ASML says it cannot fully meet AI-driven demand for its chipmaking tools in 2027 and 2028, and is seeing "strong interest" in orders for 2029 and 2030
THE SO WHAT
Tool orders stretching into 2029–2030 and ASML admitting it cannot fully meet 2027–2028 demand means structural, not cyclical, tightness in advanced fab capacity. For AI-heavy businesses, this argues for locking in long-term compute access and being realistic about how fast unit costs can fall.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechSEALSQ Presents Research on Side-Channel Attacks Against ML-KEM
Post-quantum crypto is moving from theory to implementation risk—side-channel attacks on ML-KEM mean hardware, not just math, is now in scope. If you're piloting PQC, you need your security team validating concrete implementations, not just checking an algorithm box.
Deep & Emerging TechASML Raises Full-Year Sales Outlook
ASML lifting its 2026 sales floor to €44–45B on AI demand is another confirmation that lithography capacity, not just GPUs, is a binding constraint. If your roadmap assumes cheaper, more abundant advanced-node silicon by 2027, revisit those assumptions—wafer-side bottlenecks will keep pricing and allocation tight.
Deep & Emerging TechMSETCL Said to Pick Banks for Up to $800 Million India IPO
An $800M IPO for a state power transmission utility underscores how much grid build-out capital is coming to public markets. If your growth story depends on electrification—data centers, EVs, industrial loads—watch these issuances; they’re a proxy for how fast the physical grid can catch up.
Deep & Emerging TechSingapore’s PDG Seeks $1.2 Billion Loan for Malaysia Data Center
A $1.2 billion syndicated loan for a single Malaysia data center build shows AI infra is now financed like long-lived energy assets. If you’re planning serious model training or inference capacity, start treating data center partners as part of your capital structure, not just opex vendors.