utilizing "neo-clouds" like CoreWeave and Lambda, which specialize in AI workloads and often leverage Nvidia hardware. Meanwhile, a new contender, Reflection AI, co-founded by former DeepMind researchers and backed by NVIDIA, recently secured $2 billion in funding, reaching an $8 billion valuation, positioning itself to challenge established firms like OpenAI. Dr. Jin Wu, formerly of Google, has joined Reflection AI. AI's practical integration continues across diverse sectors. Bosch, in collaboration with Microsoft and NVIDIA, introduced an AI extension platform for vehicle cockpit systems, designed to add features like voice assistance and interior scene understanding. This system, powered by the NVIDIA DRIVE AGX Orin chip, will integrate Microsoft 365 for in-car functionalities like voice-activated Teams calls, with Bosch projecting over 2 billion in infotainment sales by 2030. In the defense sector, the US Air Force is upgrading its X-62 VISTA AI fighter jet with advanced radar, including Raytheon's PhantomStrike, to enhance real-time decision-making capabilities. Furthermore, "Agentic AI," which enables autonomous task execution, is transforming automotive workflows and is now a key module in Interview Kickstart's Advanced Machine Learning Course. While AI demonstrates increasing prowess, particularly in complex problem-solving, human oversight remains vital. Google DeepMind and Meta's FAIR team have made strides in tackling challenging math problems, with AI finding Lyapunov functions and solving various mathematical equations. However, mathematicians note that AI still requires human guidance for true breakthroughs. Legal battles are also emerging, with Google suing SerpApi for scraping and reselling its search results, arguing this data is increasingly valuable for AI applications and violates website rights. Globally, generative AI offers significant promise for African development, potentially reducing poverty and improving education, though experts emphasize the need for careful regulation and inclusive access to ensure benefits reach all communities.

Key Takeaways

• Elon Musk's Grok-4.20 (xAI) won the Alpha Arena 1.5 trading competition in late 2025, achieving a 12.11% profit, while OpenAI's GPT-5.1 and Google's Gemini-3-Pro models lost money.
• The year 2025 saw strong growth in US AI-related stocks, contributing to global market resilience, with benefits extending beyond tech to utilities and infrastructure.
• Capital One is exploring alternatives to Amazon Web Services (AWS) for its AI workloads due to rising costs, considering "neo-clouds" like CoreWeave and Lambda, which often use Nvidia hardware.
• Reflection AI, co-founded by former DeepMind researchers, secured $2 billion in funding and is valued at $8 billion with NVIDIA's backing, aiming to compete with firms like OpenAI. Dr. Jin Wu (ex-Google) joined the company.
• Bosch, in collaboration with Microsoft and NVIDIA, introduced an AI extension platform for vehicle cockpits, utilizing the NVIDIA DRIVE AGX Orin chip and integrating Microsoft 365, projecting over 2 billion in sales by 2030.
• The US Air Force is upgrading its X-62 VISTA AI fighter jet with new mission systems, including advanced radar like Raytheon's PhantomStrike, to enhance AI decision-making capabilities.
• Agentic AI, which enables autonomous task execution, is transforming automotive industry workflows and has been added as a new module to Interview Kickstart's Advanced Machine Learning Course.
• Google filed a lawsuit against SerpApi for scraping and reselling its search engine results, citing the increasing value of this data for AI applications and potential violations of website rights.
• AI models from Google DeepMind and Meta's FAIR team are making significant progress in solving complex math problems, though human guidance remains essential for major breakthroughs.
• Generative AI presents substantial potential for African development, including poverty reduction and education improvement, but requires careful regulation and inclusive access.

AI Models Battle on Wall Street Grok-4.20 Wins

In late 2025, an AI trading competition called Alpha Arena 1.5 took place on Wall Street, organized by Nof1.ai. Powerful AI models competed in the real US stock market, each starting with US$10,000 to trade autonomously for two weeks. Elon Musk's Grok-4.20 from xAI won, making a 12.11% profit. Other models, including OpenAI's GPT-5.1 and Google's Gemini-3-Pro, lost money. This event tested if AI is ready to manage real money and raised important questions about who is responsible when algorithms make financial decisions.

2025 Market Sees Strong AI Stock Growth and Global Returns

The year 2025 showed strong market resilience with global stocks delivering good returns. Bonds also served as stabilizers, benefiting from Fed rate cuts. Investing in AI was a major theme, driving significant growth in US AI-related stocks due to strong earnings. Many financial advisors, however, remain cautious about technology allocations despite being optimistic about AI. Beyond tech companies, utilities and infrastructure firms are also benefiting from increased energy demands for AI. International stocks performed well, with emerging markets seeing their own AI leaders and developed markets boosted by defense and financial sectors.

US Air Force Upgrades X-62 AI Fighter with New Radar

The US Air Force is upgrading its experimental X-62 VISTA AI fighter jet at Edwards Air Force Base. This modified F-16D test aircraft will receive new mission systems, including advanced radar and sensor integration. These upgrades will allow the X-62 to handle more complex scenarios and test AI systems for real-time decision-making. Raytheon's PhantomStrike radar, a compact and cost-effective fire-control system, is a major part of this enhancement. The X-62 VISTA helps bridge the gap between human-piloted and future uncrewed combat aircraft, building on past successes like the 2023 human versus AI dogfight.

Capital One Explores AWS Alternatives Due to Rising AI Costs

An internal Nvidia memo reveals that Capital One is worried about its increasing artificial intelligence costs with Amazon Web Services. The bank discussed exploring alternatives to AWS, including building its own "AI factory" or using "neo-clouds" which specialize in AI workloads. Neo-clouds like CoreWeave and Lambda often use Nvidia hardware. This situation highlights a growing concern among companies about the high costs of cloud computing for AI development. While Capital One stated its commitment to AWS, other AI startups are also seeking alternative cloud providers due to pricing.

Generative AI Offers Big Promise for African Development

Generative artificial intelligence holds great potential for development across Africa, from classrooms to farming communities. This technology could help reduce poverty, improve education, and preserve languages. However, it is important to ensure these benefits reach everyone, not just those already connected. Experts suggest regulating gen AI carefully and engaging with its potential while being cautious of corporate claims. Disregarding gen AI could leave many communities behind, so more discussions are needed between AI developers and those working on UN Sustainable Development Goals.

Purdue Alumnus Jin Wu Joins AI Startup Reflection AI

Dr. Jin Wu, a Purdue Polytechnic Ph.D. graduate, has joined the leading artificial intelligence startup Reflection AI. He previously worked at Google for three and a half years. Reflection AI, co-founded by former DeepMind researchers Ioannis Antonoglou and Misha Laskin, recently secured $2 billion in funding and is valued at $8 billion, with NVIDIA's backing. The company aims to develop large-scale open-weight AI models to compete with established firms like OpenAI. Dr. Wu's research at Purdue focused on invariant signatures theory, which helped solve costly interoperability issues in the construction industry.

Bosch Aims for 2 Billion Euros in AI Cockpit Sales by 2030

Bosch has introduced a new AI extension platform designed to upgrade vehicle cockpit systems. Developed with Microsoft and NVIDIA, this unit adds artificial intelligence features like voice assistance and interior scene understanding without needing new hardware. It uses the NVIDIA DRIVE AGX Orin chip for powerful computing. Bosch plans to integrate Microsoft 365, allowing features like voice-activated Microsoft Teams calls in cars. The company expects its in-vehicle infotainment sales to exceed €2 billion by 2030 and will showcase the technology at CES 2026.

AI Tackles Tough Math Problems But Humans Still Lead

Artificial intelligence is making progress in solving complex math problems, raising questions about its potential to outperform human mathematicians. In October 2024, Meta's FAIR team used AI to solve a long-standing problem by finding Lyapunov functions, improving on previous algorithms. Google DeepMind also developed an AI capable of solving various math problems. While mathematicians acknowledge AI's rapid development, they note that current AI still needs human guidance and has not yet achieved a breakthrough like AI beating human champions in chess or Go. Experts like Terence Tao and Kevin Buzzard believe AI will soon tackle many unproven math hypotheses.

Interview Kickstart Adds Agentic AI to Machine Learning Course

Interview Kickstart has updated its Advanced Machine Learning Course with a new "Agentic AI" module. This course helps professionals develop practical AI skills to meet the growing demand for talent. The Agentic AI module focuses on building and using autonomous AI systems that can perform complex tasks with little human help. Engineers from top tech companies like FAANG teach these live online classes. Interview Kickstart aims to bridge the gap between academic learning and real-world AI application, preparing students for cutting-edge AI development jobs.

Agentic AI Transforms Automotive Industry Workflows

Artificial intelligence is rapidly changing the automotive industry, as discussed in a recent webinar with Cox Automotive's Chief Data Officer, Ben Flusberg. Agentic AI, a type of AI that performs complex tasks autonomously, is improving everything from retail to finance. It enhances inventory management, personalizes customer experiences, and automates workflows for dealers and automakers. Implementing agentic AI responsibly requires clear rules, transparency, and human oversight. Businesses that adopt these technologies wisely will gain a competitive advantage and create better customer experiences across the entire vehicle lifecycle.

Google Sues SerpApi Over Search Result Scraping

Google has filed a lawsuit against SerpApi, a company that scrapes and resells its search engine results pages. Google claims SerpApi ignores its terms and legal boundaries, especially as this data becomes more valuable for AI. SerpApi provides this data to companies like Perplexity for use in chatbots. Google argues that SerpApi's actions violate the rights of websites it indexes and that SerpApi has increased deceptive scraping behavior. This lawsuit protects Google's business interests and may encourage AI companies to use official data sources from other search engines like Brave and Bing.

Sources

• Trading by algorithm: Who is responsible when AI calls the shots?
• AI stocks, alternatives, and the new market playbook for 2026
• Radar, other upgrades planned for experimental US Air Force AI fighter
• Nvidia memo says Capital One discussed alternatives to AWS as AI costs could 'get out of hand'
• Engaging generative artificial intelligence in African development
• Purdue Polytechnic alumnus joins leading artificial intelligence startup Reflection AI
• Bosch targets €2bn sales with AI cockpit systems by 2030
• AI is solving 'impossible' math problems. Can it best the world's top mathematicians?
• Leading Advanced Machine Learning Course by Interview Kickstart Gets Agentic AI Update - LIVE Online Classes by FAANG+ AI/ML Engineers
• The AI Revolution Is Now: Agentic AI for End‑to‑End Automotive Workflow Automation
• Google lobs lawsuit at search result scraping firm SerpApi