Where the math is defensible.
Long-form research on live enterprise decisions. Publication is selective. Every number traces to a named source. No takes without evidence.
The African data center buildout 2026: Lagos, Nairobi, Cape Town, and Cairo
Africa hosts roughly 1.5 percent of global colocation capacity but is on track to triple installed megawatts by 2028, with Lagos, Nairobi, Cape Town, and Cairo absorbing the bulk of new builds and forcing fresh thinking on power, regulation, and inference latency.
African data center capacity is finally inflecting in 2026. Installed colocation power across the continent now sits near 700 megawatts, still a small share of the global 55 gigawatt market, but contracted pipelines suggest a path to roughly 1.8 gigawatts by 2028. South Africa, Nigeria, Kenya, Egypt, and Morocco anchor that growth. The st...
Compute behind a fence: US AI export controls in 2026
Four years of BIS rules have built a tiered global compute regime. The October 2022 baseline, the October 2023 patch, the December 2024 HBM and tooling rules, and the January 2025 AI Diffusion Framework now shape every chip flow worth tracking. DeepSeek R1 forced the harder question. Are the controls working.
United States chip export controls have moved from a narrow performance ceiling into a full architecture for who gets to compute at frontier scale. The October 7, 2022 BIS rule set the first thresholds. October 17, 2023 closed the H800 and A800 workaround. December 2, 2024 added high bandwidth memory and 24 wafer fabrication tools to the ...
AI inference cost decline 2026: the trajectory and what it forces buyers to plan for
Token prices have fallen roughly 10x per year for equivalent capability since 2023, and the buyers who treat inference as a fixed line item are mispricing every AI roadmap they own.
Inference token pricing has compressed faster than almost any input cost in modern enterprise computing, with frontier model prices falling roughly an order of magnitude per year for any fixed capability tier between 2023 and 2026. The decline is driven by the Hopper to Blackwell hardware step, kernel and serving optimizations, FP8 and FP...
AI inference economics in 2026: GPT, Claude, Gemini, and the pricing war that is rewriting the application stack
Token prices are falling roughly 10x per year at constant capability, the marginal frontier provider is now a Chinese open weight lab, and hyperscaler capex is committing 325 billion dollars to a market where the unit price keeps collapsing.
Inference is the largest unsolved cost line in enterprise AI. Output token prices for frontier general capability have fallen from 60 dollars per million in 2023 to between one and three dollars in 2026, a roughly 20 to 60 fold compression. Epoch AI estimates a 10x annual decline at constant capability, sustained for three years, driven b...
European AI sovereignty in 2026: Mistral, Aleph Alpha, Stargate UAE, and the regulatory framework war
Europe spent 2024 and 2025 trying to build a sovereign frontier through Mistral and Aleph Alpha while the United States staged a 500 billion dollar Stargate buildout and the UAE bought its way into the cap table of every meaningful Western lab. The DeepSeek shock reset the cost floor in January 2025, and the EU AI Act became the only consequential constraint on a market the bloc no longer controls.
European AI sovereignty in 2026 is shaped by three forces. First, the founder champion model has narrowed to a single survivor, Mistral AI, which raised 385 million euros in December 2023 at a 2 billion dollar valuation, then 600 million euros in June 2024 at a 6 billion dollar valuation, while Aleph Alpha announced in November 2024 that ...
Hyperscaler GPU Procurement 2026: H200 vs B200 vs GB200 in Honest Deployment Math
Blackwell is no longer a roadmap promise, it is a procurement reality, and the only honest comparison runs on workload-weighted utilization rather than peak FLOPS. The hyperscalers that win in 2026 are the ones who match SKU mix to inference share, post-training intensity, and the Rubin cadence sitting one fiscal year out.
The 2026 GPU procurement cycle is the messiest in a decade. AWS, Azure, GCP, and Meta are running three NVIDIA generations in parallel while merchant clouds (CoreWeave, Lambda, Crusoe, Nscale) chase liquid-cooled GB200 NVL72 racks at terms designed for sovereign and frontier-lab buyers. The honest math is not B200 versus H100 peak FLOPS, ...
India's Silicon Pivot: ISM 1.0 to 2.0, Tata in Dholera, and the Modi 3.0 Compute Policy Stack
Five approved fabs and ATMP plants, INR 1.55 lakh crore committed, and a planned ISM 2.0 expansion are turning India from chip importer to assembly node. Strategos maps the projects, the binding constraints, and the 2030 capacity envelope.
India's Semiconductor Mission, launched in December 2021 with an INR 76,000 crore (USD 10 billion) outlay, has converted slow paper progress into five Cabinet-approved projects worth roughly INR 1.55 lakh crore between February 2024 and 2025, anchored by Tata Electronics' 28nm fab at Dholera in Gujarat (INR 91,000 crore with Powerchip Sem...
Korea memory in 2026: Samsung versus SK Hynix, NVIDIA qualification, and the HBM share war
SK Hynix turned a two year qualification lead at NVIDIA into roughly half of the global HBM market and most of its profit pool, while Samsung is rebuilding its memory business around a delayed HBM3E ramp, an HBM4 catch up plan, and a foundry separation that signals how serious Suwon now treats the gap.
High bandwidth memory has become the single most concentrated profit pool in the AI compute stack outside NVIDIA itself. TrendForce sized the HBM market at roughly USD 16 billion in 2024 and projects USD 33 billion in 2025, with SK Hynix holding about 53 percent share, Samsung 38 percent, and Micron 9 percent in the fourth quarter of 2024...
Robotaxi economics in 2026: Waymo scales, Tesla bets on a 30 thousand dollar Cybercab, and the ride-hailing margin pool starts to crack
Waymo crossed 200 thousand paid rides per week in late 2024 across four metros, GM wound down Cruise in December 2024, Tesla unveiled the Cybercab in October 2024 with a 2026 production target, and the unit economics of US ride-hailing are about to be repriced by fleets that no longer pay drivers.
Robotaxi service is no longer a demo. Waymo One reported more than 200 thousand paid trips per week by Q4 2024 across Phoenix, San Francisco, Los Angeles, and Austin, and disclosed in February 2025 that it had served more than 4 million paid rides in 2024. GM stopped funding Cruise in December 2024 after a 10 billion dollar cumulative spe...
UAE G42, sovereign AI ambitions, and the US China tech triangulation through 2026
G42 spent 2024 and 2025 converting Abu Dhabi political capital into US compute access, Chinese hardware divorce, and a balance sheet position inside the largest AI deals on the planet. Microsoft put 1.5 billion dollars on the cap table in April 2024, the Bureau of Industry and Security wrote a Diffusion framework in January 2025 that named the UAE explicitly, and Stargate UAE pushed five gigawatts of campus capacity to Abu Dhabi. The story is no longer whether the Gulf builds sovereign AI. The story is whose chips, whose models, and whose security guarantees the buildout runs on.
G42 was incorporated in Abu Dhabi in 2018 as a Mubadala adjacent holding company chaired by Sheikh Tahnoon bin Zayed Al Nahyan, the UAE National Security Advisor and brother of the President. Through 2023 the group accumulated stakes in TikTok parent ByteDance, Huawei surveillance, BGI Genomics sequencing partnerships, and a Wuxi cloud jo...
Trump's AI Action Plan: Compute Sovereignty, Export Control Reset, and the Frontier Regulatory Vacuum
Executive Order 14179, the rescinded AI Diffusion Framework, the Stargate USD 500B compute commitment, and a renamed AI Safety Institute together redraw US AI governance around speed, capital, and selective export pressure rather than precaution.
On January 23, 2025, President Trump signed Executive Order 14179, Removing Barriers to American Leadership in Artificial Intelligence, rescinding Biden's EO 14110 and ordering an AI Action Plan within 180 days. By July 2025 the plan launched with a 60 day Request for Information that drew submissions from OpenAI, Anthropic, Google, Meta,...
AI capex met the grid: when the megawatt curve breaks
Hyperscaler capital spending crossed 500 billion dollars across 2025 and 2026 while the average US interconnection wait sits above 4 years. The constraint is no longer chips. It is megawatts on a calendar.
AI infrastructure capex has cleared 500 billion dollars across 2025 and 2026 between the four hyperscalers, NVIDIA, Oracle, and the new wave of neoclouds. The chips are arriving. The grid is not. Lawrence Berkeley National Lab puts the active US interconnection queue above 2,600 gigawatts, with median wait times above 4 years and rising. ...
Where AI productivity actually shows up: a sector decomposition
Aggregate US labor productivity grew 2.7 percent in 2024 and 1.9 percent annualized through Q4 2025. The AI signal in those numbers is real but narrow. It lives in five sectors and a dozen occupations, not in the headline.
Two years into broad enterprise AI deployment, the productivity question has moved from speculation to measurement. BLS labor productivity data show nonfarm business output per hour up 2.7 percent in 2024 and roughly 1.9 percent annualized through 2025, above the 2007 to 2019 trend of 1.4 percent. BEA multifactor productivity data lag by ...
Quebec hydropower and the new gating of AI compute
Quebec spent two decades selling itself as the cheapest, greenest place on the continent to plug in a data center. In 2026 Hydro-Quebec is throttling new connections, redesigning industrial tariffs, and forcing hyperscalers to rethink where the next gigawatt of AI training capacity actually lands.
Quebec sits on roughly 37 gigawatts of installed hydro capacity, historically the cheapest large-scale clean power in North America. From 2018 the province first banned new bitcoin mining connections, then welcomed AI campuses, and by 2024 began throttling all new large industrial loads above 5 megawatts. Hydro-Quebec's 2035 strategic pla...
The 3.2x compute curve: what GB200 actually changes about training ROI
NVIDIA says 4x. A CFO needs the number closer to 3.2x, and only above 55 percent utilization.
NVIDIA positioned GB200 NVL72 as a 4x training improvement over H100. On a workload-weighted, all-in-cost basis the number is closer to 3.2x, and only if sustained utilization clears a 55 to 60 percent threshold. This brief unpacks the curve, where the gains come from in hardware terms, and the three places the ROI tends to break in real ...