Home / Themes / AI '25: Model Developers & Platforms
Last updated
Theme thesis · 5/5 sections · Tickers 11 with notes · 3 pending
Bull / Bear Details has the investment thesis and bull/bear points. Overview is monitoring guidance (hiring, forums, second-order trends, search keywords, Google Trends, datasets).
Proprietary foundation models are becoming a critical layer of the artificial intelligence economy as developers compete to provide the reasoning, language, cod
Proprietary foundation models are becoming a critical layer of the artificial intelligence economy as developers compete to provide the reasoning, language, coding, multimodal, and agentic capabilities powering next-generation applications. Model developers and integrated technology platforms can monetize intelligence through APIs, subscriptions, enterprise licensing, cloud consumption, and proprietary applications. Research capabilities, distribution networks, computational resources, and developer ecosystems create potential competitive advantages, although rapidly improving open models, falling inference prices, and enormous development costs threaten profitability.
Expanding Demand for AI Intelligence: Increasing adoption of enterprise agents, coding assistants, multimodal applications, and reasoning-intensive services drives model usage and creates recurring revenue opportunities through APIs, subscriptions, enterprise licensing, and cloud platforms.
Model and Distribution Advantages Compound: Developers with advanced research capabilities, proprietary training data, established developer ecosystems, and large consumer or enterprise distribution networks can capture greater usage, improve model quality, and strengthen platform advantages as AI becomes embedded across digital workflows.
Inference Efficiency Improves Model Economics: Advances in model architecture, training techniques, distillation, and inference optimization reduce computing costs per task, potentially improving gross margins and making proprietary models economically viable across a wider range of consumer and enterprise applications.
Commoditization and Falling Model Prices: Open-weight alternatives, increasingly capable smaller models, aggressive competition, and declining inference prices could erode proprietary model differentiation and prevent developers from sustaining attractive pricing or margins.
Escalating Research and Compute Costs: Frontier-model training, inference infrastructure, specialized engineering talent, and continuous model upgrades require enormous capital commitments, creating risks that revenue growth fails to translate into sustainable profitability or positive free cash flow.
Regulation, Geopolitics and Platform Dependence: AI safety regulation, copyright litigation, data restrictions, export controls, and reliance on cloud infrastructure or distribution partners could constrain model development, limit international expansion, and weaken competitive positioning.
SLM benchmarks, 4-bit quantization, edge inference latency across consumer hardware and NPUs
Enterprise agentic orchestration tooling, protocol adoption, and multi-agent workflow stability
Real-world ROI from agentic rollouts, enterprise security pushback, and migration from cloud LLM APIs to on-premise SLMs
Critiques of AI SaaS pricing models ('seats plus credits'), autonomous agent reliability, and edge device hardware benchmarks
On-device NPU acceleration, computer vision integration, and embedded AI runtime optimizations
Matchedfred · access=api · map_only
Metric/field B985RC1Q027SBEA
Cadence monthly
Why it matters Directly reflects business investment in software, a core component of AI application deployment and a key driver for the theme's growth.
Signal to watch Increasing trend indicates stronger enterprise investment in AI software.
Confidence: high
Matchedfred · access=api · map_only
Cadence monthly
Why it matters Reflects the health and expansion of the services sector, a primary end-market for AI applications and agentic systems.
Signal to watch Sustained readings above 50 indicate expansion, supporting enterprise software spending.
Confidence: low
Matchedfred · access=api · map_only
Metric/field CPILFESL
Cadence monthly
Why it matters Inflation trends influence enterprise budgets and consumer purchasing power, indirectly affecting AI adoption and monetization strategies.
Signal to watch Moderating inflation could free up capital for technology investments.
Confidence: high
Matchedfred · access=api · map_only
Metric/field FEDFUNDS
Cadence monthly
Why it matters Impacts the cost of capital for businesses investing in AI infrastructure and applications, influencing investment decisions.
Signal to watch Stable or declining rates could encourage more aggressive AI investments.
Confidence: high
Not in registryaccess=scrape
Metric/field Euro Area GDP Growth
Cadence quarterly
Why it matters The EU AI Act is a significant regulatory catalyst for the theme; monitoring the economic health of the Euro Area indicates the capacity for AI adoption and compliance spending.
Signal to watch Positive GDP growth suggests a more favorable environment for AI investments and regulatory compliance.
Confidence: high
Not in registry
Metric/field Search interest for 'AI workflow automation'
Cadence daily
Why it matters Indicates growing business and developer interest in automating processes with AI, a core aspect of the theme.
Signal to watch Increasing search interest suggests rising demand and market awareness.
Confidence: medium
Not in registry
Metric/field Stars and commits for LangChain (github.com/langchain-ai/langchain)
Cadence daily
Why it matters Tracks developer activity and community engagement for key AI agent development tools, signaling innovation and adoption trends.
Signal to watch Consistent growth in stars and commits indicates active development and increasing developer interest.
Confidence: low
Not in registry
Metric/field Docket activity for 'AI governance' or 'EU AI Act'
Cadence event_driven
Why it matters Monitors the evolving regulatory landscape for AI, which can significantly impact deployment strategies and enterprise adoption.
Signal to watch Increased activity or new dockets indicate ongoing regulatory focus and potential compliance requirements.
Confidence: medium
Matched (medium)engine_advertising_brand_tracking · access=file · map_only
Metric/field Employee headcount for 'AI/ML Engineer' at Meta Platforms
Cadence irregular
Why it matters Provides insight into major theme constituents' investment in AI talent, indicating strategic focus and resource allocation.
Signal to watch Consistent growth in AI/ML roles suggests aggressive investment in AI development and deployment.
Confidence: medium
Not in registry
Metric/field Full-text search for 'agentic AI' AND 'revenue' in 8-K filings of GWRE, INTU, FICO, IBM, META, NOW, RELX, SPGI, TWLO, U, VERX, MNDY, EFX, EL.PA
Cadence daily
Why it matters Provides real-time insights into how constituent companies are discussing and monetizing agentic AI, and any associated revenue impacts.
Signal to watch Increased mentions of revenue contribution or significant customer wins related to agentic AI indicate successful monetization.
Confidence: medium
Matched (medium)spendhound_b2b_saas_ai_vendor_spend · access=file · map_only
Metric/field Enterprise spending on 'AI application software' by industry
Cadence irregular
Why it matters Directly measures enterprise investment and adoption of AI applications, validating the theme's core premise of value shifting to applications.
Signal to watch Increasing spend indicates broader enterprise adoption and monetization of AI solutions.
Confidence: high
Matched (medium)engine_advertising_brand_tracking · access=file · map_only
Metric/field Employee headcount for 'AI/ML Engineer' and 'Edge AI Specialist' across theme companies
Cadence irregular
Why it matters Offers granular insight into talent acquisition and investment in key AI and edge deployment skill sets by theme constituents.
Signal to watch Consistent growth in these specialized roles indicates strategic investment and expansion in AI capabilities.
Confidence: high
Matchedsimilarweb_website_traffic_engagement · access=file · map_only
Metric/field Web traffic and engagement for 'monday AI work platform' and 'ServiceNow Now Platform'
Cadence irregular
Why it matters Provides insights into user interest and adoption of AI-native enterprise software platforms, indicating market traction.
Signal to watch Increasing traffic and engagement metrics suggest growing user adoption and platform relevance.
Confidence: high
Matchedgartner_it_spending_market_forecasts · access=file · map_only
Metric/field Enterprise Agentic AI Adoption Rate (as per Gartner reports)
Cadence irregular
Why it matters Measures the penetration and successful deployment of task-specific AI agents and AI-native platforms in enterprise applications, confirming the theme's core premise.
Signal to watch Rising adoption rates validate the shift towards agentic AI and its integration into business processes.
Confidence: high
Not in registry
Metric/field On-Device AI Usage Share for consumer-facing AI applications (e.g., Meta AI assistant, Unity AI tools)
Cadence monthly
Why it matters Tracks the scale of Edge AI adoption versus cloud usage, especially in consumer devices and compliance-constrained applications.
Signal to watch Increasing on-device AI usage share indicates successful edge deployment and consumer acceptance.
Confidence: high
Not in registryaccess=api
Market https://kalshi.com/markets/kxmetaa
Why it matters Meta Platforms (META) is a key constituent in the 'AI '25: Apps & Edge Deployment' theme, with its core business increasingly powered by AI and agentic tools to drive user engagement and advertising revenue. A higher daily active people count indicates successful monetization of its AI-driven applications and sustained user adoption, directly validating the theme's focus on AI application value and distribution.
Series key kalshi_meta_daily_active_people_2026
Not in registryaccess=api
Market https://kalshi.com/markets/kxopensourceaiban
Why it matters The 'AI '25: Apps & Edge Deployment' theme identifies 'AI governance, security, and data quality' as critical bottlenecks and risk factors, with evolving laws like the EU AI Act creating a 'regulatory minefield.' A U.S. ban on open-source AI models would significantly impact the development and deployment strategies for many theme constituents, potentially increasing costs, limiting innovation, and altering the competitive landscape for AI applications and edge solutions.
Series key kalshi_us_opensource_ai_ban_2026
Not in registryaccess=api
Market https://polymarket.com/event/will-openai-launch-a-consumer-hardware-product-by
Why it matters The 'AI '25: Apps & Edge Deployment' theme emphasizes the shift to AI applications and monetization, including 'edge deployment strategies' and 'on-device AI.' A major player like OpenAI launching a consumer hardware product would be a significant catalyst for the broader edge AI market, demonstrating tangible productization of AI applications at the user level. This would validate the theme's premise of value shifting to applications and could impact constituent companies involved in AI hardware, software, or services for edge devices.
Series key polymarket_openai_consumer_hardware_launch
| Metric | Cadence | What It Signals | Update Source |
|---|---|---|---|
| Agent-Driven Net Promoter Score (NPS) | Biannual | Gauges end-user satisfaction with AI-assisted experiences in enterprise software | Google_Sheets |
| AI Agent-Driven Revenue Contribution | Quarterly | Shows real monetization of embedded agentic AI features | Google_Sheets |
| On-Device AI Usage Share | Semiannual | Tracks scale of Edge AI Long vs. cloud usage, especially in compliance-constrained apps | Google_Sheets |
| Catalyst | Estimated Timing | Estimated Date Start | Estimated Date End | Why It Matters | Ticker Or Theme Specific | Source Types | Contributing Tickers | Mention Count | Base Score | Source Weight | Specificity Weight | Macro Bridge | Macro Bridge Multiplier | Theme Score | Date Aggregated | Manual Override | Bridge Mention Count | Theme Base Score | Theme Importance Score | Source Type | Catalyst Source | Catalyst ID | Transcript Date |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| The ongoing and significant reduction in AI inference costs (projected 30-40% annually) will make AI application deployment increasingly viable and cost-effective, accelerating widespread adoption and monetization opportunities across various industries. | Ongoing throughout 2026 and beyond. | 2026-06-21 | 2026-12-31 | Declining inference costs are a fundamental driver of the 'AI '25: Apps & Edge Deployment' theme, shifting value from infrastructure to applications. This economic shift fuels widespread adoption and monetization opportunities for all companies deploying AI applications, making it a cross-cutting theme-level catalyst. | Theme | theme_composer | GWRE, INTU, FICO, IBM, TWLO, META, U, APP, RELX, CLS, SPGI | 11 | 0.0567 | 1.18 | 0.92 | 1.0 | 6.1517 | 2026-06-21 | False | 1 | 2.5013 | 271.5391 | Theme composer | ||||
| Key product launches and feature integrations, including Intuit's new AI-driven expert platform lineup, FICO's Mortgage Direct Licensing Program (DLP) go-live with resellers, and AppLovin's AXON Self-Serve launch, will demonstrate tangible progress in AI application and edge deployment. | Q3 2026 to Q4 2026. | 2026-07-01 | 2026-10-31 | These specific milestones validate the core thesis of value shifting to specialized AI applications and optimized hardware/infrastructure. Successful adoption and monetization of these offerings will provide concrete evidence of the theme's progression and potential for revenue generation across multiple constituent tickers. | Theme | theme_composer | INTU, FICO, APP, U, GWRE, META, SPGI | 7 | 0.0548 | 1.18 | 0.92 | Regulatory/Policy | 1.35 | 8.0309 | 2026-06-21 | False | 1 | 1.6201 | 237.4371 | Theme composer | |||
| Updates on the enterprise adoption rate of agentic AI and the realization of measurable ROI, particularly for advanced workflow management systems, will be a critical indicator of the theme's success. | Ongoing, with significant ROI realization projected for late 2026. | 2026-06-21 | 2026-12-31 | The bull case for the theme is predicated on agentic AI driving significant, measurable ROI and productivity gains. Conversely, a bear case highlights that people and processes could become a bottleneck, delaying ROI. Tracking this adoption and ROI across enterprises is crucial for validating the theme's core premise. | Theme | theme_composer | GWRE, INTU, FICO, IBM, META, TWLO | 6 | 0.0468 | 1.18 | 0.92 | 1.0 | 5.0753 | 2026-06-21 | False | 1 | 1.4136 | 153.4569 | Theme composer | ||||
| Key regulatory deadlines and enforcement actions related to AI governance, security, and data privacy, such as the implementation phases of the EU AI Act, will drive enterprise compliance efforts and shape deployment strategies for agentic AI. | The majority of the EU AI Act rules, including those for high-risk AI systems (Annex III) and transparency, will come into force and enforcement will begin on August 2, 2026. | 2026-08-02 | 2026-12-31 | This catalyst is critical as 'AI governance, security, and data quality' are identified as primary bottlenecks to widespread enterprise adoption of agentic AI. Evolving laws like the EU AI Act introduce high-risk obligations and security challenges, impacting how multiple companies (FICO, IBM, META, GWRE) develop and deploy AI applications. | Theme | theme_composer | FICO, IBM, META, GWRE | 4 | 0.051 | 1.18 | 0.92 | Regulatory/Policy | 1.35 | 7.4694 | 2026-06-21 | False | 1 | 0.9907 | 145.1991 | Theme composer | |||
| Closure of the deal combining GFiber with Astound Broadband, leading to GFiber's deconsolidation from Alphabet. | when the deal closes, which we expect to take place in Q4 | 2026-10-01 | 2026-12-31 | This strategic action will impact Alphabet's 'Other Bets' segment, potentially reducing losses from this segment and streamlining the company's portfolio. The financial impact, such as a reduction in operating loss, is a key aspect for investors. | Ticker | earnings_transcript | GOOG (ticker) | GOOG_7d775b3d | 2026-04-29 | ||||||||||||||
| Release of Hunyuan 4, a larger parameter AI model. | later this year | 2026-10-01 | 2026-12-31 | This model is expected to be more powerful, capable of beating larger models, and will provide enhanced intelligence for users, significantly boosting products powered by it and helping Tencent achieve state-of-the-art AI capabilities. | Ticker | earnings_transcript | 0700.HK (ticker) | 0700.HK_b730cc65 | 2026-08-12 | ||||||||||||||
| Initial satellite internet service launch for Amazon Leo (Project Kuiper). | this year | 2026-10-01 | 2026-12-31 | This marks the commercialization of a new, significant long-term business for Amazon, diversifying revenue streams and expanding its ecosystem with meaningful revenue commitments from enterprises and government customers. | Ticker | earnings_transcript | AMZN (ticker) | AMZN_c962c6d7 | 2026-07-30 | ||||||||||||||
| Shift in accounting for Amazon LEO costs, such as satellite manufacturing and launch services, from expensing to capitalization. | Later in 2026 | 2026-10-01 | 2026-12-31 | This accounting change will impact Amazon's reported financial results, potentially improving operating income and free cash flow metrics as significant LEO costs are no longer expensed immediately. | Ticker | earnings_transcript | AMZN (ticker) | AMZN_349295e9 | 2026-07-30 | ||||||||||||||
| Over 20 Amazon LEO satellite launches in 2026. | more than 20 launches planned in 2026 | 2026-01-01 | 2026-12-31 | These launches represent significant capital expenditures and ongoing operational costs, impacting operating income and cash flow, particularly in the North America segment. The success of these launches is critical for the LEO service. | Ticker | earnings_transcript | AMZN (ticker) | AMZN_3e5a0b8b | 2026-02-05 | ||||||||||||||
| Approximately $200 billion in capital expenditures, primarily in AWS, to meet high demand for core and AI workloads. | expect to invest about $200 billion in capital expenditures across Amazon.com, Inc., but predominantly in AWS | 2026-01-01 | 2028-12-31 | These investments are crucial for expanding AWS capacity and maintaining its leadership in cloud and AI, but will impact free cash flow and depreciation, while management expects strong return on invested capital. | Ticker | earnings_transcript | AMZN (ticker) | AMZN_2c1be643 | 2026-02-05 | ||||||||||||||
| Continued global undersupply of AI computing power, particularly in the Chinese market. | over the next 3 to 5 years | 2026-03-20 | 2031-03-19 | This bottleneck could constrain Alibaba's ability to meet accelerating AI demand for its cloud and MaaS businesses, potentially impacting revenue growth. However, as the only cloud compute company in China with proprietary chip development capabilities, it could also be a competitive advantage for Alibaba if they can secure supply. | Theme | earnings_transcript | BABA (ticker) | BABA_5ba488b1 | 2026-03-19 | ||||||||||||||
| Baidu's core AI-powered business revenue exceeding 50% of Baidu General Business revenue. | in the foreseeable future | 2026-03-01 | 2028-02-28 | This signifies a successful transition to an AI-driven business model, indicating strong growth in high-potential areas like AI Cloud, AI applications, and robotaxi, which should positively impact overall revenue growth and valuation. | Ticker | earnings_transcript | BIDU (ticker) | BIDU_b6dab023 | 2026-02-26 | ||||||||||||||
| Achievement of Xiaomi Auto's 2026 delivery target of 550,000 units. | by year-end 2026 | 2026-10-01 | 2026-12-31 | This is a key performance indicator for Xiaomi's rapidly growing EV segment, which achieved positive annual operating profit for the first time in 2025. Meeting or exceeding this target would validate the company's significant investment in EVs and demonstrate continued strong growth in a crucial new business area. | Ticker | earnings_transcript | 1810.HK (ticker) | 1810.HK_078bc81a | 2026-03-24 | ||||||||||||||
| Over 30 Amazon LEO satellite launches in 2027. | more than 30 in 2027 | 2027-01-01 | 2027-12-31 | These launches represent significant capital expenditures and ongoing operational costs, impacting operating income and cash flow, particularly in the North America segment. The success of these launches is critical for the LEO service. | Ticker | earnings_transcript | AMZN (ticker) | AMZN_11192811 | 2026-02-05 | ||||||||||||||
| Launch and strong interest in Trainium four chips. | coming in 2027 | 2027-01-01 | 2027-12-31 | This indicates continued innovation and demand for AWS's custom AI chips, reinforcing its competitive position in the AI infrastructure market and potentially driving future AWS revenue growth. | Ticker | earnings_transcript | AMZN (ticker) | AMZN_7651dd2b | 2026-02-05 | ||||||||||||||
| Xiaomi's electric vehicle business officially enters the European market. | by 2027 | 2027-01-01 | 2027-12-31 | This strategic expansion into a major international market could significantly boost Xiaomi EV's sales volume and brand presence, but also entails substantial investment and operational challenges. | Ticker | earnings_transcript | 1810.HK (ticker) | 1810.HK_44fa7333 | 2025-08-19 | ||||||||||||||
| Alibaba's quick commerce investments to generate positive economic returns for the e-commerce business as a whole. | in two years' time | 2028-03-19 | 2028-03-19 | This would signal a successful culmination of Alibaba's heavy investments in quick commerce, leading to improved overall profitability and validating its strategic importance. | Ticker | earnings_transcript | BABA (ticker) | BABA_d59f0654 | 2026-03-19 | ||||||||||||||
| Alibaba's Quick Commerce business to become profitable. | in FY '29 | 2028-04-01 | 2029-03-31 | Profitability in quick commerce would significantly improve Alibaba's overall margins and free cash flow, demonstrating a successful return on its substantial investments in this strategic area. | Ticker | earnings_transcript | BABA (ticker) | BABA_beb3b9bf | 2026-03-19 |
Earnings Summary
| Date | Type | Comment | Detail | Sentiment | Tickers |
|---|---|---|---|---|---|
| 2026-06-21 | automation_rotation | Recent earnings from Guidewire, FICO, and IBM strongly validate the AI '25 theme, showcasing tangible ROI from agentic AI via new product adoption, platform growth, and significant internal productivity savings. Intuit's challenges in its DIY segment highlight adoption bottlenecks, yet the overall trend confirms value shifting to AI applications and monetization. Proprietary data is emerging as a critical moat, despite ongoing regulatory and operational complexities. | Earnings Summary | Bullish | GWRE, INTU, FICO, IBM |
