AI Stocks & Markets

Today's Stories

1. 1

   3 AI stocks outpacing Nvidia's 11% gain in 2024

   Three AI infrastructure stocks—Monolithic Power Systems (up over 40% year to date), Astera Labs, and a third unnamed company—have outperformed Nvidia's 11% year-to-date return. Monolithic Power Systems' revenue grew 26.1% year over year in Q1, with its enterprise data segment nearly doubling; Astera Labs nearly doubled sales year over year in its first quarter. These companies provide critical support infrastructure for AI data centers—power management systems, cooling solutions, and rack-scale connectivity hardware—that hyperscalers depend on to run AI chips reliably at scale. As demand for AI infrastructure builds out, these niche suppliers may see accelerating revenue growth and position themselves to outperform broader semiconductor plays.

   Monolithic Power Systems' enterprise data and communications segments (which together drive most growth) and Astera Labs' sequential revenue momentum; the article suggests double-digit sequential growth in AI hardware has historically preceded sharp stock rallies, as it did for Micron.

2. 2

   Palantir CEO warns AI will make him 20x richer while middle class wages stagnate

   Palantir CEO Alex Karp told the MDMeets podcast that AI could make him roughly 20× wealthier (implying a fortune approaching $300 billion（約48兆円）), while middle-class workers might see their salaries merely double over the next decade. He called this disparity a "complete decoupling of unimaginable wealth and normal wealth." Karp's warning reflects a growing concern among business leaders—including BlackRock CEO Larry Fink and JPMorgan Chase CEO Jamie Dimon—that AI wealth is concentrating among a tiny group of owners of models, data, and infrastructure, while ordinary workers risk being left behind. In 2025, global billionaire wealth surged over 16% to $18.3 trillion（約2900兆円）, three times faster than the previous five-year average, underscoring how the AI boom is accelerating wealth inequality.

   Karp also criticized the "overselling of AI in this country" as "depressing," suggesting the hype around AI's benefits may be masking its real impact on income distribution. His comments echo warnings from other leaders like Nobel Prize laureate Geoffrey Hinton, who has argued that AI under capitalism will make a few people much richer while most people become poorer.

3. 3

   Nvidia Japan project signals major AI infrastructure shift

   A Japanese group is constructing an AI factory with approximately 27,500 of Nvidia's Rubin GPUs and 13,750 Nvidia Vera central processing units, powered by a 140-megawatt system. The project represents a significant infrastructure investment that extends beyond a simple chip sale, indicating growing demand for large-scale AI compute capacity in Japan and suggesting Nvidia's role in shaping how companies build AI systems globally.

   The scale of the deployment—27,500 Rubin GPUs and 13,750 Vera CPUs in a single facility—demonstrates the magnitude of computational resources now required for enterprise AI infrastructure.

4. 4

   Silicon photonics investment surges as AI data centers outgrow copper limits

   As AI clusters expand to tens of thousands of processors, copper wiring that has connected computers for decades is hitting a physical ceiling—copper cables degrade over distances longer than a meter at current signaling speeds, forcing engineers to pump in more power and generate excess heat. Silicon photonics, which sends data as light pulses through fiber instead of electrons through metal, is the fix; the optical interconnect market for AI data centers is expected to grow several times over this decade, with the broader optical transceiver market projected to jump about 60% to roughly $26 billion（約4.2兆円） in 2026. Coherent and Lumentum, which make the lasers and transceivers powering this shift, stand to benefit directly. Coherent has deepened a partnership with Nvidia on next-generation silicon photonics, while Lumentum supplies the optical components and has been expanding capacity to meet demand outpacing supply. However, both stocks have already run up on AI enthusiasm, leaving little room for disappointment, and the optical component business is cyclical and tied to a handful of huge customers whose orders can swing hard quarter to quarter.

   The optical transceiver market is projected to reach roughly $26 billion（約4.2兆円） in 2026. Both Coherent and Lumentum are treated as higher-risk, higher-reward plays on this durable trend; the optical business's volatility should be factored into position sizing.

5. 5

   TSMC outpaces ASML in AI chip boom—higher growth, lower valuation

   TSMC and ASML dominate the semiconductor value chain—TSMC manufactures advanced chips for AI, while ASML is the sole maker of the extreme ultraviolet lithography machines that produce them. In Q2, TSMC's revenue jumped 34% and gross margins expanded 910 basis points to 67.6%, while ASML's revenue grew 21% with gross margins up 30 basis points to 54%. Both companies hold near-monopolies in their segments and benefit from surging AI demand, but TSMC trades at a forward P/E of 19 times 2027 estimates versus ASML's 30.5 times. TSMC's stronger revenue growth and margin expansion, combined with lower valuation, make it the more attractive long-term buy for investors betting on the AI chip ecosystem.

   ASML plans to increase capacity by 30% this year and 30% in 2027, with potential for another 30% bump in 2028, signaling confidence in sustained demand for its lithography machines.

6. 6

   TSM earnings signal AI boom through 2030; analyst eyes Nvidia, Broadcom

   Taiwan Semiconductor Manufacturing reported Q2 revenue growth of 36% year over year (in New Taiwan Dollars) and earnings per share growth of 77%, and announced plans to invest another $100 billion（約16兆円） in Arizona fabrication facilities. CEO C.C. Wei stated that strong chip demand will persist through 2029 or 2030. TSMC's outlook signals the AI hardware buildout is far from complete, which means sustained demand for chips from companies like Nvidia and Broadcom. For investors, this suggests these three semiconductor stocks remain undervalued despite recent gains—TSMC is down about 15% from its all-time high, Nvidia more than 10%, and Broadcom more than 20%.

   Broadcom's custom AI chip division is expected to generate $100 billion（約16兆円） in AI semiconductor revenue next year, up from $10.8 billion（約1.7兆円） last quarter—a figure that signals whether the company can deliver the explosive growth priced into its forward valuation of 19 times FY 2027 earnings.