AI Stocks & Markets

Today's Stories

1. 1

   Lam Research is positioned as a strong buy because its equipment serves the memory, packaging, and AI-complexity demands that will drive semiconductor spending.

   Lam Research manufactures semiconductor fabrication equipment used in memory production, advanced packaging, and AI chip manufacturing. The company's technology addresses multiple layers of demand from artificial intelligence expansion. As AI workloads grow in complexity, chipmakers need both more powerful processors and the specialized equipment to build them—areas where Lam Research is central. Memory and packaging are foundational to every advanced chip, making the company's exposure broad across the semiconductor industry's growth.

   The company's ability to capture share in AI-driven capital spending by semiconductor manufacturers. Lam Research's relevance hinges on sustained investment in the equipment needed for next-generation memory and packaging technologies that power AI systems.

2. 2

   Super Micro Computer and Dell stocks surge on NVIDIA's new Vera Rubin supercomputing platform reveal, positioning both companies to build next-generation AI infrastructure systems.

   NVIDIA unveiled the Vera Rubin platform for scientific supercomputing at ISC High Performance 2026 in Hamburg, with Super Micro Computer and Dell named as global system builders launching custom Vera Rubin NVL4 racks. Super Micro Computer stock rose 11% and Dell rose 5% on the news. Super Micro introduced a liquid-cooled design scaling to 1,152 NVIDIA Rubin GPUs and 576 NVIDIA Vera CPUs per scalable unit, while Dell unveiled the PowerEdge XE8812 server to power the U.S. Department of Energy's next flagship supercomputer. Both companies are positioned to capture orders from the next wave of GPU-based server buildouts following NVIDIA Blackwell Ultra. Dell is already riding strong momentum—Q1 FY27 AI-optimized server revenue jumped 757% year over year to $16.13 billion（約2.6兆円）—while Super Micro is working through a $39 billion（約6.2兆円） AI server backlog after securing $7 billion（約1.1兆円） in financing. These platform announcements represent design wins that could translate to significant booked revenue once systems ship.

   These are design wins ahead of booked revenue. Dell's PowerEdge XE8812 is expected to be globally available early next year, and NVIDIA Vera Rubin NVL4-based systems are expected from manufacturers in Q4 2026. Both stocks are high-beta momentum names that have whipsawed in both directions this year, so investors should monitor whether today's gains hold through the ISC conference week.

3. 3

   Berkshire Hathaway has nearly tripled its New York Times stake under new leadership, betting on a digital subscription business that is growing faster than print is shrinking.

   Berkshire Hathaway added roughly 199% to its Class A stake in New York Times in the first full quarter under Greg Abel. The Times is a $11.7 billion（約1.9兆円） market cap publisher with 13.08 million digital subscribers generating $389.04 million（約620億円） in digital-only subscription revenue (up 16.1% year over year) and digital advertising growing 31.6%. The Times now trades at roughly 4 times sales—a valuation level no AI-era media name comes near—while producing $550.51 million（約880億円） in full-year 2025 free cash flow (up 44.36%), a zero-debt balance sheet, and four consecutive earnings beats. For a Berkshire portfolio heavy in AI winners, this appears to be ballast: a low-beta (0.94) subscription business with durable cash flow and a fortress balance sheet.

   Management expects digital-only subscription revenue up 14–17% and digital advertising up high-teens for Q2 2026. The stock sits below its 50-day moving average of $77.83 and the analyst target of $84.89, with one-year gains of 36.82% lagging Alphabet's 112.95%.

4. 4

   NTT is exploring different AI technology paths through its IOWN AI Fund and building a global ecosystem to diversify its approach beyond current industry trends.

   NTT has launched the IOWN AI Fund to pursue alternative technology directions in AI development and is working to construct a worldwide ecosystem around these efforts. The move signals NTT's strategy to avoid depending on a single technological direction and instead position itself across multiple AI pathways, which may reduce risk if any one approach encounters limitations or shifts in the market.

   NTT's ability to attract partners and developers into this ecosystem will determine whether the fund can establish meaningful alternatives to the dominant AI technology models currently shaping the industry.

5. 5

   DeepSeek's new reasoning model reaches #2 ranking among open-weights models, challenging the AI frontier as capabilities race intensifies.

   DeepSeek released a new reasoning model that has become the #2 open-weights reasoning model. The model supports 1M tokens of context (up from 128K in the prior version) and operates at 27% of the FLOPs compared with DeepSeek-V3.2. Open-weights models—AI systems where the underlying code is publicly available—are increasingly competitive with proprietary alternatives. A high-ranking open-weights reasoning model means researchers and companies outside the major AI labs now have access to capable reasoning technology, shifting where advanced AI capability is concentrated.

   The model's context window expansion to 1M tokens allows it to process substantially longer documents and conversations in a single inference run, which may expand use cases for complex reasoning tasks.

6. 6

   Cerebras approaches its first earnings as a public company with strong demand support from OpenAI and Amazon deals, while execution on chip delivery and capacity gains could drive upside.

   Wedbush analysts are preparing for Cerebras Systems' inaugural earnings report as a public company, noting that demand risk is 'almost zero' given existing deals with OpenAI and Amazon. The key variable to watch is execution—how well management delivers against its own targets. Cerebras has structural advantages that matter to investors. TSMC is expected to deliver at least modest upside to wafer output in 2026 and 2027, which could translate into incremental system sales. The company also benefits from industry-wide memory sourcing constraints, a challenge less acute for Cerebras because its chips rely on SRAM rather than high-bandwidth memory. Management built in conservatism at the IPO stage, meaning simply executing to plan could yield upside to current estimates.

   Cerebras's next-generation WSE-4 chip is flagged as a potential catalyst. While the company has not provided a formal timeline, general speculation points to a late 2026 or early 2027 launch and ramp, with the new design expected to improve both margins and revenue. Wedbush maintains a Buy rating and $270 price target on Cerebras, applying a 40x price-to-earnings multiple to its 2028 EPS estimate of $6.03, plus net cash of $28.19 per share.