Figma (FIG.US) gross profit under pressure Arm (ARM.US) revenue hits record high Wall Street admits AI returns are difficult to predict

According to Woofun AI, Figma (FIG.US) AI functionality has penetrated 80% of large customer workflows, and Arm (ARM.US) recorded a quarterly record of 1.44 billion US dollars in the second fiscal quarter, setting the tone for the industry. Despite significant increases in technology penetration, the pressure on Figma's gross margin and changes in the cost structure behind Arm's business transformation together outline the complex picture of the current AI infrastructure and application layer in the early stages of commercialization. Wall Street executives have publicly acknowledged that it is difficult to predict AI's long-term returns, and investors are beginning to require companies to prove how much profit technological advancements can turn into, rather than just relying on narrative-driven valuations.

Figma announced financial results for the third quarter of 2026. Revenue reached US$391.4 million, an increase of 43% year over year, and US$16.4 million above the previous guidance limit. The company raised its full-year revenue forecast to $1,497 million to $1,499 million, and its non-GAAP operating profit forecast to $158 million to $163 million.

However, the market did not fully buy it, and the stock price dropped by about 7% during after-hours trading. The company's revenue guidance for the fourth quarter is 402 million to 404 million US dollars, which means that the growth rate will slow to about 33% from 43% in the third quarter. This is the direct reason for the post-market pressure. The market is looking at not only gross profit margins, but also shifts in the growth rate itself. According to data compiled by Woofun AI, Figma's third quarter GAAP gross margin was 80%, and non-GAAP gross margin was 82%, which is lower than 86% in the same period last year. The decline in gross margin is mainly due to increased inference costs due to increased AI usage. The company is increasing the proportion of tasks carried out by self-developed models to improve cost efficiency, but has not disclosed specific savings. Figma has proven that large customers are willing to continue using AI, and the next step is to prove how much incremental profit this usage can generate.

If AI credit revenue grows faster than inference costs, Figma may establish a new revenue stream; conversely, AI may also become a service cost that continues to squeeze gross margins.

The usage data of AI products is the most noteworthy part of this financial report. As of October 31, over 80% of paying customers with annualized contracts worth over $10,000 spent AI credits each week, and 65% used Figma Agent every week. Overall usage of Figma MCP Server in the third quarter increased 3x month-on-month, with write-back operations increasing 4x. The company also launched the Code Layers public beta in October, which allows users to build and edit code-based prototypes directly within the design canvas. These changes are redefining Figma's product boundaries.

The core of traditional design tools is to make the interface more efficient, while Agent, MCP, and Code Layers allow software to directly participate in design, code generation, and modification. Figma is trying to evolve from a design editor to a work platform connecting design and development. Business models are also changing. In the past, Figma mainly earned revenue through seat subscriptions; now, AI credits have added a billing method linked to actual usage. The Code Layers public beta launched in October further blurred the boundaries between design and development, making 80% of high-value customers not only use AI-assisted design, but also began to rely on AI for code-level interaction.

This shift from a “tool” to a “platform” has made Figma's revenue structure more complex and placed higher demands on its cost control. Although 82% of non-GAAP gross margin is still high, it has declined from a historical high of 86%, indicating that the initial cost pressure for large-scale AI applications is showing.

Arm announced results for the second quarter of fiscal year 2027 ending September 30. Revenue reached US$1.44 billion, up 27% year over year, the highest level in the company's history for the second fiscal quarter. 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 deeper changes. In the past, the company mainly earned revenue by licensing architectures to chip designers, and now it is directly supplying self-designed Arm AGI CPUs. The company said it has received purchase orders from several customers continuing into the natural year 2027, and the first commercial deliveries are scheduled to take place in the fourth quarter of fiscal year 2027. Management anticipates that the product could contribute up to $2 billion in revenue in the 2027 and 2028 fiscal years combined. This is the company's forecast, not confirmed sales revenue. Arm also revealed that among major hyperscale cloud service providers, the share of Arm-based servers increased from 15% in 2024 to over 50% in 2025.

This statistic is for specific large cloud service providers and cannot be directly considered the global server market share. These numbers reflect the proliferation of AI infrastructure requirements. GPUs are responsible for a large amount of parallel computing, but CPUs are still responsible for general computing, task scheduling, and system management. As AI agents run programs, call tools, and execute complex workflows, CPU performance and energy efficiency will also impact the overall cost of the data center. For Arm, the opportunities and risks expand at the same time. Selling homegrown CPUs may generate more revenue per customer than licensing the architecture, but it also means that companies must bear the risk of chip delivery, supply chain, and product competition. Arm is shifting from mainly selling intellectual property rights in chip design to participating in the chip product market at the same time.

At the Hong Kong International Financial Leaders Investment Summit, Goldman Sachs (GS.US) CEO David Solomon, Citadel Securities CEO Peng Zhao, and TPG (TPG.US) founding partner Jim Coulter discussed the long-term development of AI. Some commentators pointed out that the most noteworthy aspect of this summit was not the optimism of financial institutions about AI, but rather that many executives publicly acknowledged that there is huge uncertainty about long-term predictions. Solomon believes that demand for short-term computing power is still strong, but the long-term development path is difficult to determine.

Peng Zhao raised a more direct question: technology is evolving faster than the investment cycle, and capital markets must make huge investment decisions without reliable long-term predictions. Coulter, on the other hand, reminded investors that experience during the Internet bubble showed that even if a correct judgment is that a technology will change the world, it is possible to lose money by betting on the wrong company. This reveals a structural contradiction in AI investments. Data centers, chips, and power infrastructure require capital investment years in advance, but model costs, application methods, and final business models may change within a few months. The infrastructure investment cycle does not match the technology iteration cycle.

At the summit's welcome dinner on November 2, Chinese AI companies offered a different perspective. Moonshot founder Yang Zhilin responded to the dispute between open source and closed source during the “Technology for Good” conversation hosted by Shen Nanpeng, and discussed the model's ability to recursively improve itself and the path of generalization to the physical world. Compared with simply expanding investment in computing power, embedding AI into existing industries may provide a more direct path for verifying commercial value — Hong Kong Financial Secretary Chen Maobo said at the luncheon that factories, hospitals, and logistics networks in mainland China are fertile ground for AI applications. But app value doesn't automatically equal return on investment. Model providers, cloud computing companies, chip makers, and application developers may all create value, but the final profit may not be evenly distributed. This is also the most important investment issue for the AI industry right now: the market already believes AI will transform the economy, but it's uncertain who will get the benefits that match the capital investment.

The total market capitalization of the five major US tech giants increased by nearly $800 billion in a single day. After the Democratic Party won both houses of Congress, Amazon (AMZN.US), Apple (AAPL.US), Microsoft (MSFT.US), Alphabet (GOOGL.US), and Meta (META.US) collectively rose. Investors believe that if data center construction projects are restricted, tech giants that already have huge computing power infrastructure may benefit. Amazon rose about 6.1% and Apple rose about 5.3%, in contrast to the decline of some data center operators. The clear election results were compounded by a sharp drop in oil prices. The Dow Jones Index rose 1,569 points, or 3.03%, on November 4, and the S&P 500 rose about 1.6%. However, the energy sector fell 4.77% due to a sharp drop in oil prices. It was the worst performing sector of the day. Non-essential consumption and finance led the rise.

Brent crude fell approximately 8.7% to $94.62 a barrel. The fall in energy prices has allayed market concerns about inflation, but the war in Iran and changes in US politics may still affect subsequent trends. American voters are starting to outright veto AI data center expansion. In the 2026 midterm elections, voters in many places supported candidates or policy measures to limit data center construction. Texas Democratic gubernatorial candidate Gina Hinojosa unexpectedly defeated incumbent Governor Abbott to become the state's first Latino governor and the first Democratic governor since 1990. She advocated suspending approval for new data centers, and this victory became the focus of attention in the AI infrastructure industry. The issue of data center electricity consumption, water use, and driving up electricity prices for residents is evolving from a local dispute to a national political issue.

Uber (UBER.US) is experiencing variables in the field of autonomous driving, but the direction is not a contraction of its own. Waymo Co-CEO Tekedra Mawakana has confirmed that it will end its exclusive Robotaxi partnership with Uber in Austin and Atlanta in January 2028 and use its own app to take orders, while the original agreement was supposed to run until May 2028. Uber said it was “not Uber's choice” to open an exclusive portal for Robotaxi within its app, and emphasized that it is still cooperating with many autonomous driving companies: the company's third quarter conference call said that by the end of the year, about 700 fully driverless vehicles will be launched, and there will be more than 2,000 in 2027. The real pressure comes from the supply side and the main travel industry. CEO Dara Khosrowshahi said that high fuel prices combined with limited driver supply dragged down supply growth in many international markets; travel bookings in the third quarter were 29.512 billion US dollars, up 18% year on year, but revenue was 7.520 billion US dollars, down 2% year on year, mainly due to changes in the business model in the UK market. Overall, total bookings for the quarter increased by 20%, non-GAAP earnings per share of $0.88, and a year-on-year increase of 35%, all higher than market expectations. The key to commercializing autonomous driving remains operational efficiency, not just technological maturity.

California's Proposition 40 plans to levy a one-time 5% wealth tax on residents with net assets over $1 billion. Preliminary results showed that 55.6% of the votes were against it, and the proposal faced failure. The proposal previously attracted the attention of Silicon Valley entrepreneurs and investors, and disputes focused on trade-offs between wealth taxation, capital migration, and local finance.