$GOOGL is down almost 5% after the company said expanding its use of third-party capacity in Q3 will create modest "near-term margin" pressure as it ramps infrastructure to meet surging AI demand. The added cost comes from leasing external capacity while Google builds out more of its own internal data center infrastructure.

$MSFT previewed its in-house Cobalt 200 chip at Build, underscoring the company’s push to optimize AI and cloud workloads with custom silicon. $NVDA CEO Jensen Huang also joined the conference via livestream, highlighting the deep partnership between Microsoft and Nvidia in the AI infrastructure race.

Throwback to when $NVDA CEO Jensen Huang said investing in $SPCX, Anthropic, and OpenAI today was similar to buying $AMZN, $GOOGL, or $META more than a decade ago. Now he's calling $MRVL a "$1T opportunity," highlighting how the AI infrastructure buildout extends far beyond GPUs alone. His point was that entirely new platform companies are being created in front of us.

$NBIS was initiated at Neutral with a $250 price target by D.A. Davidson, which praised the company’s strong execution, low debt profile, and diversified AI cloud strategy. They project Nebius revenue could exceed $10.6B by 2027 as demand for AI infrastructure continues accelerating. $CRWV had its price target cut to $100 from $175 by D.A. Davidson, which cited valuation, profitability, leverage, and borrowing cost concerns despite maintaining a Neutral rating. They expect CoreWeave revenue to surpass $25B by 2027, though margin pressures and financing costs above 9% remain key risks.

$ENS EnerSys launched its new AlphaCell 4.0HP+ battery designed for outdoor and edge communications deployments, targeting improved electrochemical stability and more predictable replacement cycles in harsh environments. The company says the battery is engineered to reduce premature replacement risk and lower operating costs for telecom infrastructure deployed across hard-to-access edge network locations. The company reports earnings on Wednesday after close.

$AKAM is down almost 4% AH after they announced a $2.6B convertible notes offering split evenly between 2030 and 2032 maturities, with proceeds aimed toward expanding its Cloud Infrastructure Services business, share repurchases, and hedge transactions. The company also authorized roughly $350M in buybacks tied to the offering, signaling continued focus on capital returns while funding long-term AI and cloud infrastructure growth.

$META is reportedly planning a global round of layoffs on May 20, with cuts expected to occur in three waves beginning at 4 a.m. local time, according to internal documents. The company is also said to be restructuring around AI initiatives, including moving roughly 7,000 employees into AI-focused workflow teams while reducing layers of management and implementing broader organizational changes.

$NBIS was initiated at Neutral with a $250 price target by D.A. Davidson, which praised the company’s strong execution, low debt profile, and diversified AI cloud strategy. They project Nebius revenue could exceed $10.6B by 2027 as demand for AI infrastructure continues accelerating. $CRWV had its price target cut to $100 from $175 by D.A. Davidson, which cited valuation, profitability, leverage, and borrowing cost concerns despite maintaining a Neutral rating. They expect CoreWeave revenue to surpass $25B by 2027, though margin pressures and financing costs above 9% remain key risks.

The scale of planned AI infrastructure spending going into 2026 is honestly staggering, with hyperscalers like $AMZN, $MSFT, $GOOGL, and Meta collectively pushing toward nearly $700B+ in combined AI-related capex. The important takeaway is that this buildout continues extending far beyond GPUs alone, reinforcing long-term demand across the entire AI stack tied to compute, networking, power, memory, datacenters, and $CRWV $NBIS $IREN AI cloud infrastructure.

LEOPOLD SCHENBRENNER 13F RELEASED HERES THE SITUATIONAL AWARENESS New Positions: $NVDA $AMD $INTC $TSM $ASML $MU $SMH $GLW $HIVE $TE $SHAZ Added to Existing Positions: $IREN $APLD $CLSK $RIOT $BTDR $CRWV Trimmed Positions: $SEI $CORZ $BE Exited Positions: $COHR $LBRT $TSEM $HUT $LITE $KRC $EQT

$MU more than doubled its HBM market share in just one year, helping drive explosive growth across the memory industry as AI clusters scale globally. Micron, SK hynix, and Samsung are all benefiting from surging HBM demand tied to almost $700B in AI infrastructure spending from $AMZN, $GOOGL, $META and $MSFT.

Just imagine shorting stocks during the dot-com bubble, being fundamentally right but completely wrong on timing. Great video from @WOLF_Financial Now fast forward to today, where so many funds are trying to short $NVDA, $SNDK, and $MU. That becomes a failed opportunity, both financially and mentally. If you spend more time searching for great stocks to buy instead of constantly trying to time shorts, you’ll probably be far more successful in a bull market and in a stock market that historically trends higher over time. Stan Druckenmiller is the GOAT not just because he makes a lot of money when he’s right, but because he knows when to admit he’s wrong and only loses a little when he is. Never had a DOWN year.

The most cyclical stocks appear very cheap during peak cycles. But does that mean we are at a peak? $MU Micron Technology 9.8x $A005930 Samsung Electronics 6.5x $A000660 SK hynix 4.3x $SNDK Sandisk 10.8 $WDC Western Digital 35.3x $STX Seagate Technology 27.6x $P Everpure 29.4x

Should You Be Selling Rips? IWTB with @StockSavvyShay @LogicalThesis Watch the FULL EPISODE ⤵️ Warsh's Market | $SOXX $IGV $SPY $QQQ $TLT AI Bottlenecks | $NVDA $AMD $ARM $QCOM Agents and Software | $NOW $MSFT $PLTR Cybersecurity and Robotics | $PANW $BOT $TSLA Outlook and Closing Thoughts | $TSM $AVGO $MU

$NBIS CBO and co-founder Roman Chernin basically just explained why inference infrastructure may become one of the most important layers of the AI stack over the next several years. Running open-source and alternative AI models at scale is not simply downloading weights and turning GPUs on. The real challenge is extracting reliability, efficiency, and economics from massive distributed compute systems powered by infrastructure tied to $NVDA, networking from $AVGO and $MRVL and hyperscaler compute from $AMZN, $MSFT, and $GOOGL. That is exactly why Nebius continues positioning Token Factory as a full-stack inference platform optimized for agentic workloads, open models, and large-scale AI deployment alongside the broader shift toward inference-driven AI infrastructure.