Live database — updated every 6 hours

The live research database for computer hardware.

Track every process node, every cap-ex announcement, every chip tape-out, every fab groundbreaking, every memory supply shift — across CPUs, GPUs, memory, foundries, equipment, and EDA. Built for analysts, engineers, investors, and the advertisers who need to reach them.

340+
Companies tracked
$480B
2026 cap-ex tracked
12
Process nodes monitored
24
Public sources cron'd
Core Database

Companies & Fabs

Fabless designers, IDMs, pure-play foundries, memory makers, equipment vendors, EDA tool providers, and packaging specialists.

Unlock full database →
Company Category HQ Latest Node 2026 Cap-Ex Latest Move Updated
NVIDIAFablessSanta Clara, USTSMC N3B$20B+GB300 Blackwell Ultra volume ramp — CoWoS-L allocation sold out through Q31h ago
IntelIDMSanta Clara, USIntel 18A$25BPanther Lake 18A tape-in complete; Clearwater Forest EUV milestone hit3h ago
TSMCFoundryHsinchu, TWN2 / N1.6$38BArizona Fab 22 Phase 2 N2 groundbreaking; CoWoS capacity expanding +40%2h ago
AMDFablessSanta Clara, USTSMC N3E$12BMI400 architecture confirmed for 2H 2026; CDNA 4 ramp on schedule4h ago
Samsung SemiconductorIDMSuwon, KRSF2 / SF1.4$22BTaylor TX SF2 yield issues delay major customer tape-out by 2 quarters6h ago
SK HynixMemoryIcheon, KRHBM4 16-stack$18BHBM4 16-stack passed NVIDIA qualification; sole supplier through Q2 20272h ago
ASMLEquipmentVeldhoven, NLHigh-NA EUV$10BNXE:3600D High-NA backlog 14 units; TSMC delivery confirmed Q4 20265h ago
Micron TechnologyMemoryBoise, US1-gamma DRAM$14BIdaho greenfield fab Phase 1 breaks ground; HBM4 sampling begins Q4 20268h ago
QualcommFablessSan Diego, USTSMC N3E$6BSnapdragon 8 Elite Gen 2 tape-out confirmed; 3 PC OEM wins announced9h ago
AppleFablessCupertino, USTSMC N2P$20BA20 on N2P confirmed for iPhone 18; locked 60% of TSMC N2 capacity 20267h ago
Arm HoldingsIPCambridge, UKN/A$3BCortex-X9 licensee count hits 11; CSS-2026 platform — 8 chip companies signed11h ago
BroadcomFablessPalo Alto, USTSMC N2$15BXPU revenue run-rate crosses $4B/yr; Google TPU v6 in volume production12h ago
SMICFoundryShanghai, CNN+2 (7nm equiv)$7.5B████████████████████████████locked
Applied MaterialsEquipmentSanta Clara, USN/A$6B████████████████████████████locked
SynopsysEDASunnyvale, USN/A$2.8B████████████████████████████locked
TSMC PackagingPackagingHsinchu, TWSoIC / CoWoS-L$8B████████████████████████████locked
328 more companies behind the paywall. Includes fab-level node data, CHIPS Act grant status, named deal counterparties, executive contacts, and cron-tracked supply chain changes. Start Pro — $29/mo →
Sponsor Slot — EDA vendor / fab equipment OEM / chip IP licensor Reach 20,000+ hardware engineers, tape-out planners, and procurement leads. Apply for advertiser API →
Research Feed

This week in computer hardware

Original analysis sourced from SEC filings, fab disclosures, patent databases, and earnings calls. Updated daily.

Browse archive →
Process Nodes · 6 min read · May 22, 2026

TSMC N2 ramp: why CoWoS-L is the real bottleneck, not the node itself

Advanced packaging capacity is now the binding constraint on GB300 volume. We model the CoWoS-L throughput gap and which customers get priority allocation through Q3.

Memory · 5 min read · May 21, 2026

HBM4 qualification race: SK Hynix leads, Micron closes fast, Samsung struggling

16-stack HBM4 yields at three vendors — and why the gap between first and third is wider than any previous HBM generation transition. Supply implications for MI400.

IDM · 7 min read · May 20, 2026

Intel 18A: what Panther Lake's tape-in actually tells us about yield readiness

Reading between the lines of Intel's process milestone. Three external analysts weigh in on what "tape-in complete" means for the 2H 2026 volume ramp timeline.

Equipment · 4 min read · May 19, 2026

ASML High-NA EUV: 14-unit backlog and why the 2027 N1.6 ramp depends on it

The NXE:3600D delivery schedule, tool productivity at TSMC's Fab 18, and the exposure-dose challenge that keeps process engineers up at night.

Custom Silicon · 5 min read · May 18, 2026

The XPU land grab: why every hyperscaler now has a custom chip roadmap

Broadcom's $4B XPU run-rate is just the visible tip. We map six undisclosed hyperscaler ASIC programs currently in tape-out at TSMC and Samsung.

Cap-Ex · 8 min read · May 17, 2026

$480B in 2026 semiconductor cap-ex — and 60% of it is concentrated in 4 nodes

N2, N3, HBM4, and Intel 18A are absorbing the majority of global cap-ex. The full breakdown by company, geography, and application target.

Futures Lab

2–3 year projections

Forecasts grounded in fab data, patent filings, and observable deal flow. Explicit confidence intervals. Conservative/likely/aggressive scenarios.

Projection archive →
2026 → 2028

TSMC will capture >65% of leading-edge logic revenue below 3nm

Samsung SF2 yield struggles and Intel 18A's slow external customer ramp leave TSMC with structural share dominance through N1.6. Only a credible 2nm-class equivalent from either challenger by late 2028 changes this.

Confidence 82%
2026 → 2027

HBM4 becomes the primary supply bottleneck for AI training at scale

GB300 and MI400 both require 16-stack HBM4. Combined NVIDIA + AMD demand exceeds total qualified supply through Q2 2027. SK Hynix gets first allocation; Micron online by Q3 2027.

Confidence 87%
2026 → 2028

Advanced packaging becomes the primary performance differentiator — not the node

CoWoS-L, SoIC, Foveros Direct, and EMIB are where performance is won. Every leading-edge chip in 2028 is a packaging story as much as a process node story. Substrate capacity will constrain 5+ major product launches.

Confidence 79%
2026 → 2028

Custom silicon market crosses $80B annual revenue — 4× from 2024

Hyperscaler XPU spend, sovereign AI chip programs (EU, India, UAE, Saudi), and edge inference ASICs drive acceleration. Broadcom, Marvell, and 3 undisclosed startups are the primary beneficiaries in this window.

Confidence 73%
Pricing

Pick your access tier

Free preview for everyone. Pro for analysts and engineers. Enterprise for teams. API for advertisers and AI agents. 60–70% cheaper than incumbents — with fresher data.

Free

$0 /forever
  • Top 12 companies preview
  • Weekly research digest
  • 3 article reads / month
  • Futures Lab summaries
  • Ad-supported
Create account

Pro Most popular

$29 /mo
  • Full company database (340+)
  • Process node & cap-ex data
  • Daily research feed (unlimited)
  • All Futures Lab projections + confidence
  • 1,000 API calls / month
  • CSV export
  • Email alerts on tracked companies
Start 7-day free trial

Enterprise

$499 /mo
  • Everything in Pro × 10 seats
  • Slack / Teams integration
  • Quarterly analyst briefings
  • Custom watchlists & dashboards
  • Priority data requests
  • SSO + audit logs
  • White-label dashboard option
Talk to sales

API / AI Advertiser

$0.01 /call
  • Programmatic database access
  • JSON + LLM-ready schema
  • Sponsored placement endpoint
  • Intent signals from search behavior
  • Volume discounts at 500K+ calls
  • Co-branded research slots
  • No subscription required
Get API key
For AI Agents & Advertisers

The API hook for the hardware supply chain

Every record, every projection, every research piece — structured JSON designed for both LLM ingestion and programmatic ad placement. $0.01/call, no subscription required.

# Fetch foundry capacity for targeting
curl https://api.computerpulse.io/v1/companies \
  -H "Authorization: Bearer $CP_KEY" \
  -d '{"category":"Foundry","filter":"active_node"}'

# Sponsored placement (AI advertiser hook)
POST https://api.computerpulse.io/v1/sponsored
{
  "campaign": "eda-tools-Q3",
  "target_categories": ["Fabless", "IDM"],
  "keyword_intents": ["N2 tape-out", "HBM4 qual"],
  "bid_per_impression": 0.022,
  "creative_url": "https://cdn.example.com/ad.html"
}

Why advertisers buy here

ComputerPulse readers are the decision-makers in the global chip supply chain — tape-out planners, foundry procurement managers, EDA license buyers, equipment RFP leads, and the analysts who brief them.

  • 📈 20,000+ verified hardware industry readers
  • 🤖 LLM-readable schema = your brand in AI assistants
  • 🎯 Intent signals from node search + tape-out queries
  • 💼 Direct line to engineering procurement, not marketing
  • 💰 60–70% cheaper than incumbent trade publications
Apply to advertise