BIT Research Report | AI Investment Enters the “Realization Period”—Which Targets Are Generating Real Revenue?
- Core Viewpoint: In 2026, AI investment reaches a turning point. The core logic shifts from models to the actual productivity of AI Agents, with capital expenditure surging (over $600 billion). However, sector differentiation is significant: Infrastructure (Nvidia) and vertical applications (Palantir) become key value capture points. Anthropic's revenue breaks through $30 billion, disrupting the industry landscape.
- Key Elements:
- The 2026 AI Agent market size exceeds $9 billion. Early enterprise workflows accelerate by 20-30%, and backend costs decrease by approximately 25%, marking AI's transition from an experiment to a production tool.
- Anthropic’s revenue jumped from $1 billion to $30 billion in 15 months, surpassing OpenAI. Its Claude Code tool achieved an annualized revenue of $2.5 billion, emphasizing an enterprise-first strategy.
- Nvidia's monopoly is firmly established, with the CUDA ecosystem acting as a moat; however, custom chips (TPUs/LPUs) are creating diversions. Its Vera Rubin platform is expected to reduce inference costs by another 10x.
- Cloud providers' capital expenditure exceeds $600 billion, creating a monetization gap compared to the current approximately $25 billion in direct AI revenue. If ROI disappoints, sector valuations face correction risks.
- Palantir's 2026 fiscal year revenue guidance exceeds $7.1 billion. Its AIP platform is irreplaceable in high-risk scenarios (military, healthcare), functioning as a “cockpit.”
- Anthropic is expected to IPO in October 2026. The gross profit from its AI Agent business will define the valuation center for the SaaS industry. Google Cloud's operational scale exceeds $70 billion, shielding it from AI's impact on search.
- Three core risks: the monetization gap (pressure on capital markets), enforcement of the EU AI Act in August 2026 (sharply increasing compliance costs), and approximately 62% of enterprises remaining hesitant about deep deployment due to “model hallucination.”
Author: BIT US Stock Business Special Analyst Jun
Key Data: 2025 scale of $244 billion | Cloud providers AI CapEx > $600 billion | Anthropic annualized revenue of $30 billion (surpassing OpenAI for the first time) | NVIDIA FY2026 revenue guidance of $215.9 billion
I. Core Investment: Four Pillars of the AI Value Chain
Understanding AI investing goes beyond simply looking at models; the key is to identify who captures value in the chain:
Compute Infrastructure – “Digital Oil Fields”
- Logic: Regardless of who wins, they need chips. NVIDIA (NVDA) maintains its monopoly, but custom chips (TPU/LPU) are siphoning off some market share.
- Representative: NVIDIA
Frontier Models – “Intelligent Power”
- Logic: Intense competition among leaders is evolving towards specialized division of labor.
- Representatives: OpenAI (traffic king), Anthropic (enterprise powerhouse), Meta (open-source disruptor).
AI Agents & Platforms – “Automated Brains”
- Logic: The deciding factor in 2026. AI is no longer just answering questions; it autonomously completes multi-step tasks.
- Representatives: Salesforce (Agentforce), Microsoft (Copilot Studio).
Vertical Applications & Decision Intelligence – “Hardcore Cost Reduction”
- Logic: Integrating AI into industry data to generate quantifiable ROI.
- Representative: Palantir (PLTR) - the preferred choice for defense and government AI.
II. Why 2026 is a Tipping Point "Like No Other"?
2026 marks a watershed moment for AI development, driven by three fundamental shifts:
Quantum Leap in Utility: ChatGPT solved "usability" in 2022, but AI Agents in 2026 solve "utility," transforming AI from an experiment into a productive tool for enterprise competition.
The Jevons Paradox in Action: The DeepSeek shock proved that falling AI costs didn't reduce spending; instead, lower barriers ignited a massive wave of "aggressive deployment" by global enterprises.
Scale of Capital Mobilization: This represents the largest capital mobilization for a single technology in modern economic history. Beyond tech giants, sovereign nations are also entering the fray, viewing "Sovereign AI" as central to national security and economic competitiveness.
III. Rivalry of Titans: OpenAI vs Anthropic
Anthropic: The "Cash King" of the Enterprise Market
Stunning Data: Surging from $1 billion to $30 billion in revenue in just 15 months, setting a new record in B2B software history.
Ace in the Hole: Commitment to an enterprise-first strategy; Claude Code (AI Agent coding tool) has achieved a $2.5 billion annualized revenue run rate.
OpenAI: The First Mover's Transformation Burden
Current State: Despite 900 million weekly active users, it faces staggering annual losses of approximately $14 billion. The completion of its $852 billion valuation financing round in April 2026 serves as a crucial "lifeline" to achieve cash flow positivity.
IV. The AI Agent Revolution: The Real Productivity Inflection Point
2026 is the year of the AI Agent, marking the substantive penetration of AI into business operations.
- Core Difference: Chatbots rely on human guidance (reactive questioning); AI Agents autonomously plan their path (goal-oriented, automatically calling tools, sending emails, running code, and delivering results).
- Tangible Results: The AI Agent market exceeded $9 billion in 2026. Early adopters have seen workflow speeds increase by 20%-30% and back-office operational costs decrease by approximately 25%.
- Market Leaders: Agentforce (Salesforce) has an ARR of $540 million; Copilot Studio (Microsoft), leveraging its deep ecosystem, enables seamless workplace integration for AI Agents.
V. Key Listed Companies to Watch
1. NVIDIA (NVDA) – The "Master Valve"
- Core Logic: The primary beneficiary of $600 billion in cloud provider capital expenditure.
- Moat: Not the chip itself, but the CUDA ecosystem. The newly released Vera Rubin platform aims to reduce inference costs by another 10x, making running AI Agents as cheap as turning on a light.
2. Palantir (PLTR) – The "Cockpit"
- Position: It's the application platform built on top of the infrastructure, helping governments and the military truly leverage AI Agents for decision support.
- Key Point: FY2026 revenue guidance exceeds $7.1 billion. Its AIP platform is irreplaceable in high-risk (e.g., military, medical) scenarios requiring auditability.
3. Alphabet/Google (GOOGL) – The Master Integrator
- Advantage: Possesses everything from TPU chips to the Gemini model and distribution channels reaching billions of users globally.
- Scale: Google Cloud operated at over $70 billion scale in 2025, successfully hedging the impact of AI on traditional search advertising models through enterprise-grade AI Agent services.
VI. Key Catalysts for the Second Half of 2026
The market will not only price newly listed giants but also “grade” AI’s return on investment.
- Anthropic IPO: Expected in October 2026. The gross margins disclosed in its AI Agent business prospectus will directly define the valuation benchmark for the entire SaaS industry.
- The AI Agent ROI Test: By the end of 2026, if enterprises cannot significantly improve profit margins through AI Agents, capital expenditure may face a cyclical pullback.
VII. Investment Risk Warnings
Current AI investment faces three core tests: First, the "Monetization Gap," the chasm between hyperscale cloud providers' over $600 billion in CapEx and the current ~$25 billion in direct AI revenue; if ROI disappoints, sector valuations will face sharp corrections. Second, Regulatory and Compliance Pressures, as the EU's AI Act begins phased enforcement from August 2026, significantly increasing corporate compliance costs. Third, the Technology Implementation Bottleneck, where approximately 62% of enterprises remain on the sidelines regarding deep deployment due to "model hallucinations," and Gartner predicts about 40% of early AI Agent projects may fail due to a lack of governance. Investors must be wary of valuations prematurely discounting future expectations.
Summary:
Conservative investors focus on NVIDIA or QQQ for certainty; aggressive investors research Anthropic or Palantir to capture high growth. In 2026, AI is no longer about storytelling; it’s about whose Agent can generate more profit.
BIT's US stock business directly connects with licensed brokerages, covering all core US stocks and ETFs. It supports stablecoin deposits and withdrawals, helping crypto users capture the 2026 tech dividend with one click. Services may differ based on jurisdictional restrictions and are unavailable in certain regions (including but not limited to Hong Kong).
Data as of April 2026. Sources include: Statista, Cargoson, McKinsey & Company, OpenAI (April 2026 funding announcement and CFO Sarah Friar's disclosed data), Anthropic (Series G financing in February 2026 and April revenue disclosure), Sacra, SaaStr, Remio AI, Alphabet Inc. (Q4 2025 earnings/Form 8-K), NVIDIA (FY2026 earnings/Form 8-K), Palantir Technologies (Q4 2025 earnings), PIIE (Peterson Institute for International Economics), CNBC, The Next Web, VentureBeat, The National Interest, Gartner, IDC, Joget, Tech-Insider, European Commission, Official AI Act documents, Crowell and Moring, AI Daily News.
Disclaimer: This report is for informational purposes only and does not constitute investment advice. Past performance does not guarantee future results. Investing involves risk, including the potential loss of principal. Clients should consult a qualified financial advisor before making any investment decisions.
