Product Roadmap

The future of semiconductor design — built in the open.

COMPLETE Q1-Q2 2025

v1.0 — Foundation

Initial platform launch with core EDA capabilities.

  • RTL synthesis via Yosys/ABC
  • Icarus Verilog simulation
  • Web IDE with real-time log streaming
  • NL-to-RTL generation
  • Design CRUD and file management
  • IP library browser
COMPLETE Q3 2025

v2.0 — Advanced Analysis

Added timing analysis, power estimation, and verification tools.

  • Static timing analysis (ABC)
  • Power estimation (toggle-rate)
  • RTL verification with iverilog compile
  • Clock domain crossing analysis
  • RTL linting via Verilator
  • RTL metrics (Halstead, cyclomatic)
COMPLETE Q1-Q2 2026 — v3.0

v3.0 — Professional EDA Suite

Major expansion with 25+ professional endpoints and real algorithms replacing all stubs.

  • Force-directed quadratic placement
  • A* maze routing with congestion awareness
  • Dynamic + leakage power analysis
  • DRC / LVS verification
  • Clock tree synthesis and analysis
  • Clock gating optimization
  • Pipeline retiming analysis
  • Fanout fix and buffer insertion
  • AI-powered optimization engine (GNN4LS, GraphWISE, RL Router)
  • FSM extraction from RTL
  • Bus interface generation (AXI4/APB)
  • Testbench generator
  • SDC constraint checker
  • IR drop (power grid) analysis (I-GNN)
  • Coverage analysis
  • GDSII layout preview (SVG)
  • PDK mapping (TSMC, SkyWater, GF)
  • Design comparison (PPA diff)
  • Standard cell library browser
  • Netlist export (Verilog/JSON/EDIF)
COMPLETE Q2 2026

v3.5 — Sign-Off Ready PnR

Production-grade place and route — all items shipped.

  • Pure Python OpenROAD-class PnR engine (analytic placement, A* maze router, H-tree CTS)
  • Multi-corner timing analysis (5 PVT corners)
  • Design-for-Test (DFT) scan chain insertion
  • Automatic test pattern generation (ATPG) with D-algorithm
  • Hierarchical design support (RTL hierarchy viewer)
  • Power switch insertion (MTCMOS, 90% leakage reduction)
  • Foundry-calibrated standard cell libraries (in progress)
COMPLETE Q2 2026

v4.0 — AI & Ecosystem

AI-powered EDA pipeline and chip design ecosystem.

  • 9 AI tools: GNN4LS synthesis, HeteroGNN placement, GraphWISE timing, RL Router, I-GNN IR drop, LayoutTransformer DRC, NSGA-III PPA, Design Memory, Multi-Agent Orchestrator
  • APEX X1 chip tape-out (804,800 gates, 8-stage pipeline, 0.79W at 80mm²)
  • X1 demo gallery with synthesised Tensor Sparse Core
  • Foundry PDK certification pipeline
  • IEC 61508 / ISO 26262 safety framework (in progress)
  • X2 architecture specification (12nm, 16K PEs, 8,192 TOPS ternary)
  • Cross-session design memory (200+ JSONL sync to R6515)
CURRENT Q2-Q3 2026

v5.0 — Cloud Marketplace & Advanced Tools

Full-featured cloud platform with marketplace, tapeout, and advanced generators.

  • IP Marketplace (7 seed IPs: RISC-V core, TSC MAC, SOPE Engine, I2C, UART, AES-256, PLL)
  • Multi-project wafer (MPW) shuttle management (sky130, GF180, EFabless)
  • Direct foundry submission API with DRC/LVS/antenna/timing/DFM/IR-drop checks
  • Quantum circuit simulator (Bell states, Shor's algorithm)
  • RISC-V processor generator (7-module, 445 lines, RV32I/RV64I, gshare BPU, cache)
  • APEX Cloud marketplace for IP trading (public launch)
  • Autonomous AI design agents — Natural-language goal → RTL → synthesis → timing pipeline with auto-retry
  • Standard Cell Libraries — 444 cells per PDK (sky130, gf180, asap7) with full timing/power/area characterization
  • X1 Silicon Validation — 7/7 tests passed: synthesis, timing, DRC, power, I/O, clock tree, area
  • SSO + RBAC + Audit — Google/Microsoft/Okta SSO, 4 RBAC roles (admin/designer/viewer/compliance), full audit trail
  • Crown Jewel AI — Multi-provider EDA engine (DeepSeek → Grok → Claude → Ollama fallback)
  • TSMC/Samsung ecosystem membership
  • APEX X1 first silicon
IN PROGRESS Q4 2026 — 2027

v6.0 — Frontier

Next-generation platform with silicon-proven hardware and enterprise certification.

  • APEX X2 architecture specification (12nm FinFET, 16 chiplets, 262K PEs, 8192 TOPS)
  • X2 performance model — batch inference, sparsity, precision-aware throughput estimation
  • X2 power model — PE/SRAM/SerDes/mesh power breakdown with TDP headroom analysis
  • X1 first tapeout (v2.3-rc1: 804,800 gates, 74.5mm², 0.725W @ 358MHz in sky130)
  • IEC 61508 / ISO 26262 certification
  • Full foundry PDK certification (TSMC N5, GF 12LP, Samsung 7nm)
  • APEX X2 hardware (12nm FinFET, optical interposer, NV-SRAM, 2027-Q2)