☀️ Morning Briefing
Highlighted Companies
KEY COMPANIESServiceNow posted Q1 2026 subscription revenue of $3.67B (+22% YoY) and raised full-year guidance to $15.74–15.78B, but the stock cratered on ~75bps of headwind from delayed large on-premise deals in the Middle East amid ongoing regional conflict. The miss exposed how concentrated enterprise AI pipelines remain in geopolitically unstable regions — and raised questions about whether “deal slippage” is a one-quarter story or a structural drag.
Read More →Five of the Magnificent Seven report this week, with the market’s primary test being whether AI monetization — 365 Copilot, Gemini Cloud, AWS Bedrock — is generating revenue proportional to the hundreds of billions in data center capex committed through 2026. Alphabet is expected at $2.64 EPS on $92.2B revenue (+20.6% YoY); Meta at $6.65 EPS on ~$55B (+32% YoY); AWS consensus at $36.8B (+25.6% YoY). Any cloud guidance miss will move the tape hard.
Read More →Google’s Agent2Agent interoperability protocol has cleared the proof-of-concept phase: 150 organizations are now running A2A in production, with Salesforce Agentforce handing off tasks to Google Vertex agents querying ServiceNow for IT asset data — all without shared internal architecture. This is the emerging standard battle for the enterprise AI OS layer, and every major platform vendor is now committed to it, narrowing the window for proprietary lock-in strategies.
Read More →Despite Salesforce reporting $800M Agentforce ARR and 29,000 deals (up 50% QoQ), CRM is off ~31% YTD as investors weigh whether Microsoft Copilot and OpenAI agents can erode the CRM moat that justified the valuation premium. ServiceNow is similarly off ~40%. The market is no longer rewarding AI deal volume — it wants monetization evidence and margin resilience, neither of which is yet definitive.
Read More →Microsoft reports Q3 FY2026 on April 29. Last quarter Azure grew 40% (39% constant currency) and cloud gross margin compressed to 68% as AI infrastructure scaled. The question heading into this print is whether demand growth is outpacing the capex drag — Microsoft’s CapEx rose 74% YoY in Q1 FY2026. A second consecutive margin compression quarter could reset expectations for the entire cloud sector.
Read More →AI & Semiconductors
AI & CHIPSNVIDIA has set internal estimates of $1T in combined Blackwell and Vera Rubin architecture chip sales across 2026 and 2027, with Vera Rubin now in production and first products shipping in H2 2026. Anthropic, Meta, Mistral, and OpenAI are all committed to the Vera Rubin platform for next-generation model training and long-context inference. This projection would make it the fastest revenue ramp in semiconductor history, and it underpins every hyperscaler capex forecast for the next 18 months.
Read More →A new Motley Fool analysis argues NVIDIA’s durable risk comes not from traditional chip rivals but from Google TPUs, Amazon Trainium/Inferentia, and Microsoft’s Maia — all of which are now at meaningful scale. Anthropic’s commitment to 1 million Google TPUs in a deal “worth tens of billions” is the clearest signal yet that frontier labs will bifurcate their compute stacks between NVIDIA (training) and hyperscaler ASICs (inference), compressing the total addressable market NVIDIA can claim at premium margins.
Read More →AMD CEO Lisa Su confirmed at CES that the Helios rack-scale AI platform — AMD’s direct answer to NVIDIA’s NVL72 GB200 system — enters production in 2026. Helios is already selected by OpenAI, Meta, Microsoft, and Amazon alongside NVIDIA hardware, meaning the hyperscalers are intentionally multi-sourcing to avoid single-vendor dependency. AMD’s success here depends on software ecosystem maturity (ROCm vs. CUDA) more than raw hardware performance.
Read More →Analyst data shows memory will consume roughly 30% of total hyperscaler capex in calendar 2026, a near four-fold increase from ~8% in CY23-CY24, as DRAM prices surge and HBM remains structurally undersupplied. NVIDIA reportedly receives preferential supply terms well below standard market rates — a significant competitive advantage that is difficult for cloud providers building competing AI infrastructure to replicate in the near term.
Read More →Intel’s stock jumped over 23% to a record high after guiding Q2 revenue at $13.8–14.8B versus analyst consensus of $13.04B, the largest guidance beat in years. The 18A process node — Intel’s most advanced — is drawing serious foundry customer interest from hyperscalers looking to diversify away from TSMC amid geopolitical risk. If Intel can execute 18A at scale, it becomes the first credible U.S.-based alternative for leading-edge AI chip fabrication.
Read More →Markets & Tech Stocks
S&P 500 · NASDAQ · MARKETSThe S&P 500 and Nasdaq both closed at new all-time highs on Monday, driven by the tech rally and an Iran ceasefire extension that reduced geopolitical risk premium. Futures are mixed Tuesday morning as the market pauses ahead of an extraordinarily dense earnings week: Meta, Alphabet, Amazon, Microsoft, and Apple all report by Thursday. The record close sets a high bar — any cloud guidance miss or AI monetization disappointment will be amplified by the elevated valuation backdrop.
Read More →A broad chipmaker rally last Friday, led by Intel’s blowout Q2 guidance, sent the entire semiconductor sector sharply higher in a single session. The magnitude of the moves — Intel’s largest single-day gain in years — reflects how depressed expectations had become after months of tariff anxiety and supply-chain uncertainty. The rally signals renewed institutional confidence that AI infrastructure spending is durable through 2026, not front-loaded.
Read More →With 139 S&P 500 companies through the Q1 reporting cycle, 80% have beaten estimates and aggregate earnings growth is tracking at +12% YoY — a strong season that has provided fundamental support for the index’s record run. General Motors raised 2026 guidance; Coca-Cola beat and raised. The risk now is the back half of the season, where the Mag 7’s massive capex commitments must be reconciled with margin expectations baked into current multiples.
Read More →The 247WallSt analysis this week highlights Adobe, Salesforce, and ServiceNow all significantly underperforming the S&P 500 in 2026, even as their underlying metrics — ARR growth, NRR, RPO — remain healthy. The market is repricing the category on competitive disruption risk from AI-native entrants, not on current fundamentals. For long-term investors, this creates a potential valuation dislocation; for operators, it signals that narrative risk management is now as important as execution.
Read More →The U.S. extension of an Iran ceasefire agreement sent equities higher and reduced the geopolitical risk premium embedded in tech multiples — but ServiceNow’s Q1 miss is a stark reminder that enterprise software pipelines in the Middle East carry real execution risk that macro optimism can obscure. Companies with significant government/defense exposure to the region should be stress-testing their FY2026 pipeline assumptions for continued conflict scenarios.
Read More →Supply Chain & Commodities
CHIPS · MATERIALS · FREIGHTCounterpoint Research confirmed memory price inflation of 80–90% quarter-over-quarter entering 2026 as HBM and high-capacity DDR5 demand from AI infrastructure absorbs capacity that previously served consumer electronics. This isn’t a short-cycle correction — it reflects a structural reallocation of wafer capacity that DRAM makers are not rushing to reverse, as HBM carries 3–5x the margin of standard DRAM and demand visibility is better anchored to multi-year hyperscaler capex commitments.
Read More →HBM’s share of global DRAM wafer starts has reached 23%, up sharply from low single digits two years ago, producing a cascading shortage in standard DRAM and LPDDR for smartphones and PCs. Producing one gigabyte of HBM requires 3–4x the wafer capacity of DDR5, meaning every HBM die allocated to an AI accelerator displaces multiple units of consumer memory. IDC has flagged a potential crisis for smartphone and PC OEMs if HBM allocation continues to crowd out commodity DRAM through H2 2026.
Read More →Reports indicate NVIDIA receives HBM at preferential rates well below standard market pricing, likely through long-term supply agreements with SK Hynix and Samsung locked in during Blackwell’s design phase. This creates a durable cost and availability moat: hyperscalers building competing AI accelerators (Trainium, Maia, TPU) are procuring memory at spot or near-spot rates that are now 80–90% higher than a year ago, directly compressing the economics of their custom silicon programs.
Read More →Tom’s Hardware’s deep-dive on the chipmaking supply chain confirms that leading-edge fab capacity additions remain deliberately constrained — “nobody’s scaling up,” per one analyst — as TSMC, Samsung, and Intel Foundry navigate $10B+ facility investments with multi-year lead times. The conservative posture reflects trauma from the 2022–23 oversupply crash, but it means the current AI-driven demand surge has no rapid supply relief valve, keeping prices elevated and delivery timelines stretched well into 2027.
Read More →TrendForce projects total DRAM sector capex of $61.3B in 2026, up 14% from $53.7B in 2025, with the incremental spend heavily weighted toward HBM process development and advanced packaging rather than commodity DRAM bit supply growth. This means the market is investing to serve AI rather than to relieve the consumer electronics shortage — a deliberate choice that keeps high-margin HBM tight and commodity DRAM structurally constrained simultaneously, compounding the pricing cycle that is already running well above historical norms.
Read More →