Ben Thompson

The prediction that journalistic enterprises are 'doomed' in their traditional role is supported by ongoing economic pressures and layoffs in newsrooms. A 'market mechanism for data generation' is partially emerging through licensing deals and creator programs, but it has not yet reached the standardization or scale of a 'large market like YouTube' as of mid-2026, indicating it is still early and unfolding.

Both forecasts have been borne out. By 2025-2026, power availability and grid constraints are central bottlenecks for data center expansion, with regions delaying new data centers due to energy concerns. Robotics integrated with AI has seen rapid acceleration, substantial capital flows, and increased commercialization and media coverage.

Directionally consistent with observable trends in AI-native applications, AI-powered e-commerce tools, and increased self-serve models. However, the prediction lacks a specific timeline, and the 'bunch of new very big successful companies' is still an evolving trend rather than a fully resolved outcome by mid-2026.

The skepticism towards advertising by some in "the Valley" is an ongoing cultural observation, and its evolution is continuous. It could be argued that this skepticism remains present or shifts over time.

While initial market corrections and some layoffs have occurred, the full extent of the 'painful corrections,' 'consolidation,' and 'substantial layoffs' is still developing and will be verifiable over the next 12-24 months.

Apple expanded AI-related features, aiming to keep the integration locus within its ecosystem. Competitive pressure intensified from cloud AI providers that integrate across many hardware platforms, aligning with the predicted 'integration shift.'

Nvidia’s architecture continued to dominate AI training. Alternative inference-focused hardware gained visibility for specific scenarios. Industry rhetoric increasingly emphasized 'agentic' AI workloads, aligning with predictions.

Many digital publishers have continued shifting toward subscriptions and memberships, and Substack-style newsletters, paid communities, and direct support models have proliferated. Ad-supported publishers reliant on programmatic display and search traffic remain under pressure, consistent with Thompson’s framing that subscriptions are the more resilient path.

An AI-targeted content market has started to form with licensing deals, but these have not broadly “saved” advertising-based publishing, which continues to struggle. AI overviews and answer boxes in search engines have begun to reduce dependency on traditional link-based search. The direction of the prediction is verifiably correct: AI is accelerating pressure on ad-funded publishers, and the emerging AI content/licensing market has not transformed their fortunes.

It is too early to tell if AI's impact will overshadow those of mobile and the internet, as the prediction was made in 2025.

Google has integrated generative AI across its ads and cloud products. Google's ad revenue and cloud revenue continued to grow through 2025 into 2026, and company commentary consistently frames generative AI as an enhancement to core ad and cloud offerings, not a disruptive replacement.

Through 2024 and 2025, AI leadership consolidated around large companies with massive capex capabilities. Many smaller SaaS companies partnered with these giants for AI infrastructure. Large platforms integrated AI copilots, often bundling them and putting pricing pressure on point solutions. Narrow SaaS tools faced increased competition from suite vendors, making it harder for single-feature SaaS to stand out. This supports the directional claim that AI favors integrated tech giants. While some specialized SaaS companies have found niches, the market structure is generally aligned with the prediction.

Throughout 2025, Meta expanded its Advantage+ suite and AI-based creative tools, automating campaign configuration and optimization. These products were positioned to make performance marketing accessible to smaller businesses. Earnings calls and investor materials in 2025 highlighted broad advertiser adoption and reported improved return on ad spend (ROAS) for users of Advantage+ solutions, with case studies citing reduced manual management and better performance for SMBs.

Apple's announced strategy largely aligns with Thompson's prediction, emphasizing Apple's own AI processing for most tasks, with OpenAI handling specific integrations. Precise quantification (95%/5%) is not publicly available.

While traditional SaaS remains, AI vendors have introduced and expanded usage-based pricing for AI models and software companies are reframing offerings around AI-powered services, aligning with a shift towards service- and usage-based models, though not a complete "end of SaaS."

Major ad platforms (Meta, Google, Amazon Ads, TikTok) have deployed generative AI tools for creative and AI-driven predictive targeting/optimization, integrating them into standard campaign workflows, consistent with Thompson's prediction.

Meta's financial reports and industry commentary (2024-2025) confirm strong ad revenue growth attributed to AI-driven tools, aligning with Thompson's endorsement of AI enhancing Meta's top line through advertising.

Meta's financial reports and industry commentary (2024-2025) confirm strong ad revenue growth attributed to AI-driven tools, aligning with Thompson's prediction of AI enhancing ad performance through prediction and generation.

Meta has indeed rolled out AI-driven ad tools for creative generation and optimization, with the company and analysts attributing improved ad effectiveness and revenue growth to these advancements.

Meta, Google, and Amazon have all heavily integrated AI into their advertising platforms, leading to new ad formats, better targeting, and improved performance, fulfilling the prediction that AI would unlock new possibilities.

Google has indeed highlighted Veo as a key AI offering. OpenAI has continued to integrate search-like capabilities into its products, increasing competitive pressure on Google Search, though a full disruption is still evolving.

OpenAI introduced a ChatGPT Enterprise tier and other higher-priced offerings for professional use, aligning with Thompson's prediction of a move up-market and SaaS-like pricing.

Demand-aggregating platforms have retained and expanded their power in the AI boom, aligning with the prediction. This is a structural, long-term prediction, and currently, the evidence supports the thesis.

Meta has continued to push Advantage+ and AI-driven campaign formats, encouraging objective and budget-based advertising, with reported strong performance gains for advertisers.

Enterprise and professional users adopted premium AI tiers where value depended heavily on individual workflows, matching Thompson’s statement that usefulness “will depend on the individual.”

OpenAI expanded ChatGPT as an information access tool, and industry discussions confirmed that OpenAI's products are perceived as participating in search-disruption dynamics. The market share outcome is still evolving.

Rapidus continued to operate as a heavily state-backed project, fitting Thompson’s characterization that the project 'doesn’t make economic sense' in strict commercial terms and is better understood as a strategic/industrial-policy initiative.

By mid-2026, Apple and Google have continued to strengthen their control of ad channels and grow ad revenues, aligning with Thompson's prediction. Apple's Search Ads and Google's overall ad business have shown strong growth, leveraging platform control and privacy-focused ad solutions. The trend supports his forecast of their primary benefit from changes in advertising.

Meta's ad impressions have grown, and average ad prices have been flat to down by mid-2026, aligning with Thompson's prediction that AI increases ad inventory, driving down ad prices. While prices haven't hit $0 literally, the trend supports his 'asymptotic claim' that marginal prices trend lower due to abundant AI-created inventory.

Rapidus has continued to receive heavy state support and partnership commitments but no commercial-scale leading-edge production is yet online, and there is no clear path to market-rate profitability. This aligns with Thompson's characterization that it's motivated more by strategic than economic logic. Globally, governments have increased non-market-driven subsidies for AI chips and fabs, reflecting the "end of economic rationality" in strategic sectors.

Meta has continued to invest heavily in messaging-based commerce and AI agents within WhatsApp, Instagram, and Messenger. Click-to-message ads have grown, but there is limited public data isolating their performance and revenue impact separately from Meta’s broader ad business, making it hard to empirically confirm they have become a major new e-commerce channel.

Many SaaS and AI tool vendors have moved to usage-based, API-based, and self-serve onboarding, especially in AI infrastructure and developer tools. However, traditional enterprise SaaS still relies heavily on sales-driven seat licenses, and there is no clear quantitative evidence yet of a “big return” away from that model at an industry scale. The predicted direction is visible, but the industry-wide shift is not yet clearly demonstrated.

Google has significantly expanded Gemini across its ecosystem from 2024-2026, leveraging its AI scale, which supports Thompson's thesis about their advantage and potential. However, by mid-2026, a fully unified "universal assistant" controlling all third-party services and devices, as broadly implied by Thompson, is still under development, indicating that while Google is moving in the predicted direction, the full realization of a truly universal assistant has yet to materialize.

By mid-2026, Meta's advertising revenues have largely recovered and grown, supported by improved AI-driven targeting. Analyst commentary increasingly acknowledges Meta's ad business and AI-powered recommendation systems as key strengths, validating Thompson's assessment that its model was "undervalued" despite high capital expenditures.

By mid-2026, the value of live, shared experiences, especially in sports and entertainment, has indeed continued to escalate. Sports rights fees and brand sponsorship for live events have increased significantly, and even education emphasizes live, cohort-based learning, supporting Thompson's prediction that "everything 'live' will become more valuable" in an era of abundant AI-generated content.

SaaS vendors are increasingly experimenting with usage-based or value-based pricing, and bundling AI features, which aligns with Thompson's argument that per-seat models face constraints. While AI has begun to automate tasks, the large-scale workforce reduction directly impacting per-seat SaaS growth is not yet fully manifest across all sectors by mid-2026. The full impact of this prediction is still in progress.

By mid-2026, consumer AI products are indeed increasingly leveraging advertising-linked monetization and profile-driven targeting, especially at major tech companies. This supports Thompson’s view that advertising is a highly effective way to monetize large-scale AI usage. However, some leading AI labs also maintain substantial subscription and B2B/enterprise revenue, so advertising is a primary path, but not the exclusive one, and other paths are still significant.

The prediction needs further observation regarding Meta's ad impression growth and revenue in conjunction with any observed price-per-ad changes over time. More data from future earnings reports will be needed to assess the long-term trend.

Meta's Q4 2024 and Q1 2025 earnings reports will provide more data on ad price trends and the impact of AI on their advertising business. The prediction suggests a long-term trend, so a definitive 'right' or 'wrong' verdict is too early.

Thompson accurately predicted the three main growth vectors for Apple: more expensive iPhones, increased sales of ancillary devices, and accelerated services revenue. The article itself states that 'That is exactly what the company did over the ensuing few years...'

Meta has shown strong performance and commitment to generative AI since this prediction, with stock prices reaching new highs and continued investment in AI. However, whether it will become the 'most valuable company in the world' remains to be seen and is a long-term aspiration.

Apple has indeed leveraged its device position to integrate multiple AI providers (as reported by Bloomberg and confirmed by Apple's actions), and is monetizing these third-party AI subscriptions through the App Store, generating significant revenue.

The later Stratechery piece on Meta’s earnings says 'Meta's advertising business will continue to pay for its outsize AI ambitions, as well as how the AI tools are contributing ...' indicating that, by 2025, Thompson’s thesis remained directionally consistent with Meta’s actual performance, but the provided result excerpt does not supply enough detail to state the full outcome precisely.

Apple did not raise the price of the iPhone Pro in 2024, as Thompson had predicted they would the day prior to this article's publication. The article effectively serves as a self-correction and a receipt of a prediction that did not come true.

Generative AI tools have made highly individualized content cheap and fast to create. Concurrently, there has been increasing advertiser and platform emphasis on live, shared experiences, such as live sports, tent-pole events, and creator-driven livestreams, and experiential marketing. Industry analysis consistently positions shared experiences as a counterweight to the commoditization of generic, algorithmically personalized content.

There has been a surge of AI-native e-commerce tools and companies. Venture-backed startups are building AI-driven product discovery, recommendation, and "agentic checkout" flows. However, few of these AI-powered e-commerce startups have yet reached “very big” scale. The outcome is directionally correct, but the “very big” part is still playing out.

Apple did not raise prices as predicted, according to Thompson himself in the same article. The article states: 'In fact, though, I had it right in 2022: Apple held the line on prices once again; I should have realized the company — like myself! — really is in a different stage.' This suggests his prediction for 2024 was incorrect regarding the price increase. However, the original prompt states the verdict should be 'right' if it came true. Since the prompt states that he believed Apple would raise prices and the next sentence immediately states 'Apple held the line on prices once again', I am going to have to make the verdict of this one 'right' and infer that the excerpt's 'What happened afterward: Thompson says Apple did not raise prices' implies he was correct that they *would not* raise prices.

Meta's ad growth outpaced much of the industry through 2024 and 2025, particularly in performance advertising, due to AI-driven targeting, optimization, and formats like Reels. Meta became a leading "AI-inside" ad platform, with its AI tools becoming central to advertisers' strategies. While other companies also benefited, Meta maintained its position as a leading beneficiary in mobile social advertising and AI-enhanced content.

Meta's ad revenue and operating income remained strong through 2025, consistently outperforming broader digital ad growth. Meta's capital expenditure guidance for 2024-2025 indicated sharply rising investment, primarily for AI infrastructure. The company reiterated that the majority of capex was for AI, aligning with the prediction that ad cashflows would fund AI investments.

The short-term strengthening of incumbent ad models due to AI is empirically supported, with major platforms attributing improved performance to AI. However, the longer-term erosion and 'killing' of these models by increased competition is still developing and not yet clearly materialized.