By the TensorMax editorial team
· Drawing from sources across the AI industry
Today's top story
market event
$80.6B
Micron Technology briefly surpassed Meta and Tesla in market valuation, closing the week at roughly $1.27 trillion.
Why it matters. Micron's market value briefly surpassing Meta and Tesla to reach $1.27 trillion, with a 236% stock increase in a month, underscores the strategic stake of AI memory demand in the tech industry. The company's revenue quadrupled to $41.45 billion, driven by the AI data centre buildout, which has created a shortage of system memory chips. This surge has significant implications for the industry, with major players like Nvidia, Microsoft, and Google buying up large quantities of memory, forcing other companies to hoard it as well, and driving up consumer electronics prices.
Micron Technology briefly surpassed Meta and Tesla in market valuation, closing the week at roughly $1.27 trillion. The stock has soared over 236% in the past month alone, reaching $1,132 a share. This surge followed blockbuster third-quarter earnings, with revenue quadrupling year on year to $41.45 billion and profits jumping from $1.88 billion to $28.2 billion. Micron forecast fourth-quarter revenue of between $49 billion and $51 billion, further fueling Wall Street's bullishness. The driver behind this growth is the AI data centre buildout, which has created a shortage of system memory chips, particularly High-Bandwidth Memory. A single AI server requires significantly more memory than a laptop, and major players like Nvidia, Microsoft, Amazon, Google, Meta, and Oracle are buying up large quantities, forcing other companies to hoard memory as well. The shortage has already driven up consumer electronics prices and is predicted to persist into 2027. To mitigate the risks of a bust cycle, Micron has signed 16 long-term supply agreements, including a multi-year deal with Anthropic covering HBM, DRAM, and SSD supply. According to William Blair analyst Sebastien Naji, demand growth continues to outpace the rate that new cleanroom space can come online, and the company's long-term agreements provide improving revenue visibility. Whether Micron can sustain this growth without a bust cycle remains an open question, with memory having historically been one of the most cyclical businesses in tech. Goldman Sachs has pegged the 2026 DRAM supply-demand gap at 4.9%, the most severe shortage in 15 years.
More from today
model release
$6,220
Why it matters. Meta's launch of Muse Spark, a new internal AI model, is a strategic move to reduce its dependence on external AI providers like Google, which has capped Meta's access to its Gemini model due to compute constraints, affecting several Google clients, including Meta, which has been forced to shift workloads to its own model, with Google itself spending over $180 billion on capex this year and still unable to serve all of its customers' demand, including paying SpaceX $920 million a month for access to 110,000 Nvidia GPUs, highlighting the AI compute shortage that is reshaping relationships between the industry's largest companies, with Meta redirecting billions toward AI infrastructure, including a capex guidance of $115 to $135 billion for 2026, to build internal alternatives capable of handling critical workloads like content moderation at scale, a transition accelerated by the Gemini restrictions, which has also led to Meta telling staff to use AI tokens more efficiently, as the company looks to reduce its external dependence and build its own AI infrastructure, a move that is consistent with the broader industry pattern, where demand for AI compute is growing faster than even the
research paper
Why it matters. The strategic stake of this study lies in its revelation of a significant shift in AI nudification target demographics, with non-celebrity individuals now accounting for 55.8% of targets, indicating an expansion of harm from public figures to individuals within users' own social circles. This shift, driven by the dominance of open-source models such as Stable Diffusion and Wan, which generate 42.7% of images and 66.5% of videos respectively, underscores the urgent need for interventions in platform governance, technical safeguards, and affected individual protection, particularly given that a small cohort of active producers, with the most prolific producing 780 items, drives community engagement and shapes target demographics.
model release
Why it matters. The US clearance of Anthropic's Mythos 5 AI model for wider use, with restrictions on Fable 5 expected to be lifted soon, marks a significant development in the AI industry. With a 5/5 impact rating, this signal has substantial implications for the market. The clearance of Mythos 5, which has demonstrated high performance capabilities, is likely to have a ripple effect on competitors and the broader AI landscape. As the AI market continues to evolve, the ability of models like Mythos 5 to balance intelligence, speed, and cost will be crucial in determining their success.
research paper
Why it matters. The Inattentional Gap in task-conditioned language and vision models poses a significant strategic stake as it can lead to accidents arising from unspecified hazards, despite models scoring near-perfectly on benchmark safety evaluations. Research by Kwan Soo Shin has shown that conditioning models on a narrow task suppresses their reporting of co-present, safety-critical signals, a phenomenon that persists even in larger models and across different tasks, including radiology and driving text scenarios. This dissociation between measured benchmark safety and real-world safety can have severe consequences, highlighting the need for more comprehensive evaluation methods.
model release
Why it matters. The release of Z.ai's GLM-5.2 model, which matches Anthropic's Mythos in cybersecurity capabilities, poses a significant threat to the US's competitive edge in the AI market. With a reported ability to match top US models in some bug-finding scenarios, GLM-5.2 has the potential to reset the AI race, giving China a major advantage in the field. This development is particularly concerning given the US government's restrictions on the use of advanced AI models, which may be handing China an unintended advantage.
funding event
$520B
Why it matters. The strategic stake of this $520 billion chip plant project is to bolster South Korea's position in the increasingly competitive global artificial intelligence industry. With the country's president, Lee Jae Myung, deeming the move as coming at a 'decisive moment', it underscores the importance of staying ahead in the AI transformation of the global economy. This massive investment plan, amounting to 800 trillion won, is a significant commitment to expanding chipmaking capacity, which will be crucial in determining the country's competitiveness in the AI sector.
Catch up quick
-
South Korea announces a colossal investment plan in AI worth over €1,000 billion
-
China's all-CPU LineShine tops the Top500 with custom 304-core Arm CPUs and HBM
-
US government requests OpenAI to limit GPT-5.6 launch and assess model capabilities before full release
-
Nearly 400 local US newspapers sue OpenAI and Microsoft over AI copyright infringement
-
Normal Computing announces tape-out of the world's first thermodynamic computing chip, CN101
-
The US Office of Management and Budget proposes a rule change that could give political appointees veto power over $1 trillion in federal grants for scientific research
-
Claude Code AI coding tool vulnerable to indirect prompt injection, allowing attackers to gain full control of developers' machines via hidden malware in GitHub repositories
-
Netflix releases GenPage, a generative model for constructing personalized homepages
-
Gusto builds a new product line using Claude Code, a permanent Zoom room, and almost no traditional product process
-
South Korea, Samsung, and SK Hynix to invest $590B in new chip complex
-
Rocket Lab to acquire Iridium for $8B
-
Shares in chipmakers SK Hynix and Samsung surge 310% and 183% respectively in the first half of 2026, driving South Korea's Kospi index up 125%
-
NASA is launching a rescue mission to save its aging Swift observatory from reentering the atmosphere, with the help of Katalyst Space Technologies' autonomous robotic spacecraft, Link.
-
Memory prices expected to rise 40%-50% in Q3 2026 and 30%-40% in Q4
-
Online spending across all retailers in the US hit $26.4B during Amazon's four-day Prime Day event, up 9.3% YoY
Also on the desk
Cutting-room floor.Hit reply with one thing you'd delete from today's brief. Filler, repeated beats, a story you skimmed — that's the feedback we act on.
|
|