SanDisk has emerged as one of the S&P 500's best-performing stocks, rising more than 650% in 2026 driven by surging demand for AI memory solutions. The company's shares surged 11% on September 18 alone. Rival Micron Technology also benefited from the AI memory trend, with its stock up 45% over the past year, though SanDisk has significantly outperformed it. SanDisk focuses on flash memory and storage, while Micron offers Optane SSDs for high-end applications and Ultra SSDs for budget-conscious users.
Google is expanding its AI footprint in multiple directions. The company is adding more languages to its AI Brief closed beta and introducing a new reporting feature in Google Ads that gives advertisers a unified view of their Search ads journey. Separately, Google met with South Korean officials on September 9 to discuss selling its tensor processing units directly to the government, which would allow it to compete with Nvidia in the government AI chip market. Google's chips handle complex AI tasks like self-driving cars and medical imaging.
Investment in AI companies continues to accelerate. Bessemer Venture Partners raised $5.75 billion to invest in AI-native firms, with partner Byron Deeter noting that these companies are scaling faster than any previous category of technology. DHH SpA, which grew revenue from €3.9 million in 2016 to €41.8 million in fiscal 2025, announced plans to invest more in AI infrastructure and European data sovereignty. Palo Alto Networks also invested $17 million in developing its new AI security testing service.
AI safety and security are drawing increased attention from both governments and industry leaders. OpenAI, Anthropic, and Hugging Face leaders will brief the UN Security Council on the risk of losing control over advanced AI systems, marking the first time the Council has focused on this risk alone. Palo Alto Networks launched Unit 42 Continuous Frontier AI Defence, a service that constantly tests AI model security using multiple models including Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6-Cyber. During internal testing, the company found a year's worth of security exposures in just three weeks.
Research applying reliability theory to Google DeepMind's AI control defenses found that control stacks can have cubic, quadratic, or linear rare-failure suppression depending on failure domains. Birnbaum importance helps identify which component improvements offer the most reliability gains. Meanwhile, a new LSTM ensemble model predicts river runoff in Iran with R-squared values ranging from 74.95% to 91.42%, helping manage water resources and plan agriculture. Close O Matic also launched an autonomous AI sales platform called Omni Rocket with agent Closora that processes transactions in real time across voice, SMS, and email.
Key Takeaways
- SanDisk stock rose over 650% in 2026, driven by AI memory demand, while Micron's stock gained 45% over the past year
- Google is selling tensor processing units directly to South Korea's government, competing with Nvidia in the government AI chip market
- Bessemer Venture Partners raised $5.75 billion to invest in AI-native companies scaling faster than any prior tech category
- OpenAI, Anthropic, and Hugging Face leaders will brief the UN Security Council on AI safety for the first time
- Palo Alto Networks launched continuous AI security testing using Claude Mythos 5 and GPT-5.6-Cyber, finding a year's worth of exposures in three weeks
- Google is expanding AI Brief with new languages and a unified reporting tool for Google Ads search journeys
- DHH SpA grew revenue from €3.9 million in 2016 to €41.8 million in 2025 and plans increased AI infrastructure investment
- Reliability theory research on Google DeepMind's AI control shows control stacks can have cubic, quadratic, or linear rare-failure suppression
- An LSTM ensemble model predicts river runoff in Iran with R-squared values between 74.95% and 91.42%
- Close O Matic launched an autonomous AI sales platform using Omni Rocket engine and Closora agent for real-time deal processing
Micron and Sandisk battle for AI memory stock gains
Micron and Sandisk are two major players in the AI memory market. Sandisk has outperformed Micron stock over the past year. Micron offers Optane SSDs for high-end applications while Sandisk offers Ultra SSDs for budget-conscious users. Both companies provide high-performance, low-power AI memory solutions.
SanDisk stock soars over 650% in 2026 AI memory boom
SanDisk Corporation stock rose more than 650% in 2026 after an 11% surge on September 18. The company became one of the S&P 500's best-performing stocks. Micron Technology also benefited from the AI memory trend with its stock up 45% in the past year. SanDisk's gains were driven by growing demand for AI memory solutions.
Sandisk and NVIDIA compete as top AI stock buys
Sandisk and NVIDIA are two AI stocks gaining investor attention. NVIDIA is more heavily invested in AI with GPUs used in deep learning and natural language processing. NVIDIA reported net income of $1.4 billion in the first quarter of 2023. Sandisk focuses on flash memory and storage but reported a net loss of $143 million in the same period.
Reliability theory applied to Google DeepMind AI control
Researchers applied reliability theory to Google DeepMind's defenses against rogue AI deployment. The study found that control stacks can have cubic, quadratic, or linear rare-failure suppression depending on failure domains. Birnbaum importance helps identify which component improvements offer the most reliability gains. The results provide guidance on what to separate, improve, measure, and test in AI control systems.
LSTM ensemble model predicts river runoff in Iran
Researchers developed a daily runoff prediction model using Long Short-Term Memory networks for the Ajichay River in northwest Iran. The model combines five LSTM units trained on time intervals from 2 to 6 days and uses Particle Swarm Optimization for fine-tuning. Testing on 2017-2018 data showed R-squared values from 74.95% to 91.42%. The model helps manage water resources, control floods, and plan agriculture.
Google expands AI Max with new languages and reporting tools
Google is adding more languages to the AI Brief closed beta and introducing a new reporting feature in Google Ads. The new tool gives advertisers a single view of their Search ads journey. It shows which search terms triggered ads, which creative assets users saw, and where they landed on a website. The company says more details and availability will come later this year.
Bessemer Venture Partners raises $5.75B to invest in AI companies
Bessemer Venture Partners raised $5.75B to invest in AI-native companies. The firm is known as a leading venture capital investor. Partner Byron Deeter said AI-native companies are scaling faster than any category of technology the firm has backed before. He also noted that VC firms need to keep increasing their funding to meet demand.
DHH reveals decade of growth and plans for AI investment
DHH SpA shared its September 2026 investor presentation on September 23. The company grew from €3.9 million in revenue in 2016 to €41.8 million in fiscal 2025. It now plans to invest more in AI infrastructure and European data sovereignty. The company operates digital services for small and medium businesses across Southern Europe.
OpenAI and Anthropic to brief UN Security Council on AI safety
OpenAI, Anthropic, and Hugging Face leaders will brief the UN Security Council on AI safety. The meeting focuses on the risk of losing control over advanced AI systems. It is the first time the Council has focused on this risk alone. The meeting takes place as global debate over AI safety and responsible development continues to grow.
Close O Matic launches autonomous AI sales platform
Close O Matic launched an end-to-end AI sales platform designed to close and process transactions in real time. The system uses its central engine Omni Rocket and agent Closora to guide buyers from agreement to signed deal. It works across voice, SMS, and email. The platform is built on a Six-Step Top-Producer Sales Psychology Framework created by founder Mark Anthony.
Google offers AI chips to South Korea for national use
Google met South Korean officials on September 9 to discuss selling its tensor processing units directly to the government. This move lets Google sell AI chips to governments instead of only renting them through cloud services. The chips handle complex AI tasks like self-driving cars and medical imaging. This could give Google a new revenue source. It also puts Google in competition with Nvidia in the government AI chip market.
Palo Alto launches always-on AI security testing
Palo Alto Networks launched Unit 42 Continuous Frontier AI Defence, a service that constantly tests AI model security. It scans for weaknesses and fixes them before attackers can exploit them. The service uses multiple AI models including Anthropic's Claude Mythos 5 and OpenAI's GPT-5.6-Cyber. Palo Alto Networks invested USD $17 million in research and development for this service. The company found a year's worth of security exposures in just three weeks during internal testing.
Sources
- Micron vs. Sandisk: Which AI Memory Stock Has More Room to Run?
- This AI Memory Stock Is Up More Than 650% in 2026. Micron Investors Should Pay Attention
- Sandisk vs. NVIDIA: Which AI Stock Is the Better Buy Now?
- Reliability Theory for AI Control
- A Multi-Timestep LSTM Ensemble regressor for Enhanced Short-Term Runoff Prediction
- We’re bringing AI Brief to more languages and adding a new AI Max reporting feature.
- VC firm Bessemer now has another $5.75B to invest in (what else?) AI
- DHH September 2026 presentation: decade of growth, AI investment ahead
- LIVE: OpenAI and Anthropic to brief Security Council as AI guardrails debate intensifies
- Close O Matic Launches Autonomous AI Sales System to Eliminate the Final Transaction Gap
- Google Is Taking Its AI Chips Outside the Cloud. Nvidia Just Got a New Sovereign-AI Rival
- Palo Alto launches continuous AI defence for clients
Comments
Please log in to post a comment.