Highlighted Companies
KEY COMPANIESSalesforce reported full-year FY2026 revenue of $41.5B (+10% YoY) with Agentforce closing 29,000 deals in Q4 alone — up 50% quarter-on-quarter — and delivering 2.4 billion agentic work units across Agentforce and Slack. The $800M ARR for Agentforce proves enterprise adoption is real. Yet the stock has fallen 38% in 2026 as investors focus on FY2027 guidance of $45.8–46.2B, implying just 10–11% growth — a deceleration that signals customers are buying AI but rationing expansion budgets. The Agentforce story is winning on volume; the pricing power thesis remains unproven.
Read More →Microsoft and OpenAI announced a sweeping restructuring of their partnership on April 27. The headline: Azure exclusivity is gone. OpenAI can now distribute its models through any cloud provider — including Amazon and Google — a structural shift that expands OpenAI’s addressable market while reducing Microsoft’s strategic moat. Microsoft retains a license to OpenAI IP through 2032 and continues receiving revenue-share payments through 2030. The AGI clause — which would have frozen Microsoft’s access if OpenAI achieved AGI — was removed. Microsoft stays a major shareholder. The deal reframes the relationship from exclusive vendor lock-in to strategic co-investor, and puts OpenAI on a clear path to its planned IPO.
Read More →Anthropic is in active discussions to deploy Microsoft’s Maia 200 custom AI accelerator as part of a broader chip diversification strategy. This follows Anthropic’s $5B investment from Microsoft and its $200B Google Cloud infrastructure commitment announced earlier this month. The Maia 200 talks signal that hyperscaler-custom silicon is becoming a credible complement to NVIDIA GPUs — particularly for inference workloads where cost-per-token economics matter more than raw training throughput. If Anthropic standardizes on Maia for portions of its Claude inference stack, it would validate Microsoft’s silicon strategy and accelerate the commoditization of GPU inference.
Read More →AI & Semiconductors
AI & CHIPSGoogle Cloud Next 2026 opened with CEO Thomas Kurian declaring the AI pilot era over. Vertex AI was rebranded as the Gemini Enterprise Agent Platform, absorbing Agentspace into a single product. The eighth-generation TPU — the first to feature two distinct chips — was designed specifically for agentic AI workloads. Workspace Studio, a no-code agent builder across Gmail, Docs, Sheets, Drive, and Meet, was announced. The Agent-to-Agent (A2A) protocol went live, enabling AI agents from different vendors to coordinate tasks. Forrester called it Google’s clearest full-stack AI bet against OpenAI and Anthropic to date.
Read More →Salesforce and Google Cloud announced a deep integration at Google Cloud Next 2026 that allows Agentforce agents to operate across both platforms with full context — pulling CRM data from Salesforce while reasoning with Gemini models running on Google Cloud infrastructure. Gemini-powered reasoning for Agentforce is generally available as of May 2026. The partnership is the clearest evidence yet that the AI agent ecosystem is converging around interoperability rather than walled gardens, with Salesforce’s 150,000-customer distribution combined with Google’s model and compute stack creating a formidable joint offering against Microsoft Copilot.
Read More →Anthropic opened public beta access to its autonomous managed agent framework, enabling long-running workflows in coding, finance, and legal domains where sub-agents coordinate and evaluate work using rubric-based outcomes. Separately, Anthropic is paying $1.25 billion per month to SpaceX for computing capacity through May 2029 — a $45B compute commitment that dwarfs any single infrastructure deal in AI history. Combined with the $200B Google Cloud commitment and the $900B valuation fundraise, Anthropic is making the most aggressive infrastructure bets of any private company ever, predicated on the view that Claude enterprise deployments will generate returns at sufficient scale.
Read More →Markets & Tech Stocks
S&P 500 · NASDAQ · MARKETSWith 93% of S&P 500 companies reported, Q1 2026 earnings growth has hit 28% year-over-year — the strongest result since Q4 2021. 83% of companies beat EPS estimates and 81% beat revenue estimates (vs. a 70% five-year average). Tech leads: standout EPS beats include NVIDIA ($1.87 vs. $1.75), Microsoft ($4.27 vs. $4.06), Sandisk ($23.41 vs. $14.62), and Intel ($0.29 vs. $0.02). Analysts project continued double-digit growth through Q4 2026, with full-year 2026 EPS growth consensus at 21.3%. The AI capex supercycle is flowing directly through to earnings — equipment suppliers and hyperscalers alike.
Read More →The S&P 500 entered 2026 at record highs, plunged in February on tariff shock and geopolitical uncertainty, hit a brief bear market in April, then staged a strong eight-week recovery through May 22. The index has now reclaimed prior highs at 7,473. Analysts are debating whether the recovery holds: the bull case rests on AI earnings delivering 20%+ growth through 2026 and the Fed cutting rates in H2; the bear case points to 30-year yields near 5%, a potential credit event from $100B in big-tech bond issuances, and Moody’s U.S. credit downgrade signaling that fiscal risk is being mispriced. The path to year-end depends on whether AI earnings growth exceeds the cost of capital.
Read More →The Big Four hyperscalers have collectively issued $100 billion in bonds year-to-date to help fund their combined $725B capex program, and bond market investors have responded by purchasing record levels of Credit Default Swaps — effectively insurance against potential defaults. The dynamic reflects a fundamental tension: AI capex is consuming a historic share of operating cash flow (hyperscalers projected to spend 92% of OCF on capex in 2026, vs. 41% in 2023), requiring external financing at a moment when the 30-year yield recently hit its highest point since 2007. If AI ROI takes longer than expected to materialize, the financing structure becomes a vulnerability.
Read More →Supply Chain & Commodities
CHIPS · MATERIALS · FREIGHTAmazon ($200B), Alphabet ($175–185B), Meta ($115–135B), and Microsoft ($120B+) collectively plan $725B in capital expenditure for 2026, up 77% from $410B in 2025 — the single largest concentrated infrastructure cycle in the history of technology. Roughly 75%, or $450B, is directly AI-related (GPU servers, AI-specific data centers, networking, cooling). Oracle is targeting an additional $50B. The scale is so large that traditional infrastructure metrics have become inadequate: hyperscalers are now analyzed as sovereign-scale infrastructure builders, not cloud software companies.
Read More →The Information reported that Anthropic’s Google Cloud commitment totals $200 billion over five years for 5 gigawatts of compute capacity — a figure that dwarfs any prior enterprise cloud contract. At $40B per year, Anthropic’s Google Cloud spend alone would rank as a major enterprise technology budget at corporate scale. The commitment is partly the result of Anthropic’s $5B investment from Google. The deal cements Google as the primary infrastructure layer for the world’s most valuable private AI company, provides Google Cloud with a demand anchor for its aggressive data center buildout, and creates a single-point-of-concentration risk for Anthropic’s compute strategy.
Read More →Dell’Oro Group reports that global data center capital expenditure is on track to exceed $1 trillion in 2026, growing more than 50% from $726B in 2025. As recently as 2025, Dell’Oro had forecast this milestone for 2029. The acceleration is entirely AI-driven: the power grid, CoWoS advanced packaging, and HBM memory remain the three binding physical constraints on even faster deployment. Microsoft’s disclosure that $25B of its 2026 capex increase is attributable specifically to rising memory and storage component prices underscores how HBM scarcity is inflating the cost of every AI training and inference cluster being built.
Read More →