{"id":113877,"date":"2026-08-11T04:45:11","date_gmt":"2026-08-11T04:45:11","guid":{"rendered":"https:\/\/www.europesays.com\/ch\/113877\/"},"modified":"2026-08-11T04:45:11","modified_gmt":"2026-08-11T04:45:11","slug":"hkex-tech-index-adds-minimax-ubs-says-enterprise-ai-spending-isnt-slowing-in-house-software-builds-become-mainstream-biggo-finance","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ch\/113877\/","title":{"rendered":"HKEX Tech Index Adds MiniMax; UBS Says Enterprise AI Spending Isn&#8217;t Slowing, In-House Software Builds Become Mainstream \u2014 BigGo Finance"},"content":{"rendered":"<p>Hong Kong Exchanges and Clearing (HKEX, 00388.HK) is set to welcome a new tech powerhouse, while the investment logic for global enterprises in artificial intelligence (AI) is undergoing a critical shift. MiniMax (00100.HK) will be added to the HKEX Tech 100 Index after the market close on Wednesday (August 12), becoming a focal point for investors. At the same time, UBS&#8217;s latest enterprise survey reveals that corporate AI spending is not retreating despite cost optimization efforts, but traditional Software-as-a-Service (SaaS) vendors are facing strong headwinds from enterprises building their own AI systems.<\/p>\n<p>According to an HKEX announcement, MiniMax qualifies under the fast-entry rule and will be officially included in the HKEX Tech 100 Index and the HKEX Tech &amp; US Tech 100 Index after the market close on August 12, effective August 13. The exchange noted that as a newly listed tech company in the Hong Kong market, MiniMax&#8217;s rapid inclusion reflects its established market representation and further enriches the HKEX tech indices&#8217; coverage of innovative and technology enterprises. Previously, MiniMax was added to the Stock Connect program on August 6, providing mainland Chinese investors with more options to participate in innovative and tech companies listed in Hong Kong.<\/p>\n<p>Buoyed by the news, MiniMax opened more than 4% higher on Monday (August 11), rising as much as 4.8% intraday to HK$337.8 (approximately $43.06), before paring gains to close up 0.56% at HK$324.2, with 3.048 million shares traded and turnover of HK$1 billion (approximately $128.0 million).<\/p>\n<p>Enterprise AI Spending Shifts to &#8220;Calculating ROI,&#8221; But Budgets Aren&#8217;t Contracting<\/p>\n<p>As MiniMax basks in the index spotlight, the global AI industry is undergoing a deeper structural transformation. UBS analyst Karl Keirstead noted in an enterprise AI survey report released on August 10 that a large number of enterprises have moved from the initial &#8220;token maximization&#8221; phase to a new stage characterized by token optimization, where efficiency and return on investment (ROI) are prioritized. Crucially, this shift &#8220;has not yet translated into a pullback in AI spending.&#8221;<\/p>\n<p>The report shows that the core concern for enterprises has shifted from &#8220;Can AI work?&#8221; to &#8220;Who is using it, how much are they using, and is it worth it?&#8221; Multiple companies have deployed usage tracking systems; once an employee reaches a certain quota, requests automatically switch to lower-tier models with slower response speeds. Some enterprises have even begun restricting employees&#8217; ability to create AI agents. Sensitivity to consumption-based billing models is also rising, as it directly links AI usage to costs.<\/p>\n<p>More granular optimization is occurring at the model invocation level: providing models with more refined context rather than filling the entire context window; reducing repeated calls to RAG systems; and lowering per-task costs through scripting, routing, and prompt compression. One enterprise offered frontier models to customers at a fixed price, only to find that model usage costs eroded profit margins, potentially forcing future pricing structure adjustments.<\/p>\n<p>However, overall budgets have not seen significant contraction. Enterprises generally fear missing the window of opportunity and prefer to push forward with deployment first, then gradually install guardrails. Karl Keirstead emphasized that &#8220;calculating ROI&#8221; is not synonymous with &#8220;halting AI investment,&#8221; at least within this sample of surveyed enterprises.<\/p>\n<p>&#8220;Build&#8221; Logic Dominates; The Inflection Point Traditional SaaS Vendors Are Waiting For Hasn&#8217;t Arrived<\/p>\n<p>Salesforce (CRM) founder Marc Benioff has repeatedly expressed the view that enterprises building their own AI will ultimately fail and will turn to buying mature SaaS vendors&#8217; AI applications. ServiceNow (NOW) holds a similar logic. However, the UBS survey shows that reality has not yet sided with SaaS vendors.<\/p>\n<p>Many Fortune 500 companies still prefer to build their own AI products, driven by three main reasons: highly customized internal processes that off-the-shelf SaaS AI features may not fit; retaining full control over data, workflow orchestration layers, and agent systems; and some external products being overpriced or failing to deliver expected results.<\/p>\n<p>The survey includes a notable case: one enterprise plans to build its own ontology layer using Turtle files, streaming SAP or PTC Windchill data in real-time into a Databricks data lake, undergoing multiple layers of cleansing before making it available for agents to build an action engine, hoping to replace Palantir (PLTR) solutions. Reasons cited include lower costs, easier talent recruitment, and the future ability to open the technology to customers. Another enterprise directly stated that there are currently no SaaS vendor AI features worth buying, as they can develop plug-ins themselves.<\/p>\n<p>This does not mean procurement has completely disappeared. In specific scenarios such as code review, security, observability, SIEM, and workflow, enterprises still purchase mature products. But for software investors, the real signal lies in whether enterprises are willing to pay for these features continuously and forgo the build option. If this preference does not reverse, application software companies&#8217; AI products will struggle to generate high ROI.<\/p>\n<p>Data Management Layer Holds Firm; Databricks Repeatedly Mentioned<\/p>\n<p>In the enterprise AI technology stack, data management companies have a significantly stronger presence than the application software layer. Databricks was mentioned multiple times in the survey, commonly appearing in scenarios involving real-time data lakes, data governance, and AI tool foundations. Snowflake (SNOW) is used for building agents, with surveyed enterprises explicitly pointing out that large language models (LLMs) cannot directly replace Snowflake\u2014LLMs excel at processing language, not at mathematically processing and transforming millions of data points. A more sensible approach is to place models like Claude on top of Snowflake for anomaly observation.<\/p>\n<p>Feedback on Palantir is more mixed. On one hand, it still holds a place in enterprise AI and ontology layers; on the other, some customers are attempting to replace Palantir with self-built AI ontology layers. A large defense contractor mentioned that internal teams have already begun reducing Palantir usage and predicted the company will face different challenges in the coming years. However, the UBS report cautions that such cases remain early signals and are insufficient to extrapolate into a broad trend.<\/p>\n<p>The database layer is also seeing competitive pressure. Some customers are choosing Redis and Postgres for AI workloads, intensifying external attention on whether MongoDB (MDB) can fully participate in AI workloads.<\/p>\n<p>AI-Native Startups Have Entered Corporate Lists; Large SaaS Presence Limited<\/p>\n<p>A noteworthy detail from the survey is that the number of listed application software companies mentioned by enterprises is limited. Atlassian (TEAM), CrowdStrike (CRWD), and Salesforce all appeared, but with low frequency. Instead, a group of small, private AI-native companies are more active on procurement lists: CodeRabbit for code review, Opik for LLM observability, Onyx for agent security, ArmorPoint for AI SIEM, Workfabric for digital workflow twins, and LiteLLM connected to Langfuse for model routing and prompt compression.<\/p>\n<p>These companies are not targeting the macro &#8220;enterprise AI platform&#8221; narrative but rather very specific business pain points. In code review, for example, AI-generated code far outpaces human review capacity; enterprises choose CodeRabbit to prevent security vulnerabilities from entering production environments. In security, AI can prioritize vulnerabilities by combining information such as firewall configurations and user permissions, helping security teams identify truly critical threats.<\/p>\n<p>The survey also points out that AI is rapidly enhancing attack capabilities, and enterprise security-related spending is expected to increase accordingly. Unlike other AI application scenarios, the budget logic for security spending is more rigid, driven by preventing attack surface expansion rather than experiencing new features.<\/p>\n<p>Model Routing Emerges; Chinese Open-Source Models Not Yet Mainstream for Large Enterprises<\/p>\n<p>A natural path to reducing AI invocation costs is routing some tasks from frontier closed-source models to open-source models. However, among this sample of predominantly large non-tech enterprises, OpenAI GPT and Anthropic Claude remain the mainstream choices. Some enterprises explicitly stated they are still using frontier models, though tech-native companies show higher acceptance of Chinese open-source models. Large traditional enterprises, however, place greater emphasis on security, compliance, and controllability, particularly government, defense, and related supply chain companies.<\/p>\n<p>Currently, model cost reduction is advancing along three parallel paths: distributing different tasks to different models, using small models for low-value requests, and optimizing invocation chains to reduce token waste. Chinese open-source models may be one option, but they have not yet become a priority choice for most large enterprises.<\/p>\n<p>Cloud Dominance Unchanged; On-Premise Narrative Lacks Customer-Side Support<\/p>\n<p>Enterprise AI infrastructure remains highly concentrated on AWS, Azure, and Google Cloud. Surveyed enterprises rarely mentioned large-scale expansion of on-premise AI hardware and software stacks. Automotive data, predictive maintenance, agent applications, data lakes, and AI workflows on Snowflake and Databricks are mostly built around the three major clouds.<\/p>\n<p>One enterprise in the survey plans to migrate all data from Azure to Google Cloud, citing the latter&#8217;s more competitive pricing. However, the migration itself is not simple, involving a data transfer cycle of at least one year, as well as compatibility issues between Power BI users, Excel scripts, and Google Sheets. Microsoft-related spending is expected to be reduced but not eliminated. This indicates price competition exists among cloud vendors, but a substantive trend of moving AI from the cloud back on-premise has not yet been sufficiently supported by customer-side evidence.<\/p>\n<p>MiniMax&#8217;s Commercialization Momentum Recognized by Institutions<\/p>\n<p>Returning to the Hong Kong stock market, MiniMax&#8217;s rapid rise similarly reflects the accelerating commercialization trend of AI. China Merchants Securities noted that the company&#8217;s annual recurring revenue (ARR) grew from approximately $100 million in December 2025 to about $430 million in April 2026, with token consumption increasing 152% between February and June 2026. After the release of the M3 model, average daily token consumption on the OpenRouter platform rose from 4.71T to 6.99T, with coding package consumption growing more than tenfold.<\/p>\n<p>Institutional analysts note that the development of leading overseas players similarly indicates that agent workflows, enterprise customer penetration, and token-based billing are becoming important directions for large model commercialization, and MiniMax&#8217;s existing product layout aligns well with industry trends.<\/p>\n<p>From index inclusion to industry research, both threads point to the same trend: the AI industry is moving from a &#8220;token-burning&#8221; infrastructure phase into a &#8220;calculating ROI&#8221; deep commercialization zone. For investors, those who can prove their commercial value at this inflection point will have the opportunity to stand out in the next stage of competition.<\/p>\n","protected":false},"excerpt":{"rendered":"Hong Kong Exchanges and Clearing (HKEX, 00388.HK) is set to welcome a new tech powerhouse, while the investment&hellip;\n","protected":false},"author":2,"featured_media":113878,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[129],"tags":[7801,55461,27118,37384,40739,28122,7802,27570,27044,223],"class_list":["post-113877","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ubs","tag-anthropic","tag-databricks","tag-google-cloud","tag-hong-kong-exchanges-and-clearing","tag-minimax","tag-openai","tag-palantir","tag-salesforce","tag-snowflake","tag-ubs"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@ch\/117075058566949608","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/posts\/113877","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/comments?post=113877"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/posts\/113877\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/media\/113878"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/media?parent=113877"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/categories?post=113877"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ch\/wp-json\/wp\/v2\/tags?post=113877"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}