Figma's AI features have entered the daily workflows of 80% of its large customers, but gross margins remain under pressure. Arm posted a record quarterly revenue of $1.44 billion, with data center royalties doubling. At the Hong Kong financial summit, Wall Street executives openly acknowledged the difficulty of predicting AI's long-term returns. The AI industry's growth continues, but investors are beginning to demand that companies prove how much profit technological progress can actually translate into.
1 | Figma revenue grows 43%, commercialization enters profit validation phase
Figma reported its fiscal 2026 third-quarter results, with revenue reaching $391.4 million, up 43% year-over-year and $16.4 million above the upper end of its prior guidance. The company raised its full-year revenue outlook to $1.497 billion to $1.499 billion, and its non-GAAP operating profit outlook to $158 million to $163 million. But the market was not fully convinced, and the stock fell notably at one point in after-hours trading.
AI product usage data was the most noteworthy part of this earnings report. As of October 31, among paid customers with annualized contract value exceeding $10,000, more than 80% consumed AI credits weekly, and 65% used Figma Agent weekly. Overall usage of Figma MCP Server tripled quarter-over-quarter in the third quarter, with write-back operations growing 4x. The company also launched Code Layers in public beta in October, allowing users to build and edit code-based prototypes directly on the design canvas.
These changes are redefining Figma's product boundaries. The core of traditional design tools is to help people draw interfaces more efficiently, while Agent, MCP, and Code Layers allow the software to directly participate in design, code generation, and modification. Figma is trying to evolve from a design editor into a work platform connecting design and development.
The business model is also changing. In the past, Figma mainly generated revenue through seat subscriptions; now, AI credits have added a charging method tied to actual usage. But growth in AI usage also means more inference costs. Figma's third-quarter GAAP gross margin was 80%, and non-GAAP gross margin was 82%, the latter lower than 86% in the same period last year. The company is increasing the share of tasks handled by its own models to improve cost efficiency, but has not disclosed specific savings.
Figma has already proven that large customers are willing to keep using AI. What it needs to prove next is how much incremental profit this usage can create. If AI credit revenue grows faster than inference costs, Figma may establish a new revenue source; otherwise, AI could also become a service cost that continuously squeezes gross margins.
(Source: Figma Q3 2026 official earnings report / Reuters)
2 | Arm revenue grows 27%, data center royalties double, competition in AI infrastructure is expanding to CPUs
Arm reported results for the second quarter of fiscal year 2027 ended September 30, with revenue reaching a record $1.44 billion, up 27% year-over-year. Licensing and other revenue was $773 million, up 50% year-over-year; royalty revenue was $670 million, up 8% year-over-year. Among them, data center royalty revenue doubled year-over-year.
Arm's business is undergoing a deeper transformation. The company previously generated revenue mainly by licensing its architecture to chip designers, and now it has begun directly offering its own designed Arm AGI CPU. The company said it has secured purchase orders from multiple customers extending through calendar year 2027, with initial commercial deliveries planned for the fourth quarter of fiscal year 2027. Management expects the product could contribute up to $2 billion in revenue combined across fiscal years 2027 and 2028. This is a company forecast, not confirmed sales revenue.
Arm also disclosed that among major hyperscale cloud service providers, the share of Arm-based servers rose from 15% in 2024 to over 50% in 2025. This statistic only covers specific large cloud service providers and cannot be directly regarded as global server market share.
These figures reflect the diffusion of AI infrastructure demand. GPUs handle massive parallel computing, but CPUs still handle general-purpose computing, task scheduling, and system management. As AI Agents need to run programs, call tools, and execute complex workflows, CPU performance and energy efficiency will also affect the overall cost of data centers.
For Arm, both opportunities and risks are expanding. Selling self-designed CPUs may bring greater per-customer revenue than architecture licensing, but it also means the company must bear chip delivery, supply chain, and product competition risks. Arm is shifting from mainly selling chip design intellectual property to also participating in the chip product market.
(Source: Arm official earnings report and shareholder letter / Wall Street Journal)
3 | Wall Street executives admit difficulty predicting AI returns, technological progress is outpacing capital's ability to judge
At the Hong Kong International Financial Leaders Investment Summit, financial figures including Goldman Sachs CEO David Solomon, Citadel Securities CEO Peng Zhao, and TPG founding partner Jim Coulter discussed the long-term development of AI. A commentary published by Reuters Breakingviews at 06:00 Beijing time on November 5 noted that the most noteworthy aspect of the summit was not financial institutions' optimism about AI, but that multiple executives publicly acknowledged enormous uncertainty in long-term forecasts.
Solomon believes short-term compute demand remains strong, but the long-term development path is difficult to judge. Peng Zhao raised a more direct question: the pace of technological development has outpaced investment cycles, and capital markets must make massive investment decisions without reliable long-term forecasts. Coulter reminded investors that lessons from the dot-com bubble show that even correctly judging that a technology will change the world can still result in financial losses from betting on the wrong companies.
This reveals a structural contradiction in AI investment. Data centers, chips, and power infrastructure require capital investment years in advance, but model costs, application methods, and ultimate business models can change within months. The investment cycle for infrastructure does not match the technology iteration cycle.
At the summit, Chinese AI companies offered another perspective. Moonshot founder Yang Zhilin discussed open-weight models and China's manufacturing application advantages. Compared with simply scaling up compute investment, embedding AI into existing industries may provide a more direct path to validating commercial value.
But application value does not automatically equal investment returns. Model providers, cloud computing companies, chipmakers, and application developers may all create value, but the ultimate profits may not be distributed evenly.
This is also the most important investment question facing the AI industry today: the market already believes AI will transform the economy, but it cannot yet determine who will capture returns commensurate with capital investment.
(Source: Reuters)
Also worth knowing ↓
U.S. tech giants added approximately $750 billion in market value in a single day. After Democrats won both chambers of Congress, shares of Amazon, Apple, Microsoft, Alphabet, and Meta all rose collectively. Investors believe that if new data center construction projects are restricted, tech giants that already own massive compute infrastructure could actually benefit. Amazon rose about 7.4%, Apple nearly 7%, in contrast to declines in some data center operators. (Source: Reuters)
Falling U.S. oil prices pushed the Dow up more than 1,500 points. On November 4, the Dow Jones Industrial Average rose 1,569 points, and the S&P 500 gained about 1.6%. Brent crude fell about 8.7% to $94.62 per barrel. Retreating energy prices eased market concerns about inflation, but the Iran war and U.S. political changes could still affect subsequent trends. (Source: Associated Press)
U.S. voters are beginning to directly reject AI data center expansion. In the 2026 midterm elections, voters in multiple locations supported candidates or policy measures that restrict data center construction. Texas Democratic gubernatorial candidate Gina Hinojosa's upset victory became a focus of attention for the AI infrastructure industry. The issues of data center power consumption, water use, and upward pressure on residential electricity prices are evolving from local disputes into a national political issue. (Source: Axios / Reuters)
Uber slows down large deals in the autonomous driving space. According to Wall Street Journal market reports, Uber is scaling back its autonomous driving partnership footprint, while the company faces insufficient driver supply and pressure on revenue from its core ride-hailing business. The key to autonomous driving commercialization remains operational efficiency, not just technological maturity. (Source: Wall Street Journal)
California's proposed 5% billionaire wealth tax faces failure. Preliminary midterm election results show insufficient support for this one-time wealth tax proposal targeting the super-rich. The proposal had previously drawn attention from Silicon Valley entrepreneurs and investors, with controversy centered on the trade-offs among wealth taxation, capital migration, and local fiscal finances. (Source: Business Insider)
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
