51 articles
Amkor Technology reported record results on surging AI server packaging demand, even as semiconductor stocks fell sharply in late July 2026. GlobalFoundries, Kalray, and Skyworks-Qorvo are each expanding capacity or partnerships to meet data center and AI infrastructure needs.
KLA raised its September-quarter revenue guidance to $4 billion and lifted its 2026 wafer fab equipment market outlook to the low-$150 billion range. The company also raised its advanced packaging revenue target to $1.1 billion and announced a 10-for-1 stock split, pointing to sustained AI chip and HBM memory production growth.
Alphabet increased its capital expenditure forecast to $205 billion, citing AI infrastructure buildout. The stock declined after the announcement as investors weighed near-term margin pressure against long-term AI revenue potential.
Four AI hyperscalers plan roughly $650 billion in combined 2026 capital spending, with Alphabet raising its own 2026 capex outlook. Vertiv and Nvidia are co-developing 800-volt DC power platforms to keep pace with data center energy demand, pulling infrastructure vendors into multi-year AI buildout commitments.
Major cloud providers say early AI infrastructure spending is paying off, a signal they plan to keep expanding capex. That reinforces Nvidia's position as the primary chip supplier and supports Dell's fiscal 2027 AI server revenue guidance.
The semiconductor ETF SOXX has risen 79% year-to-date as AI chip demand splits between datacenter GPU buildout and a fast-emerging edge inference tier. InspireSemi's Thunderbird I targets underserved datacenter verticals while Phison's aiDAPTIV collaboration with Intel expands AI memory capacity for local on-device workloads. U.S. export controls on Chinese hardware are compressing competitive timelines across the board.
Nvidia has extended its CUDA platform into quantum computing through CUDA-Q and NVQLink, enabling quantum-classical hybrid systems. The company also launched a generative AI model targeting quantum error correction. This dual positioning insulates Nvidia from compute paradigm risk that pure-play quantum stocks face.
Broadcom collapsed 19.5% in two days following its Q3 FY26 guidance release on June 3, while NVIDIA dropped 6% on June 5. The moves signal a structural shift: AI semiconductor valuations now require guidance that beats consensus by a wide margin, not merely meets it. NVIDIA, AMD, and Marvell face the same dynamic heading into Q2 earnings season.
Adobe, Microsoft, Canonical/Red Hat, Cadence, Siemens, Synopsys, NYSE, CrowdStrike, and Palantir all announced NVIDIA partnerships within a 48-hour window at GTC Taipei. The multi-sector cluster spans creative software, chip design, financial infrastructure, cybersecurity, and enterprise OS. Historical patterns link this type of clustering to NVIDIA data center revenue beats one to two quarters later.
Dell reported $34.1 billion in AI server orders in Q4 FY2026, with its Infrastructure Solutions Group guiding to $13 billion in AI server revenue for Q1 FY2027—implying year-over-year growth above 100%. The figures point to a hypergrowth phase in enterprise AI infrastructure spending with no current signs of demand saturation. Dell's backlog size positions it as a leading vendor as enterprises accelerate GPU-dense data center buildouts.
Intercontinental Exchange has announced a planned GPU Compute Futures product, moving AI infrastructure from a procurement challenge into a regulated financial instrument. The product aims to create price discovery for GPU compute — a resource that underpins nearly all commercial AI development. If adopted, it could open AI infrastructure investment to a broader class of institutional investors.
Analog Devices posted $3.16B in Q1 FY2026 revenue, up 30% year-over-year, while Applied Materials beat earnings expectations in Q2 FY2026. Both companies are riding sustained hyperscaler capital spending on AI infrastructure, with ADI stock up 57.3% year to date. TSMC's entry into the EPIC Platform as a founding partner signals continued fab expansion ahead.
Nebius Group stock jumped 15.72% on Q1 2026 earnings day after the company raised its year-end contracted power target to 4+ gigawatts. A newly secured Pennsylvania AI factory accounts for 1.2 GW of that target. Goldman Sachs reiterated a $205 buy price target, citing three simultaneous tailwinds behind the quarter's revenue beat.
Corning's stock has surged 137.4% year to date following an expanded long-term partnership with NVIDIA. The deal commits Corning to building three new manufacturing facilities to supply fiber optic and specialty glass components for AI data center infrastructure. With 39% of businesses already running generative AI in production, demand for high-bandwidth optical interconnects is accelerating.
Each gigabyte of HBM consumes three times the wafer capacity of standard DRAM, creating a structural supply ceiling that SK hynix and Samsung both project will persist through end of 2027. Gartner forecasts a 125% full-year DRAM price increase for 2026. Micron stock has already surged 162% year-to-date.
Nvidia deployed DGX GB300 AI supercomputing systems to the U.S. Naval Postgraduate School as U.S.-Iran military tensions escalate. The timing suggests accelerated military AI infrastructure buildout, potentially signaling broader defense sector adoption of advanced GPU clusters for computational warfare applications.
Intel has partnered on a massive semiconductor fabrication project while ARM targets $15B in chip revenue, positioning the industry for AI infrastructure expansion. The buildout faces headwinds from financial pressures forcing Wolfspeed to restructure $97M in debt and Chinese rare earth export restrictions complicating supply chains.
The Invesco QQQ ETF delivered a 46% annual return in 2025, driven primarily by AI-focused mega-cap holdings that pushed top constituents above $3 trillion in market capitalization. The concentration mirrors the 1995 internet proliferation phase, raising questions about valuation sustainability as AI infrastructure spending dominates tech sector performance.
Traditional telecom operators are repositioning as AI infrastructure providers through significant capital investments in data center capacity and partnerships with AI compute companies. The shift represents direct competition with hyperscalers in the AI infrastructure market. Operators are targeting new revenue streams from AI services in 2027-2028.
Hardware makers are embedding security and networking directly into data center chips to handle AI workloads without performance penalties. NVIDIA's BlueField-3 DPUs now run full firewall software, while infrastructure operators prepare legacy facilities for GPU-intensive operations. The push reflects enterprise AI adoption driving demand for purpose-built compute infrastructure.