HSBC raises 2026-2030e global AI revenue TAM by 38%, but sees intensifying divergence of strong B2B and weak B2C

Overview

This report updates HSBC’s forecast for the global AI total addressable market. The core change is that B2B enterprise AI adoption is accelerating much faster than previously expected, while B2C consumer AI user growth, subscription conversion, and advertising monetization are weaker than prior assumptions. The report believes compute costs and scale laws are driving the Western LLM market toward a highly concentrated structure. OpenAI is still likely to become one of the oligopolists, but it has recently lagged competitors such as Anthropic and Gemini in B2B and some user-growth metrics.

Core views

First, the global AI industry revenue forecast for 2026-2030e is raised by 38% overall, lifting the 2030 AI industry revenue forecast to USD920bn. Second, the B2B AI revenue forecast is raised by 74%, with Agentic AI, enterprise applications, and coding tools as the main incremental drivers, and Anthropic viewed as leading in B2B. Third, the B2C AI revenue forecast is lowered by 20%, as OpenAI’s user growth slows and the low-priced ChatGPT Go tier causes ARPU dilution, weighing on subscription and advertising revenue. Fourth, OpenAI’s estimated funding gap falls to USD77bn, but this assumes financing, equity value, and the path of compute investment all materialize as forecast.

Analysis framework

The report uses a TAM re-estimation framework, dividing AI revenue into two major segments: B2B and B2C. For B2B, it focuses on enterprise applications, Agentic AI, foundation-model enterprise market share, and AI coding-tool revenue. For B2C, it breaks revenue into subscriptions, advertising, user scale, paid conversion rates, and ARPU. Its analysis of OpenAI’s funding needs starts from net cash, FCF losses, completed equity financing, potential proceeds from disposing AMD shares, and compute-cost assumptions to derive the 2026-2030e funding gap.

Methodology notes

• Market size forecastingTAM segment re-estimation

  Split AI industry revenue into B2B and B2C, then separately update assumptions for adoption rates, user counts, pricing, and revenue.

  The B2B upward revision comes from accelerating enterprise AI, Agentic AI, and coding tools; the B2C downward revision comes from slowing user growth, low-priced subscription tiers, and uncertainty around monetizing ad inventory.

• Competitive landscapeLLM oligopoly structure assessment

  High compute costs, scale laws, and financing capability create barriers, driving concentration in the Western foundation-model market.

  The report argues that more compute leads to stronger models, larger user scale, and more revenue; absent regulatory intervention, leading vendors are more likely to recover sunk costs and achieve long-term RoIC above WACC.

• Corporate funding needsOAI funding gap bridge

  Derive OpenAI’s funding gap through 2030 from net cash, FCF losses, equity financing, and the value of saleable investments.

  HSBC cuts its estimate of OAI’s 2026-2030e FCF loss from USD278bn to USD221bn and, together with additional financing and a higher AMD share value, lowers the estimated funding gap from USD154bn to USD77bn.

• User monetizationSubscription and advertising revenue split

  Assess B2C AI revenue through user numbers, paid conversion rates, ARPU, and ad load.

  The report believes lower-priced tiers such as ChatGPT Go may help increase subscriber counts, but could also cause higher-priced Plus users to downgrade, thereby lowering blended ARPU and affecting total revenue.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• OpenAI / OAI

  The report focuses on estimating its revenue, FCF, compute costs, and funding gap.

  Strengths

  It remains the leader in B2C chatbots and is considered likely to become one of the Western LLM oligopolists; it also has important compute arrangements with Microsoft, AMD, Amazon, CoreWeave, and others.

  Weaknesses

  User growth is slowing, while B2C subscription and advertising monetization are under pressure; in B2B it lags Anthropic, and part of Azure sales revenue is no longer attributed to OAI.

  Comparison

  Compared with Anthropic, OAI is stronger in B2C but less leading in B2B; compared with Gemini, OAI still has a stronger brand and user base.

  Risks

  The funding gap still reaches USD77bn, and compute costs remain high; if an IPO or equity financing does not go smoothly, user growth continues to slow, or competition intensifies, valuation and the funding path may come under pressure.

• Anthropic

  Viewed by the report as an important leader in B2B AI and coding tools.

  Strengths

  B2B revenue is growing quickly, with Claude and Claude Code driving a significant increase in ARR, and the company has already filed an IPO-related registration statement.

  Weaknesses

  Its consumer-side revenue share remains lower than OpenAI’s, and long-term scaling still depends on enterprise adoption and financing markets.

  Comparison

  Ahead of OpenAI in B2B and coding tools; still behind ChatGPT in B2C brand and user scale.

  Risks

  If enterprise AI adoption slows, model-quality gaps converge, or competitors cut prices, revenue growth and market share may be affected.

• AMD.US

  Related to OAI’s 6GW GPU leasing arrangement; OAI can sell AMD shares after vesting to supplement funding.

  Strengths

  AMD’s share price rose from USD203.43 on March 31, 2026, to USD466.38 on June 5, 2026; the report estimates this could contribute USD8bn to improving OAI’s funding gap.

  Weaknesses

  Details on OAI-related GPU capacity buildout are limited, and HSBC conservatively models the first batch of capacity as contributing starting in 2027.

  Comparison

  Compared with NVIDIA, AMD provides a potential GPU alternative and equity upside mechanism within the OAI agreement.

  Risks

  Share-price volatility, uncertainty around capacity delivery, and changes in the AI-chip competitive landscape will affect potential value.

• NVIDIA

  As the leading GPU supplier, it is a major beneficiary of compute expansion for Western large models.

  Strengths

  The report believes Western models can access the most advanced NVIDIA GPUs, and larger compute scale drives improvements in model quality.

  Weaknesses

  HSBC does not include OAI’s 10GW LOI with NVIDIA in its model, because the probability of it converting into a definitive agreement is viewed as limited.

  Comparison

  Compared with Chinese model vendors subject to export restrictions, U.S. frontier models have an advantage in access to high-end GPUs.

  Risks

  Order fulfillment, regulatory restrictions, customer capex cycles, and alternative chip solutions will affect the degree of ongoing benefit.

• Microsoft / Amazon / Alphabet

  Cloud and AI distribution partners that affect OpenAI’s revenue sharing, compute procurement, and AI product commercialization.

  Strengths

  Microsoft’s new agreement with OpenAI caps revenue sharing; related transactions with Amazon and Alphabet are seen as partially offsetting changes in Azure sales.

  Weaknesses

  The agreement structures are complex, and revenue attribution and cost-sharing have a major impact on OAI’s long-term profitability path.

  Comparison

  Microsoft is the most deeply integrated into the OpenAI ecosystem; Amazon and Alphabet are more reflected as alternative sources of compute and distribution.

  Risks

  Changes in cloud-service contracts, revenue-share caps, RPO delivery, and competitive relationships could alter how benefits are distributed among the parties.

• B2B AI TAM

  The report’s main direction of upward revision.

  Strengths

  Enterprise AI, Agentic AI, coding tools, and consulting/integration services create clearer paid use cases.

  Weaknesses

  Enterprise process integration is complex, and deployment cycles and ROI validation still take time.

  Comparison

  Compared with B2C, B2B has higher revenue quality and a more positive growth revision.

  Risks

  If enterprise adoption falls short of expectations, tool commoditization rises, or price competition intensifies, the TAM upward revision may fail to materialize.

• B2C AI TAM

  The consumer AI market segment that the report lowers.

  Strengths

  Products such as ChatGPT, Claude, and Gemini have large-scale user reach and still have room for future subscription and advertising monetization.

  Weaknesses

  User switching costs are low, competition from free products is strong, and lower-priced subscription tiers may dilute ARPU.

  Comparison

  Compared with B2B, B2C has larger user scale but lower revenue certainty.

  Risks

  Stagnant user growth, weaker-than-expected paid conversion, constrained ad load, or regulatory restrictions could continue to lower revenue forecasts.

Key data

• 2030 AI industry revenue forecastUSD920bnThe previous forecast was USD728bn.
• 2026-2030e cumulative AI industry revenue revision+38%Driven jointly by a sharp upward revision in B2B and a modest downward revision in B2C.
• 2026-2030e B2B AI revenue forecast revision+74%Mainly driven by accelerating adoption of Agentic AI, enterprise applications, and coding tools.
• 2030 B2B AI TAMUSD708bnThe report says this grows from USD20bn in 2025 to USD708bn in 2030.
• 2026-2030e B2C AI revenue forecast revision-20%Mainly due to OpenAI user growth coming in below expectations and pressure on subscription conversion rates and ARPU.
• OAI 2026-2030e funding gapUSD77bnPreviously USD154bn; the report says it has been halved.
• OAI 2026-2030e FCF loss forecastUSD221bnPreviously USD278bn.
• OAI 2026-2030e compute cost assumptionUSD687bnAssumes compute reaches about 11GW by 2030, versus about 2GW currently.
• Anthropic ARR estimateUSD47bnHSBC’s latest estimate for May 2026, above the prior model’s USD20bn.
• OpenAI B2C 2030 revenue share assumption37%Previously assumed at 44%; the revenue share in May 2026 was about 57%.

Impact & implications

The implications for the AI value chain are that enterprise AI, B2B distribution of foundation models, AI coding tools, and compute infrastructure remain the clearest growth areas; consumer chatbots still have massive user scale, but revenue quality and monetization efficiency are below earlier optimistic assumptions. For listed assets, AMD, NVIDIA, Microsoft, Amazon, Alphabet, cloud service providers, and AI infrastructure suppliers may continue to benefit from compute and enterprise AI demand, but these benefits depend on the financing capacity, order delivery, and market-share changes of private model companies such as OpenAI and Anthropic.

Risks

• B2B AI adoption may come in below HSBC’s latest assumptions, causing the TAM upward revision to fail to materialize.
• B2C AI user growth, subscription conversion rates, and advertising monetization may continue to be weaker than expected.
• Although OpenAI’s funding gap has been lowered, it still requires support from an IPO, equity financing, asset disposals, or other financing methods.
• High compute investment is key to improving model quality, but it may also create long-term FCF pressure and financing dependence.
• Competition in the LLM market is intense, and Anthropic, Gemini, Meta, Chinese models, and Europe’s Mistral AI may alter the market-share landscape.
• Regulation, export restrictions, data compliance, and AI safety requirements may change model-training costs, market access, and revenue paths.
• If lower-priced subscription tiers trigger downgrades by higher-priced users, this could further reduce OpenAI’s and the industry’s B2C ARPU.

What to watch

• Whether OpenAI moves forward with an IPO or a new round of equity financing, and the valuation and financing size involved.
• Anthropic’s IPO progress, ARR disclosure, and growth in Claude enterprise customers.
• Whether ChatGPT WAU/MAU resumes growth, and the impact of ChatGPT Go on Plus-user downgrades.
• The actual paid penetration of B2B Agentic AI and coding tools in enterprises.
• Delivery progress on OpenAI’s compute and cloud-service agreements with Microsoft, AMD, Amazon, CoreWeave, and others.
• AI chip demand, supply, and order fulfillment for AMD and NVIDIA.
• Changes in user growth, model quality, and enterprise share for Gemini, Claude, and Chinese models.
• Commercial validation of AI ad inventory, ad ARPU, and advertising experiences in consumer chatbots.