Signal Tracker

Forward-looking signals on AI, semiconductors, and quantum computing — confidence-scored, source-cited, outcomes recorded publicly. 9 signals published since April 2026. Last updated Jul 24, 2026.

The Discovery & Routing layer is the unclaimed strategic layer of agentic commerce

Artificial Intelligence · Active · Confidence 4/5 · Horizon 12–18 months · Added Apr 25, 2026 · Updated Jul 11, 2026

Thesis: Above identity, policy, and intent, a discovery and routing layer is forming in agentic commerce that no player has yet formally named or governed. Current registries (x402 Bazaar, MPPscan, AgentCash) are optimized exclusively for outcome transactions — reliability, cost, latency — with no commercial implementation able to distinguish outcome intent from experience intent. Whoever controls mode-aware discovery controls demand routing for agent-mediated commerce.

Strategic implication: Merchants whose advantage lies in brand and experience are structurally invisible to current agentic discovery infrastructure. Payment networks, model providers, and merchant platforms will contest this layer within 12–18 months; strategy teams at consumer-facing companies should treat discovery-layer positioning as a first-class architectural decision, not a marketing question.

Outcome-optimized agent rails will commoditize brand-led commerce categories that pass through them

Artificial Intelligence · Active · Confidence 3/5 · Horizon 18–24 months · Added Apr 25, 2026 · Updated Jul 11, 2026

Thesis: Agentic commerce volume is scaling fast ($262B agent-influenced holiday 2025 sales per Salesforce; $8B direct agentic transaction value in 2026 per Juniper, trending toward multi-trillion by 2030–31). Because current discovery infrastructure ranks purely on cost, reliability, and trust score, demand routed through agents will systematically strip brand equity of commercial relevance in categories where consumers would have chosen experience.

Strategic implication: Consumer brands and merchants should model what share of their category flows through agent-mediated channels by 2028 and whether their differentiation survives mode-blind routing. Waiting for the infrastructure to mature means accepting the ranking logic others define.

Enterprise AI lock-in has migrated from the model layer to cloud, data, and orchestration layers

Artificial Intelligence · Active · Confidence 4/5 · Horizon 12–18 months · Added May 9, 2026 · Updated Jul 11, 2026

Thesis: GPT-4-class capability is commodity (token costs down ~300x since 2023), so lock-in has moved to three compounding layers: cloud agent runtimes (AgentCore, Gemini Enterprise, Azure AI Foundry), data control planes (Databricks Unity/MCP, Snowflake Cortex), and workflow orchestration (Agentforce, Copilot Studio). Enterprises making these three decisions independently across business units are accumulating multi-camp lock-in with compounding migration costs.

Strategic implication: Strategy teams must map their camp exposure across all three layers now, before production commitments harden. The camp question belongs in M&A diligence: a target's AI stack portability is now a valuation variable.

MCP + A2A become the de facto portability standard for enterprise agent architectures

Artificial Intelligence · Active · Confidence 4/5 · Horizon 6–12 months · Added May 9, 2026 · Updated Jul 11, 2026

Thesis: MCP adoption in enterprise production went from 31% to 78% of AI teams in a year, SDK downloads hit 97M/month, and both MCP and A2A now sit under the Linux Foundation's Agentic AI Foundation with all major vendors as co-founders. Open-protocol-native architectures will become the benchmark against which vendor lock-in is measured — but protocol portability will not equal operational portability (memory, workflow config, behavioral calibration remain sticky).

Strategic implication: Enterprises should require MCP/A2A compatibility in AI vendor contracts as a floor, while budgeting for the fact that switching costs live above the protocol layer. Vendors' MCP support claims need diligence on what is actually portable.

The quantum stack is vertically integrating — capital is sorting real-revenue players from story stocks

Quantum Computing · Active · Confidence 4/5 · Horizon 12–18 months · Added Mar 28, 2026 · Updated Jul 24, 2026

Thesis: March 2026 marked visible convergence: SEEQC's cryogenic control breakthrough (Nature Electronics), IBM's quantum-centric supercomputing reference architecture, IonQ crossing $130M revenue (60% commercial) and acquiring SkyWater foundry for $1.8B, and the UK's £2B ProQure procurement program. The sector is splitting between vertically integrating revenue generators and capital-rich, revenue-poor speculation vehicles; expect consolidation and supply-chain lock-up to accelerate.

Strategic implication: Enterprises planning quantum posture should evaluate vendors on commercial revenue mix and supply-chain control, not qubit-count announcements. Sovereign procurement programs (UK model) will reshape vendor timelines and geography of capability.

Q-Day compression to ~2029 forces PQC migration mandates to tighten within 12 months

Quantum Computing · Resolved · Confidence 4/5 · Horizon 6–12 months · Added Mar 28, 2026 · Updated Jul 24, 2026

Thesis: Google moved its Q-Day estimate to 2029 (from 2040s consensus), and March 2026 papers compressed quantum attack resource requirements by an order of magnitude. With harvest-now-decrypt-later already operational and enterprise PQC migrations taking 3–5 years, regulators and sector bodies will tighten migration deadlines within the next 12 months, creating immediate commercial demand for PQC tooling.

Strategic implication: Any organization holding data sensitive beyond 2029 should have cryptographic inventory underway now. For strategy teams: PQC readiness is becoming a counterparty-risk and vendor-selection criterion, not just an internal IT project.

Outcome: Confirmed within the 12-month window. On June 22, 2026 the White House signed EO 14412 'Securing the Nation Against Advanced Cryptographic Attacks,' imposing the first binding federal PQC deadlines: key establishment on high-value/high-impact systems by Dec 31, 2030 and digital signatures by Dec 31, 2031, plus a FAR rule (due ~Dec 19, 2026) pulling covered federal contractors inside the same 2030 requirement. OMB executed two days later via memo M-26-15 (June 24): agency migration plans due Oct 22, 2026, TLS 1.3 mandatory by Jan 2, 2030, and NIST's 2035 endpoint preserved as the final phase.

PQC tooling market expands 5–7x in five years, led by crypto-agility platforms

Quantum Computing · Active · Confidence 3/5 · Horizon 18–24 months · Added Apr 24, 2026 · Updated Jul 11, 2026

Thesis: The mandatory, regulation-driven nature of PQC migration creates a multi-year market expanding 5–7x even on conservative estimates. The winning category is crypto-agility platforms and migration tooling (discovery, key management, primitive rotation) rather than any single algorithm or hardware vendor, because large institutions cannot afford infrastructure rebuilds.

Strategic implication: The clearest near-term indicator is which institutions start cryptographic inventory now. Companies enabling discovery and management of cryptographic exposure have a regulation-driven tailwind independent of quantum hardware outcomes.

The Quantum-AI flywheel reaches commercial viability within 18–24 months

Quantum Computing · Active · Confidence 3/5 · Horizon 18–24 months · Added May 23, 2026 · Updated Jul 11, 2026

Thesis: AI and quantum have formed a bidirectional reinforcing loop: AI solves quantum's hardest engineering problems (AlphaQubit error-correction decoding running in production on Willow), while quantum accelerates specific AI workloads (QPA for LLM fine-tuning; UCL's 20-qubit system delivering 20% better accuracy at 100x less memory). With measurable results and capital lining up (Goldman $500M, Novo €188M), hybrid quantum-classical systems reach commercial viability in 18–24 months.

Strategic implication: Financial services and pharma face the earliest exposure — their core problems are structurally quantum-native. Mid-market firms waiting for 'general availability' will be late; the actionable move now is quantum-ready data infrastructure and talent familiarity, not production deployment.

Guaranteed frontier AI capacity reprices upward while commodity token prices fall — workload bifurcation becomes the majority enterprise AI architecture

Artificial Intelligence · Active · Confidence 4/5 · Horizon 12–18 months · Added Jul 15, 2026 · Updated Jul 15, 2026

Thesis: AI compute is splitting into two products with opposite price curves. Guaranteed frontier capacity (reserved GPU access) is repricing upward under physical constraint (record-low data center vacancy, power scarcity) and political constraint (first statewide moratorium signed July 2026, 100+ local moratoria) — AWS's January 2026 15% increase on reserved H200 capacity broke a two-decade cloud deflation pattern. Meanwhile commodity per-token prices for capability-equivalent models keep falling (~50x/year median). The spread drives workload bifurcation: frontier reasoning stays on repricing cloud capacity; standard workloads route to commodity or local execution via the orchestration/gateway layer that shipped in 2026 (Apple on-device routing; Microsoft hybrid platform; LiteLLM/Portkey/Kong gateways).

Strategic implication: Enterprises should split AI budgets into two separately-forecast lines (commodity tokens vs. guaranteed capacity), stress-test 2027 reserved-capacity renewals at +15–30%, and pilot hybrid routing on one high-volume low-sensitivity workload to establish their own local-vs-cloud cost curve before vendor defaults harden.