Technology · 10.3

Cloud Services & Data Centres

Quantified news impact based on deduplicated, quality-reviewed events in a single 60-day window.

Average net impact+27weighted per event

Average gross impact32absolute weighted average

Events61same period

Confidence76%model average

Raw intensity63before weighting

Previous period0net impact

Gross changeInsufficient base

Diffusion speed1.0 / day

Concentration2%

Direction disagreement17%

Source independence97%23 sources

IMPACT TREND

60-day effective impact

Positive 52 · Negative 6 · Mixed 3

+360-36

5 Jun0 events

10 Jun0 events

15 Jun0 events

20 Jun0 events

25 Jun0 events

30 Jun0 events

5 Jul0 events

10 Jul0 events

15 Jul0 events

20 Jul0 events

+21

25 Jul38 events

+36

30 Jul23 events

Current periodPrevious period

TIME HORIZON

Impact duration

Immediate3%63 cumulative effective impact

Short term23%452 cumulative effective impact

Medium term65%1279 cumulative effective impact

Long term10%188 cumulative effective impact

CO-IMPACT

Related industries

EVENT EVIDENCE

Events ranked by effective impact

Page 3 of 7 · 61 events

Macro4

Macro31 Jul, 21:26positive

Goldman Sachs Forecasts $765 Billion AI Spending by 2026 as Cash Flows Dwindle; Amazon Capex Raised to $220

Goldman Sachs forecasts that six major tech companies will spend $765 billion on AI in 2026, with total rising to nearly $1.2 trillion in 2027. Amazon raised its full-year capex forecast to $220 billion, while its free cash flow turned negative $7.6 billion. Meta's cash generation fell 91% year-over-year, and Alphabet posted its first quarter of negative free cash flow. Memory chip shortages are driving cost increases, with Tesla CEO Elon Musk calling memory pricing 'crazy' and Amazon CEO Andy Jassy citing 'price increases' for memory chips as a factor in capex guidance.

Effective: +43
Intensity: 70
Confidence: 80%
Horizon: Immediate

Macro28 Jul, 06:55positive

29 Countries Sign Agreement to Establish World AI Cooperation Organization at 2026 Shanghai WAIC

At the 2026 Shanghai World Artificial Intelligence Conference, 29 countries signed an agreement to establish the World AI Cooperation Organization. China will export computing power, technology, and industry experience to help developing countries build data centers, communication networks, and other digital infrastructure. The cooperation covers Asia, Africa, and Latin America, focusing on talent training, infrastructure, technology exchange, and industrial applications.

Effective: +41
Intensity: 70
Confidence: 80%
Horizon: Medium term

Macro27 Jul, 21:52positive

SASAC Releases Second Batch of AI Scenarios and Datasets, Launches Open-Source Community 2.0 and Smart Factory

At the 2026 World AI Conference, China's State-owned Assets Supervision and Administration Commission (SASAC) released the second batch of high-value AI scenarios and high-quality datasets from central state-owned enterprises. It also launched the AI open-source “Huanxin Community” 2.0 and initiated the Joint Foundation Project for an Intelligent Software Factory. SASAC said it will accelerate computing infrastructure construction and promote open-source collaboration.

Effective: +37
Intensity: 70
Confidence: 75%
Horizon: Short term

Macro28 Jul, 08:00positive

2026 Lujiazui Forum Focuses on Finance-Tech Innovation; Tech SME Loan Balance 4.03 Trillion Yuan

The 2026 Lujiazui Forum focused on aligning financial services with technological innovation. By the end of the first quarter of 2026, 303,300 technology-oriented SMEs had obtained loans, a coverage rate of 50.4%. The outstanding loan balance in domestic and foreign currencies reached 4.03 trillion yuan, up 20.9% year on year.

Effective: +36
Intensity: 70
Confidence: 80%
Horizon: Medium term

Industries4

Industries31 Jul, 15:41positive

US tech giants ramp up AI infrastructure spending in July, annual capex plans near $800 billion

In July, major US tech companies rolled out full-year capital expenditure plans totaling close to $800 billion, directing funds toward AI servers, data centers, proprietary chips and memory capacity. The Philadelphia Semiconductor Index initially dropped 11% in one week, with nearly 70% of semiconductor names down over 20% from their highs, before a late-month earnings-driven rebound lifted the index 8.19% in a single session. Alphabet raised its annual capex to $195–205 billion, Microsoft kept its $190 billion plan and Meta lifted its range to $130–145 billion. Intel raised its full-year spending to $20 billion and Tesla reaffirmed more than $25 billion in capex. Memory chip makers and other semiconductor firms secured large cloud orders, driving sharp rallies; SanDisk surged nearly 26% in one day.

Effective: +41
Intensity: 70
Confidence: 70%
Horizon: Short term

Industries27 Jul, 14:51positive

Week of July 27, 2026: Microsoft, Meta, Amazon, Apple to Report Earnings; Microsoft Plans $190 Billion Capex

In the week of July 27, 2026, Microsoft, Meta, Amazon, and Apple will report quarterly earnings. Microsoft plans $190 billion (1900 hundred million) in capital expenditure for 2026, with Azure revenue growth at 39-40%. Amazon's capex approaches $200 billion, with AWS growth at 28%. Meta raised its 2026 capex guidance to $125-145 billion. Apple, with smaller capex, will see Tim Cook's last earnings call as CEO. Google Cloud revenue grew 82%, with backlog of $514 billion; Microsoft's backlog exceeds $600 billion, and Amazon's exceeds $360 billion.

Effective: +40
Intensity: 82
Confidence: 75%
Horizon: Medium term

Industries27 Jul, 11:41positive

Huawei, Sugon Launch 1024-Card Supernodes; Chinese Internet Giants to Scale Deployment by H2 2026

Chinese major internet firms plan to scale up deployment of supernode solutions starting in the second half of 2026. Huawei and Sugon have already unveiled 1024-card systems. Supernodes integrate multiple GPUs into a single logical supercomputer, boosting efficiency and reducing costs. Sugon's 100,000-card cluster at the National Supercomputing Internet Zhengzhou node achieves 90% utilization. The technology delivers 30% to 50% performance gains in PD-separated inference versus traditional setups.

Effective: +39
Intensity: 78
Confidence: 75%
Horizon: Medium term

Industries27 Jul, 10:15positive

Enterprise AI Spending Shifts to Multi-Model Allocation as Token Costs Surge 10-100 Times

Enterprises are shifting from indiscriminate use of frontier AI models to multi-model fine-grained allocation based on task difficulty, driven by rapid adoption of low-cost open-source models. Token costs have surged 10 to 100 times in the short term, with examples including a financial institution's commitment rising from $10 million to $60 million in three months. Chinese open-source models from Zhipu, Moonshot AI, Alibaba, and DeepSeek are widely trialed due to costs at 10% of training and 10-20% of API pricing of overseas peers, while maintaining 20-40% gross margins. OpenRouter data shows US enterprise token usage of Chinese models has stayed above 30% since February 2026, reaching 46% in some periods.

Effective: +38
Intensity: 74
Confidence: 85%
Horizon: Short term

Companies2

Companies27 Jul, 12:11positive

Moonshot AI Unveils Kimi K3 Open-Source Model with 2.8 Trillion Parameters, Raises $3.5 Billion in Series F

Moonshot AI released the Kimi K3 model weights and technical report on July 27, 2026, alongside three key infrastructure technologies: MoonEP, FlashKDA and AgentEnv. The 2.8-trillion-parameter mixture-of-experts model features native visual understanding and a 1-million-token context window. It achieved top open-source scores on benchmarks including the Artificial Analysis Intelligence Index v4.1 and Frontend Code Arena, though it remains slightly behind Claude Fable 5 and GPT-5.6 Sol. On July 29, the company completed a Series F round of over US$3.5 billion at a US$50 billion pre-money valuation. Rapid adoption followed across Huawei Ascend, Alibaba Cloud and global providers like Nebius and Cursor.

Effective: +41
Intensity: 65
Confidence: 85%
Horizon: Short term

Companies31 Jul, 02:53positive

Meta's $700 Billion AI and Cloud Commitments: $349.3 Billion in Irrevocable Contracts, $347 Billion in Future

Meta has committed nearly $700 billion in future spending on artificial intelligence data centers, cloud computing, and related infrastructure. The company disclosed $349.3 billion in irrevocable contract commitments, primarily for third-party cloud services, servers, and network equipment. Additionally, it holds $347 billion in future lease commitments not yet on the balance sheet, with $68 billion added in July alone. Payments for these leases will begin in 2027 and 2028, covering data centers, colocation facilities, and some network infrastructure beyond existing active leases.

Effective: +38
Intensity: 70
Confidence: 80%
Horizon: Medium term